<?xml version="1.0" ?><!DOCTYPE pmc-articleset PUBLIC "-//NLM//DTD ARTICLE SET 2.0//EN" "https://dtd.nlm.nih.gov/ncbi/pmc/articleset/nlm-articleset-2.0.dtd"><pmc-articleset><article xml:lang="en" article-type="review-article" dtd-version="1.4"><processing-meta base-tagset="archiving" mathml-version="3.0" table-model="xhtml" tagset-family="jats"><restricted-by>pmc</restricted-by></processing-meta><front><journal-meta><journal-id journal-id-type="nlm-ta">Eur Respir Rev</journal-id><journal-id journal-id-type="iso-abbrev">Eur Respir Rev</journal-id><journal-id journal-id-type="pmc-domain-id">4270</journal-id><journal-id journal-id-type="pmc-domain">err</journal-id><journal-id journal-id-type="publisher-id">ERR</journal-id><journal-title-group><journal-title>European Respiratory Review</journal-title></journal-title-group><issn pub-type="ppub">0905-9180</issn><issn pub-type="epub">1600-0617</issn><publisher><publisher-name>European Respiratory Society</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="pmcid">PMC13014283</article-id><article-id pub-id-type="pmcid-ver">PMC13014283.1</article-id><article-id pub-id-type="pmcaid">13014283</article-id><article-id pub-id-type="pmcaiid">13014283</article-id><article-id pub-id-type="pmid">41881453</article-id><article-id pub-id-type="doi">10.1183/16000617.0151-2025</article-id><article-id pub-id-type="publisher-id">ERR-0151-2025</article-id><article-version-alternatives><article-version article-version-type="pmc-version">1</article-version><article-version article-version-type="VoR" vocab="JAV" vocab-identifier="http://www.niso.org/publications/rp/RP-8-2008.pdf">Version of Record</article-version></article-version-alternatives><article-categories><subj-group subj-group-type="heading"><subject>Reviews</subject></subj-group><subj-group subj-group-type="hwp-journal-coll"><subject>4</subject></subj-group></article-categories><title-group><article-title>Biomarker tests of progression from tuberculosis infection to disease: a systematic review</article-title></title-group><contrib-group><contrib contrib-type="author"><name name-style="western"><surname>Tingley</surname><given-names initials="K">Kylie</given-names></name><xref rid="af1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Li</surname><given-names initials="A">Angela</given-names></name><xref rid="af2" ref-type="aff">2</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">https://orcid.org/0000-0001-9272-0985</contrib-id><name name-style="western"><surname>MacLean</surname><given-names initials="EL">Emily L.</given-names></name><xref rid="af3" ref-type="aff">3</xref><xref rid="af4" ref-type="aff">4</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Skidmore</surname><given-names initials="B">Becky</given-names></name><xref rid="af5" ref-type="aff">5</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Maredia</surname><given-names initials="N">Nawal</given-names></name><xref rid="af2" ref-type="aff">2</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Ravikumar</surname><given-names initials="R">Rajalakshmi</given-names></name><xref rid="af2" ref-type="aff">2</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Alvarez</surname><given-names initials="GG">Gonzalo G.</given-names></name><xref rid="af2" ref-type="aff">2</xref><xref rid="af6" ref-type="aff">6</xref><xref rid="af7" ref-type="aff">7</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Matteelli</surname><given-names initials="A">Alberto</given-names></name><xref rid="af8" ref-type="aff">8</xref></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid" authenticated="false">https://orcid.org/0000-0001-6641-0094</contrib-id><name name-style="western"><surname>Sulis</surname><given-names initials="G">Giorgia</given-names></name><xref rid="af2" ref-type="aff">2</xref><xref rid="af6" ref-type="aff">6</xref><xref rid="cor1" ref-type="corresp"/></contrib></contrib-group><aff id="af1"><label>1</label>Independent Epidemiologist, River Bourgeois, NS, Canada</aff><aff id="af2"><label>2</label>School of Epidemiology and Public Health, Faculty of Medicine, University of Ottawa, Ottawa, ON, Canada</aff><aff id="af3"><label>3</label>NHMRC Clinical Trials Centre, Faculty of Medicine and Health, University of Sydney, Sydney, Australia</aff><aff id="af4"><label>4</label>Sydney Infectious Diseases Institute, University of Sydney, Sydney, Australia</aff><aff id="af5"><label>5</label>Independent Information Specialist, Ottawa, ON, Canada</aff><aff id="af6"><label>6</label>Ottawa Hospital Research Institute, Ottawa, ON, Canada</aff><aff id="af7"><label>7</label>Division of Respirology, Department of Medicine, The Ottawa Hospital, Ottawa, ON, Canada</aff><aff id="af8"><label>8</label>Institute of Infectious and Tropical Diseases, Department of Clinical and Experimental Sciences, WHO Collaborating Centre for TB Prevention, University of Brescia, Brescia, Italy</aff><author-notes><corresp id="cor1">Corresponding author: Giorgia Sulis (<email>gsulis@uottawa.ca</email>)
</corresp></author-notes><pub-date pub-type="collection"><month>1</month><year>2026</year></pub-date><pub-date pub-type="epub"><day>25</day><month>3</month><year>2026</year></pub-date><volume>35</volume><issue>179</issue><issue-id pub-id-type="pmc-issue-id">505159</issue-id><elocation-id>250151</elocation-id><history><date date-type="received"><day>26</day><month>6</month><year>2025</year></date><date date-type="accepted"><day>21</day><month>1</month><year>2026</year></date></history><pub-history><event event-type="pmc-release"><date><day>25</day><month>03</month><year>2026</year></date></event><event event-type="pmc-live"><date><day>26</day><month>03</month><year>2026</year></date></event><event event-type="pmc-last-change"><date iso-8601-date="2026-03-26 09:25:13.630"><day>26</day><month>03</month><year>2026</year></date></event></pub-history><permissions><copyright-statement>Copyright ©The authors 2026</copyright-statement><copyright-year>2026</copyright-year><license><ali:license_ref xmlns:ali="http://www.niso.org/schemas/ali/1.0/" specific-use="textmining" content-type="ccbynclicense">https://creativecommons.org/licenses/by-nc/4.0/</ali:license_ref><license-p>This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact <email xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="mailto:permissions@ersnet.org">permissions@ersnet.org</email></license-p></license></permissions><self-uri xmlns:xlink="http://www.w3.org/1999/xlink" content-type="pmc-pdf" xlink:href="ERR-0151-2025.pdf"/><abstract><sec><title>Background</title><p>Tuberculosis (TB) remains a global health threat, with millions of new infections annually. While most infected individuals remain asymptomatic, a subset progresses to active disease. Effective tests predicting TB progression are urgently needed to enhance prevention strategies and reduce the global TB burden.</p></sec><sec><title>Objective</title><p>To evaluate biomarker-based tests designed to predict progression from TB infection to active disease.</p></sec><sec><title>Methods</title><p>We conducted a systematic literature search of four electronic databases for studies published between 2016, when the first trial evaluating biomarker-based tests for TB progression was published, and 2024. Quality assessment was performed using QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies 2). Data on study design, population, biomarkers, and diagnostic accuracy were extracted and synthesised narratively.</p></sec><sec><title>Results</title><p>19 studies were included, reporting 70 biomarker-based tests for predicting TB progression. These included 23 gene signatures, 14 protein signatures and 33 cytokines/chemokines. Heterogeneity in study populations, methodologies and outcome definitions precluded direct comparisons. Many studies lacked transparency in reporting key population characteristics, reference standards and diagnostic accuracy outcomes, limiting clinical applicability. Most tests demonstrated only moderate predictive accuracy, with no single biomarker approach emerging as a definitive tool for clinical use.</p></sec><sec><title>Conclusions</title><p>No biomarker-based test to predict TB progression is currently ready for clinical implementation. Standardised methodologies, larger validation studies and improved reporting transparency are necessary to advance TB diagnostics and improve early detection efforts. Future research should prioritise refining tests for TB progression by improving their predictive accuracy, identifying appropriate target populations and evaluating their cost-effectiveness and feasibility for integration into TB prevention programmes, particularly in high-burden settings.</p></sec></abstract><abstract abstract-type="short" id="abstract-1"><title id="title-1">Shareable abstract</title><p><bold>Current biomarker-based tests for predicting tuberculosis progression show promise but lack consistency and clinical readiness. Standardised methodologies, transparent reporting and large-scale validation studies are crucial to advancing TB diagnostics.</bold>
<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://bit.ly/3NFuokV" ext-link-type="uri">https://bit.ly/3NFuokV</ext-link></p></abstract><funding-group><award-group id="funding-1"><funding-source>
<institution-wrap><institution>Canada Research Chairs</institution><institution-id institution-id-type="doi">http://dx.doi.org/10.13039/501100001804</institution-id></institution-wrap>
</funding-source><award-id>Tier 2 CRC in Communicable Disease Epidemiology (P</award-id></award-group><award-group id="funding-2"><funding-source>
<institution-wrap><institution>Canadian Institutes of Health Research</institution><institution-id institution-id-type="doi">http://dx.doi.org/10.13039/501100000024</institution-id></institution-wrap>
</funding-source><award-id>CIHR Fellowship n. 472823 (MacLean)</award-id></award-group><award-group id="funding-3"><funding-source>
<institution-wrap><institution>New Diagnostics Working Group (NDWG)</institution></institution-wrap>
</funding-source><award-id>NDWG is a multi-sector group that aims to progress</award-id></award-group></funding-group><custom-meta-group><custom-meta><meta-name>pmc-status-qastatus</meta-name><meta-value>0</meta-value></custom-meta><custom-meta><meta-name>pmc-status-live</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-status-embargo</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-status-released</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-open-access</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-olf</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-manuscript</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-legally-suppressed</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-pdf</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-supplement</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-pdf-only</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-suppress-copyright</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-real-version</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-scanned-article</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-preprint</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-in-epmc</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-license-ref</meta-name><meta-value>CC BY-NC</meta-value></custom-meta></custom-meta-group></article-meta></front><body><sec id="s1"><title>Background</title><p>In 2024, tuberculosis (TB) stood once again as the leading cause of death from a single infectious agent globally, with 10.7 million new cases and 1.23 million fatalities annually [<xref rid="C1" ref-type="bibr">1</xref>]. Despite the alarming global burden of TB infection (TBI), with an estimated 1.7 billion individuals infected with <italic toggle="yes">Mycobacterium tuberculosis</italic> [<xref rid="C2" ref-type="bibr">2</xref>], of whom only a small percentage (5% to 15%) will develop active TB disease during their lifetime [<xref rid="C3" ref-type="bibr">3</xref>, <xref rid="C4" ref-type="bibr">4</xref>]. Most people with TBI remain asymptomatic and do not progress to active disease – a condition previously referred to as “latent TB infection” [<xref rid="C4" ref-type="bibr">4</xref>]. However, for those who do progress to active TB, the consequences can be severe, leading to illness, disability and even death if not promptly diagnosed and treated. In recognition of the persistent threat posed by TB, the World Health Organization (WHO) launched the End TB Strategy in 2015, which aims to reduce TB incidence and mortality [<xref rid="C2" ref-type="bibr">2</xref>, <xref rid="C5" ref-type="bibr">5</xref>, <xref rid="C6" ref-type="bibr">6</xref>]. Additionally, WHO guidelines on TB preventive treatment (TPT) emphasise identifying and treating people who have TBI at highest risk of disease progression, such as individuals living with HIV, close contacts of TB cases and other vulnerable groups, given that the safety profile of current TPT regimens does not support indiscriminate treatment of all infected individuals [<xref rid="C7" ref-type="bibr">7</xref>]. In this context, biomarker-based tests that reliably predict progression could play a pivotal role in refining risk stratification and targeting preventive therapy to those most likely to benefit, thereby optimising both clinical outcomes and programmatic resources.</p><p>Accurate diagnosis of TBI is critical for TB prevention. However, currently available tests, tuberculin skin tests (TSTs) and interferon-gamma release assays (IGRAs), exhibit a remarkably low positive predictive value (PPV) (below 10%) for future disease development [<xref rid="C8" ref-type="bibr">8</xref>]. This limitation, coupled with the potential toxicity of preventive treatment regimens [<xref rid="C9" ref-type="bibr">9</xref>], restricts programmatic management of TBI to less than 2% of the overall reservoir [<xref rid="C10" ref-type="bibr">10</xref>]. To address this gap, research has increasingly focused on identifying biomarkers that can distinguish individuals with TBI who are likely to progress to active TB, in the absence of clinical signs/symptoms or microbiological evidence of disease. Biomarker-based tests hold promise for improved risk stratification and more targeted delivery of TPT [<xref rid="C11" ref-type="bibr">11</xref>–<xref rid="C13" ref-type="bibr">13</xref>]. In 2016, Z<sc>ak</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C14" ref-type="bibr">14</xref>] described the first prototype of a biomarker-based test for TB progression, namely a whole blood RNA signature including 16 genes. The authors demonstrated that this signature predicted TB progression with 66.1% sensitivity and 80.6% specificity [<xref rid="C14" ref-type="bibr">14</xref>], a performance very close to the characteristics of a WHO-recommended target product profile for a test of progression [<xref rid="C15" ref-type="bibr">15</xref>]. Despite nearly a decade having passed since the publication of this landmark study [<xref rid="C14" ref-type="bibr">14</xref>], none of the numerous candidate tests developed since then have reached the market. The inadequacies of existing immunological tests, combined with the absence of commercially available biomarker-based tools, underscores a critical gap in this field.</p><p>In recent years, several reviews have explored the diagnostic and predictive capabilities of various biomarker-based tests for TB. However, most have focused on detecting active TB rather than predicting progression from TBI to active disease [<xref rid="C16" ref-type="bibr">16</xref>]. While a few studies have attempted to synthesise evidence on the prediction of progression to active TB, the available literature remains limited. For instance, in 2020, M<sc>ulenga</sc>
<italic toggle="yes">et al.</italic> [<xref rid="C17" ref-type="bibr">17</xref>] conducted a systematic review comparing the performance of host blood transcriptomic signatures for diagnosing and predicting progression to TB disease in HIV-negative adults and adolescents. Their analysis included 20 studies evaluating 25 signatures for diagnosis or prediction of progression. However, the authors concluded that the evidence was insufficient to determine if host blood mRNA signatures could be used as standalone diagnostic or predictive tests [<xref rid="C17" ref-type="bibr">17</xref>]. Since the publication of this review over 5 years ago, additional evidence has emerged regarding biomarker-based tests for predicting TB progression.</p><p>We conducted a systematic review of the literature published since 2016 to 1) identify and describe biomarker-based tests with potential for market entry and 2) evaluate the performance and limitations of the identified tests in predicting progression from TBI to active TB.</p></sec><sec sec-type="methods" id="s2"><title>Methods</title><sec id="s2a"><title>Protocol and registration</title><p>The protocol for this systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO ID CRD42024554184) and adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses of Diagnostic Test Accuracy Studies (PRISMA-DTA) guidelines [<xref rid="C18" ref-type="bibr">18</xref>]. A completed PRISMA checklist is provided in <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://publications.ersnet.org/lookup/doi/10.1183/16000617.0151-2025#supplementary" ext-link-type="uri">appendix 1</ext-link>.</p></sec><sec id="s2b"><title>Search strategy</title><p>An experienced information specialist (B. Skidmore) developed the search strategy in collaboration with the research team. The Medline strategy was peer-reviewed by a senior information specialist (K. Campbell) using the PRESS Checklist [<xref rid="C19" ref-type="bibr">19</xref>]. Searches were conducted across multiple databases, including Ovid Medline® ALL, Embase Classic+Embase, Cochrane Central Register of Controlled Trials and Web of Science (core collection). Search terms included controlled vocabulary such as “latent tuberculosis”, “immunologic tests” and “predictive value of tests”, alongside relevant free text such as “LTBI”, “biomarker” and “forecast.” Animal-only records and case studies were excluded. There were no language restrictions. Searches were restricted to publications from 1 January 2016 to 6 April 2024. We applied this date restriction because the first study unveiling a prototype test for predicting TB progression dates to approximately 9 years ago [<xref rid="C14" ref-type="bibr">14</xref>]. EndNote (version 9.3.3, Clarivate Analytics) was used for deduplication and records were managed and screened using Covidence (Veritas Health Innovation Ltd.). The full search strategy is provided in <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://publications.ersnet.org/lookup/doi/10.1183/16000617.0151-2025#supplementary" ext-link-type="uri">appendix 2</ext-link>.</p><p>In addition to bibliographic database searches, we hand-searched reference lists of included studies and reviewed grey literature by searching clinical trial registries (<italic toggle="yes">e.g.</italic> ClinicalTrials.gov) and abstracts from various relevant conferences (<italic toggle="yes">e.g.</italic> the World Union Conference) from 2020 to 2024. We also explored selected Grey Matters [<xref rid="C20" ref-type="bibr">20</xref>] sources to ensure broad coverage of relevant literature.</p></sec><sec id="s2c"><title>Eligibility criteria</title><p>The review included experimental and observational studies reporting on the features, performance and challenges of biomarker-based tests aimed at evaluating the risk of progression from TBI to active TB. Progression from TBI to active TB was defined as the transition from an asymptomatic infection, where the bacteria are present but inactive in the body and undetectable through conventional microbiological tests, to an active disease state characterised by clinical symptoms, bacterial replication, positive microbiological tests and the potential to transmit the infection to others. This progression is often confirmed through clinical, microbiological or radiological evidence of active TB, such as positive sputum culture, chest radiograph findings consistent with TB or other diagnostic tests demonstrating active disease (<italic toggle="yes">e.g.</italic> Xpert MTB/Rif or Ultra).</p><p>Eligible study designs were randomised and nonrandomised trials, cohort, case–control, and cross-sectional studies. Reviews, commentaries and perspective pieces discussing or reporting on the challenges encountered in the development, evaluation or future implementation of one or more eligible tests were retained to gather further contextual information but were not included in the results of this review. Protocols, economic analyses, modelling studies and case reports were excluded.</p><p>The inclusion and exclusion criteria according to the PICO (population, intervention, comparison, outcome) framework were as follows:
<list list-type="bullet"><list-item><p>Population: no restrictions were applied based on population demographics.</p></list-item><list-item><p>Intervention: biomarker-based or immunological tests aimed at predicting progression to active TB disease.</p></list-item><list-item><p>Comparison: no exclusion criteria were applied regarding comparators, recognising the absence of standard reference tests for predicting progression to active TB disease.</p></list-item><list-item><p>Outcomes: the primary outcomes included diagnostic accuracy measures, such as sensitivity, specificity, PPV, negative predictive value (NPV) and area-under-the-curve (AUC), for biomarker-based tests predicting progression to active TB.</p></list-item></list></p></sec><sec id="s2d"><title>Screening and selection process</title><p>Records retrieved from database searches were imported into Covidence, while grey literature was catalogued separately. Screening involved an initial review of titles and abstracts, followed by full-text evaluation conducted independently by two reviewers among K. Tingley, A. Li, N. Maredia and R. Ravikumar. To expedite the title and abstract screening process, a liberal accelerated approach [<xref rid="C21" ref-type="bibr">21</xref>] was employed, whereby any citation marked as “include” or “unsure” by either reviewer advanced to full-text screening. Pilot testing for title and abstract screening was conducted using 25 randomly selected titles and abstracts to ensure high inter-rater reliability. Similarly, full-text screening was piloted on 10 randomly selected full-text articles to maintain reliability. Discrepancies between reviewers were resolved through discussion or, if needed, adjudication by a senior team member. All full-text exclusions were systematically documented and justified to ensure reproducibility.</p></sec><sec id="s2e"><title>Assessment of methodological quality</title><p>The methodological quality of included studies was evaluated using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool [<xref rid="C22" ref-type="bibr">22</xref>]. One reviewer performed the quality assessment, and a second reviewer verified the results. We conducted more than one QUADAS-2 evaluation if different reference standards or distinct patient selection procedures were used for the same test within the same study. Assessments were summarised graphically by domain (patient selection, index test, reference standard, flow and timing). Any disagreements were resolved through discussion or by a third reviewer.</p></sec><sec id="s2f"><title>Data extraction</title><p>A single reviewer among K. Tingley, A. Li, N. Maredia and R. Ravikumar extracted data for each study using a customised form built in Airtable (Formagrid Inc.), an online collaborative spreadsheeting platform and tailored to the specific objectives of the review. A second reviewer verified all extracted data. The form was pilot tested on four randomly selected studies ensuring its consistency and reliability. Extracted data included publication details, study setting, design, sample size and participant characteristics, alongside details on diagnostic criteria, test types and performance measures. Contextual information regarding test development and implementation challenges was also gathered. <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://publications.ersnet.org/lookup/doi/10.1183/16000617.0151-2025#supplementary" ext-link-type="uri">Appendix 3</ext-link> includes the data extraction form.</p></sec><sec id="s2g"><title>Data synthesis</title><p>Although a meta-analysis was initially planned, the studies included in this review demonstrated substantial clinical and methodological heterogeneity, as well as incomplete and nontransparent reporting, which precluded data pooling. We therefore conducted a descriptive and narrative synthesis to summarise the key findings.</p><p>For the descriptive synthesis, data were extracted on the main study characteristics and diagnostic performance measures as described above. Studies were grouped by biomarker test type, with findings on sensitivity, specificity, PPV, NPV and AUC where available. Due to heterogeneity in study design, patient populations, and methodologies, direct comparisons were not feasible; however, we organised data into tables to illustrate the performance range of each test.</p><p>Given the variability and limited transparency in reporting, we did not evaluate the biomarker tests against the ASSURED criteria [<xref rid="C23" ref-type="bibr">23</xref>] or WHO target product profiles [<xref rid="C24" ref-type="bibr">24</xref>]. Nonetheless, the narrative synthesis provides insights into the strengths and limitations of identified tests and highlights key evidence gaps and areas for further research.</p></sec></sec><sec sec-type="results" id="s3"><title>Results</title><sec id="s3a"><title>Study selection</title><p>A total of 5582 studies were identified from electronic database searches, with an additional 28 studies retrieved from grey literature sources (<xref rid="F1" ref-type="fig">figure 1</xref>). After removing duplicates, 5517 articles remained for title and abstract screening. Of these, 5339 studies were determined to be irrelevant to our review questions, leaving 178 studies for full-text review. During the full-text screening, 159 studies were excluded for specific reasons: 48 lacked relevant outcome data, 76 focused on the wrong stage of the TB spectrum (<italic toggle="yes">e.g.</italic> diagnosis of TBI) or on discrimination of TB stages rather than progression to active TB and 25 were excluded due to ineligible study design (<xref rid="F1" ref-type="fig">figure 1</xref>). Ultimately, 19 studies met all eligibility criteria and were included in the review. Of these, two studies [<xref rid="C25" ref-type="bibr">25</xref>, <xref rid="C26" ref-type="bibr">26</xref>] were excluded from the quantitative data synthesis due to incomplete reporting of key test characteristics. A complete list of excluded studies with reasons for exclusion is provided in <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://publications.ersnet.org/lookup/doi/10.1183/16000617.0151-2025#supplementary" ext-link-type="uri">appendix 4</ext-link>.</p><fig position="float" id="F1" orientation="portrait"><label>FIGURE 1</label><caption><p>Study selection. TB: tuberculosis.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="ERR-0151-2025.01.jpg"/></fig></sec><sec id="s3b"><title>Study characteristics</title><p>Our review included 19 studies, the majority of which (16/19, 85%) were published after 2019 (<xref rid="TB1" ref-type="table">table 1</xref>). Nearly half of the corresponding authors were based in the USA (9/19, 47%), followed by South Africa (2/19, 11%), China (2/19, 11%), the United Kingdom (2/19, 11%) and a handful of other countries (<xref rid="TB1" ref-type="table">table 1</xref>). Most studies (16/19, 84%) were supported by nonindustry funding sources. Reported conflicts of interest were relatively evenly split, with 42% (8/19) of studies disclosing potential conflicts (<xref rid="TB1" ref-type="table">table 1</xref>).</p><table-wrap position="float" id="TB1" orientation="portrait"><label>TABLE 1</label><caption><p>Descriptive characteristics of included studies (n=19)</p></caption><table frame="hsides" rules="groups"><colgroup span="1"><col align="left" span="1"/><col align="center" span="1"/><col align="center" span="1"/></colgroup><thead><tr><th align="left" rowspan="2" colspan="1">Study characteristic</th><th align="center" colspan="2" rowspan="1">Frequency</th></tr><tr><th align="center" rowspan="1" colspan="1">n</th><th align="center" rowspan="1" colspan="1">%</th></tr></thead><tbody><tr><td align="left" colspan="3" rowspan="1">
<bold>Year of publication</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> 2016–2018</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr><tr><td align="left" rowspan="1" colspan="1"> 2019–2021</td><td align="center" rowspan="1" colspan="1">10</td><td align="center" rowspan="1" colspan="1">53</td></tr><tr><td align="left" rowspan="1" colspan="1"> 2022–2024</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">32</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Country of corresponding author</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> China</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" rowspan="1" colspan="1"> Germany</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> India</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> Korea</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> Mexico</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> South Africa</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" rowspan="1" colspan="1"> UK</td><td align="center" rowspan="1" colspan="1">9</td><td align="center" rowspan="1" colspan="1">47</td></tr><tr><td align="left" rowspan="1" colspan="1"> USA</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Study sponsor<sup>#</sup></bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Industry</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> Nonindustry</td><td align="center" rowspan="1" colspan="1">16</td><td align="center" rowspan="1" colspan="1">84</td></tr><tr><td align="left" rowspan="1" colspan="1"> Not reported</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Conflicts of interest</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Authors report conflicts</td><td align="center" rowspan="1" colspan="1">8</td><td align="center" rowspan="1" colspan="1">42</td></tr><tr><td align="left" rowspan="1" colspan="1"> Authors report no conflicts</td><td align="center" rowspan="1" colspan="1">11</td><td align="center" rowspan="1" colspan="1">58</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Study design</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Case–control study</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr><tr><td align="left" rowspan="1" colspan="1"> Cohort study</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">68</td></tr><tr><td align="left" rowspan="1" colspan="1"> Cross-sectional study</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> Other (<italic toggle="yes">e.g.</italic> nested case–control study, quasi-experimental study)</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Data/sample collection</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Prospective</td><td align="center" rowspan="1" colspan="1">12</td><td align="center" rowspan="1" colspan="1">63</td></tr><tr><td align="left" rowspan="1" colspan="1"> Retrospective</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">32</td></tr><tr><td align="left" rowspan="1" colspan="1"> Not reported</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Sampling method</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Consecutive</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr><tr><td align="left" rowspan="1" colspan="1"> Convenient</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> Random</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> Not reported</td><td align="center" rowspan="1" colspan="1">14</td><td align="center" rowspan="1" colspan="1">74</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Sample size</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> 50–99</td><td align="center" rowspan="1" colspan="1">4</td><td align="center" rowspan="1" colspan="1">21</td></tr><tr><td align="left" rowspan="1" colspan="1"> 100–499</td><td align="center" rowspan="1" colspan="1">5</td><td align="center" rowspan="1" colspan="1">26</td></tr><tr><td align="left" rowspan="1" colspan="1"> 500–999</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" rowspan="1" colspan="1"> ≥1000</td><td align="center" rowspan="1" colspan="1">7</td><td align="center" rowspan="1" colspan="1">37</td></tr><tr><td align="left" rowspan="1" colspan="1"> Not reported</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Country for participant recruitment<sup>#</sup></bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Single country</td><td align="center" rowspan="1" colspan="1">11</td><td align="center" rowspan="1" colspan="1">58</td></tr><tr><td align="left" rowspan="1" colspan="1"> Multi-country</td><td align="center" rowspan="1" colspan="1">8</td><td align="center" rowspan="1" colspan="1">42</td></tr><tr><td align="left" rowspan="1" colspan="1"> Brazil</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr><tr><td align="left" rowspan="1" colspan="1"> Haiti</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> India</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> UK</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr><tr><td align="left" rowspan="1" colspan="1"> China</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr><tr><td align="left" rowspan="1" colspan="1"> South Africa</td><td align="center" rowspan="1" colspan="1">10</td><td align="center" rowspan="1" colspan="1">53</td></tr><tr><td align="left" rowspan="1" colspan="1"> Gambia</td><td align="center" rowspan="1" colspan="1">5</td><td align="center" rowspan="1" colspan="1">26</td></tr><tr><td align="left" rowspan="1" colspan="1"> Ethiopia</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr><tr><td align="left" rowspan="1" colspan="1"> Moldova</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> Uganda</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" rowspan="1" colspan="1"> USA</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" rowspan="1" colspan="1"> Venezuela</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Study population<sup>#</sup></bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Confirmed TB</td><td align="center" rowspan="1" colspan="1">8</td><td align="center" rowspan="1" colspan="1">42</td></tr><tr><td align="left" rowspan="1" colspan="1"> Healthy controls</td><td align="center" rowspan="1" colspan="1">8</td><td align="center" rowspan="1" colspan="1">42</td></tr><tr><td align="left" rowspan="1" colspan="1"> LTBI</td><td align="center" rowspan="1" colspan="1">12</td><td align="center" rowspan="1" colspan="1">63</td></tr><tr><td align="left" rowspan="1" colspan="1"> Household contacts of TB cases</td><td align="center" rowspan="1" colspan="1">7</td><td align="center" rowspan="1" colspan="1">37</td></tr><tr><td align="left" rowspan="1" colspan="1"> Other special population (<italic toggle="yes">e.g.</italic> HIV co-infection)</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" rowspan="1" colspan="1"> Not reported</td><td align="center" rowspan="1" colspan="1">2</td><td align="center" rowspan="1" colspan="1">11</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Number of biomarker tests per study</bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> 1</td><td align="center" rowspan="1" colspan="1">7</td><td align="center" rowspan="1" colspan="1">37</td></tr><tr><td align="left" rowspan="1" colspan="1"> 2</td><td align="center" rowspan="1" colspan="1">5</td><td align="center" rowspan="1" colspan="1">26</td></tr><tr><td align="left" rowspan="1" colspan="1"> 3</td><td align="center" rowspan="1" colspan="1">1</td><td align="center" rowspan="1" colspan="1">5</td></tr><tr><td align="left" rowspan="1" colspan="1"> 4+</td><td align="center" rowspan="1" colspan="1">6</td><td align="center" rowspan="1" colspan="1">32</td></tr><tr><td align="left" colspan="3" rowspan="1">
<bold>Types of test included<sup>#</sup></bold>
</td></tr><tr><td align="left" rowspan="1" colspan="1"> Host: RNA (<italic toggle="yes">e.g.</italic> transcriptomic signatures)</td><td align="center" rowspan="1" colspan="1">11</td><td align="center" rowspan="1" colspan="1">58</td></tr><tr><td align="left" rowspan="1" colspan="1"> Host: proteins</td><td align="center" rowspan="1" colspan="1">5</td><td align="center" rowspan="1" colspan="1">26</td></tr><tr><td align="left" rowspan="1" colspan="1"> Pathogen: cytokines/chemokines</td><td align="center" rowspan="1" colspan="1">3</td><td align="center" rowspan="1" colspan="1">16</td></tr></tbody></table><table-wrap-foot><p>LTBI: latent tuberculosis infection; TB: tuberculosis. <sup>#</sup>: Categories are not mutually exclusive.</p></table-wrap-foot></table-wrap><p>Cohort studies were the predominant study design (13/19, 68%). Sample collection was prospective in most studies (12/19, 63%), though sampling methods were frequently not reported, with only 26% specifying methods like consecutive or random sampling (<xref rid="TB1" ref-type="table">table 1</xref>). Study sample sizes varied substantially, with 7/19 (37%) including over 1000 participants. Studies were conducted in both single-country (11/19, 58%) and multi-country settings (8/19, 42%), with common recruitment sites including South Africa (10/19, 53%), Gambia (5/19, 26%) and the UK (3/19, 16%) (<xref rid="TB1" ref-type="table">table 1</xref>). Study populations typically included confirmed TB cases (8/19, 42%) and individuals with TBI (12/19, 63%) (<xref rid="TB1" ref-type="table">table 1</xref>). Studies often reported on more than one unique biomarker-based test (12/19, 63%) (<xref rid="TB1" ref-type="table">table 1</xref>).</p></sec><sec id="s3c"><title>Quality of included studies</title><p>We assessed the quality of 17 out of 19 eligible studies. The two studies excluded from data synthesis due to incomplete reporting on key test characteristics were also excluded from quality assessment. For two of the 17 assessed studies, we conducted more than one QUADAS-2 evaluation, as each study used different reference standards or had distinct patient selection procedures within its design warranting separate quality assessments. A total of 19 quality assessments were completed.</p><p>The QUADAS-2 assessment identified notable concerns regarding both bias and applicability, largely due to gaps in transparency and reporting. The risk of bias was most frequently rated as “unclear” across the domains of patient selection, index test, and flow and timing, affecting 13 out of 19 assessments for patient selection and nine for index test blinding (<xref rid="F2" ref-type="fig">figure 2</xref>). This lack of detail, especially regarding recruitment methods, blinding procedures and participant follow-up, limited our ability to fully assess potential sources of bias in many studies. Only a minority of studies were rated as having a “low” risk of bias across all domains, indicating that overall methodological transparency was insufficient in much of the sample (<xref rid="F2" ref-type="fig">figure 2</xref>).</p><fig position="float" id="F2" orientation="portrait"><label>FIGURE 2</label><caption><p>Risk of bias assessment using QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies 2).</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="ERR-0151-2025.02.jpg"/></fig><p>In terms of applicability, the risk was generally low, with most studies using appropriate patient populations, index tests and reference standards for TB diagnostic and prognostic evaluation (<xref rid="F2" ref-type="fig">figure 2</xref>). However, a few studies received “unclear” ratings, particularly concerning reference standards, due to either variations in protocols or limited information on the test's clinical applicability (<xref rid="F2" ref-type="fig">figure 2</xref>). These “unclear” ratings underscore a common challenge across studies included in the systematic review, namely insufficient detail that hinders a robust assessment of study rigor and real-world relevance. The full QUADAS-2 assessment for each study is provided in <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://publications.ersnet.org/lookup/doi/10.1183/16000617.0151-2025#supplementary" ext-link-type="uri">appendix 5</ext-link>.</p></sec><sec id="s3d"><title>Performance of biomarker-based tests</title><p>Our review identified 64 unique biomarker-based tests for predicting progression to active TB, including 23 gene signatures (<xref rid="TB2" ref-type="table">table 2</xref>), 14 protein signatures (<xref rid="TB3" ref-type="table">table 3</xref>) and 33 cytokines/chemokines (<xref rid="TB4" ref-type="table">table 4</xref>).</p><table-wrap position="float" id="TB2" orientation="portrait"><label>TABLE 2</label><caption><p>Summary of diagnostic performance of host RNA biomarkers for predicting progression to active tuberculosis (ATB) (n=10 studies)</p></caption><table frame="hsides" rules="groups"><colgroup span="1"><col align="left" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/></colgroup><thead><tr><th align="left" rowspan="1" colspan="1">First author [ref.], year</th><th align="center" rowspan="1" colspan="1">Index test name</th><th align="center" rowspan="1" colspan="1">Reference standard used</th><th align="center" rowspan="1" colspan="1">Group 1</th><th align="center" rowspan="1" colspan="1">Group 2</th><th align="center" rowspan="1" colspan="1">Timing</th><th align="center" rowspan="1" colspan="1">Sensitivity % (95% CI)</th><th align="center" rowspan="1" colspan="1">Specificity % (95% CI)</th><th align="center" rowspan="1" colspan="1">PPV % (95% CI)</th><th align="center" rowspan="1" colspan="1">NPV % (95% CI)</th><th align="center" rowspan="1" colspan="1">AUC (95% CI)</th></tr></thead><tbody><tr><td rowspan="12" colspan="1">
<bold>V<sc>argas [<xref rid="C27" ref-type="bibr">27</xref>] 2023</sc></bold>
</td><td align="center" rowspan="6" colspan="1">Gene signature (comprising 42 genes)<sup>#</sup></td><td align="center" rowspan="6" colspan="1">NR</td><td align="center" rowspan="6" colspan="1">Progressors (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="6" colspan="1">Nonprogressors (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1"><3 months to disease</td><td align="center" rowspan="1" colspan="1">87.3 (81.5–93.1)</td><td align="center" rowspan="1" colspan="1">91.7 (89.9–93.5)</td><td align="center" rowspan="1" colspan="1">17.7 (14.5–20.9)</td><td align="center" rowspan="1" colspan="1">99.7 (99.6–99.8)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><12 months to disease</td><td align="center" rowspan="1" colspan="1">78.6 (73.4–83.9)</td><td align="center" rowspan="1" colspan="1">80.3 (77.7–83.0)</td><td align="center" rowspan="1" colspan="1">7.5 (6.6–8.5)</td><td align="center" rowspan="1" colspan="1">99.5 (99.4–99.5)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><18 months to disease</td><td align="center" rowspan="1" colspan="1">77.3 (72.3–82.3)</td><td align="center" rowspan="1" colspan="1">78.3 (75.6–81.0)</td><td align="center" rowspan="1" colspan="1">6.8 (6.0–7.6)</td><td align="center" rowspan="1" colspan="1">99.4 (99.3–99.5)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><24 months to disease</td><td align="center" rowspan="1" colspan="1">74.7 (69.7–79.6)</td><td align="center" rowspan="1" colspan="1">78.3 (75.6–81.0)</td><td align="center" rowspan="1" colspan="1">6.6 (5.8–7.3)</td><td align="center" rowspan="1" colspan="1">99.3 (99.3–99.4)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><6 months to disease</td><td align="center" rowspan="1" colspan="1">82.1 (76.4–87.8)</td><td align="center" rowspan="1" colspan="1">83.8 (81.3–86.2)</td><td align="center" rowspan="1" colspan="1">9.3 (8.1–10.6)</td><td align="center" rowspan="1" colspan="1">99.6 (99.5–99.7)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><30 months to disease</td><td align="center" rowspan="1" colspan="1">74.2 (69.3–79.2)</td><td align="center" rowspan="1" colspan="1">78.3 (75.6–81.0)</td><td align="center" rowspan="1" colspan="1">6.5 (5.8–7.3)</td><td align="center" rowspan="1" colspan="1">99.3 (99.3–99.4)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="6" colspan="1">Gene signature (comprising 18 genes)<sup>¶</sup></td><td align="center" rowspan="6" colspan="1">NR</td><td align="center" rowspan="6" colspan="1">Progressors (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="6" colspan="1">Progressors (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1"><3 months to disease</td><td align="center" rowspan="1" colspan="1">83.3 (76.8–89.8)</td><td align="center" rowspan="1" colspan="1">91.6 (89.8–93.4)</td><td align="center" rowspan="1" colspan="1">16.8 (13.8–19.9)</td><td align="center" rowspan="1" colspan="1">99.6 (99.5–99.7)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><6 months to disease</td><td align="center" rowspan="1" colspan="1">74.0 (67.5–80.5)</td><td align="center" rowspan="1" colspan="1">86.9 (84.7–89.2)</td><td align="center" rowspan="1" colspan="1">10.4 (8.8–11.9)</td><td align="center" rowspan="1" colspan="1">99.4 (99.3–99.5)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><12 months to disease</td><td align="center" rowspan="1" colspan="1">74.8 (69.2–80.4)</td><td align="center" rowspan="1" colspan="1">81.0 (78.4–83.6)</td><td align="center" rowspan="1" colspan="1">7.4 (6.5–8.4)</td><td align="center" rowspan="1" colspan="1">99.4 (99.3–99.5)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><18 months to disease</td><td align="center" rowspan="1" colspan="1">78.4 (73.5–83.4)</td><td align="center" rowspan="1" colspan="1">74.9 (72.0–77.8)</td><td align="center" rowspan="1" colspan="1">6.0 (5.3–6.6)</td><td align="center" rowspan="1" colspan="1">99.4 (99.3–99.5)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><24 months to disease</td><td align="center" rowspan="1" colspan="1">77.1 (72.2–81.9)</td><td align="center" rowspan="1" colspan="1">74.9 (72.0–77.8)</td><td align="center" rowspan="1" colspan="1">5.9 (5.3–6.5)</td><td align="center" rowspan="1" colspan="1">99.4 (99.3–99.5)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1"><30 months to disease</td><td align="center" rowspan="1" colspan="1">77.3 (72.5–82.0)</td><td align="center" rowspan="1" colspan="1">74.9 (72.0–77.8)</td><td align="center" rowspan="1" colspan="1">5.9 (5.3–6.5)</td><td align="center" rowspan="1" colspan="1">99.4 (99.3–99.5)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="6" colspan="1">
<bold>X<sc>in [<xref rid="C28" ref-type="bibr">28</xref>] 2022</sc></bold>
</td><td align="center" rowspan="1" colspan="1">hsa-miR-16-5p</td><td align="center" rowspan="6" colspan="1">NR</td><td align="center" rowspan="6" colspan="1">ATB (n=73)<break/> More than 50% ≥60 years old<break/> Predominately male</td><td align="center" rowspan="6" colspan="1">LTBI (n=77)<break/> More than 50% ≥60 years old<break/> Predominately male<break/> Higher BMI than ATB group</td><td align="center" rowspan="6" colspan="1">Followed-up to 5 years</td><td align="center" rowspan="1" colspan="1">68.49 (57.14–78.00)</td><td align="center" rowspan="1" colspan="1">58.44 (47.29–68.79)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">hsa-miR-16-5p+co-variables</td><td align="center" rowspan="1" colspan="1">73.97 (62.89–82.66)</td><td align="center" rowspan="1" colspan="1">90.91 (82.40–95.53)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">hsa-miR-16-5p+hsa-miR-451a</td><td align="center" rowspan="1" colspan="1">38.36 (28.05–49.83)</td><td align="center" rowspan="1" colspan="1">90.91 (82.40–95.53)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">hsa-miR-451a</td><td align="center" rowspan="1" colspan="1">38.36 (28.05–49.83)</td><td align="center" rowspan="1" colspan="1">90.91 (82.40–95.53)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">hsa-miR-451a+co-variables</td><td align="center" rowspan="1" colspan="1">80.82 (70.34–88.22)</td><td align="center" rowspan="1" colspan="1">79.22 (68.88–86.78)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">hsa-miR-451a+hsa-miR-16-5p+co-variables</td><td align="center" rowspan="1" colspan="1">76.71 (65.83–84.92)</td><td align="center" rowspan="1" colspan="1">83.12 (73.23–89.86)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="5" colspan="1">
<bold>L<sc>ong [<xref rid="C29" ref-type="bibr">29</xref>] 2021</sc></bold>
</td><td align="center" rowspan="5" colspan="1">Gene signature (comprising 10 genes)<sup>+</sup></td><td align="center" rowspan="5" colspan="1">NR</td><td align="center" rowspan="2" colspan="1">Nonprogressor (within 0–1 year) (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="2" colspan="1">Progressor (within 0–1 year) (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1">At exposure</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.66</td></tr><tr><td rowspan="1" colspan="1">6 months after exposure</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.72</td></tr><tr><td rowspan="3" colspan="1">Nonprogressor (within 1–2 years) (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="3" colspan="1">Progressor (within 1–2 years) (n=NR)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1">At exposure</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.48</td></tr><tr><td rowspan="1" colspan="1">6 months after exposure</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.52</td></tr><tr><td rowspan="1" colspan="1">18 months after exposure</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.81</td></tr><tr><td rowspan="4" colspan="1">
<bold><sc>de</sc> A<sc>raujo [<xref rid="C35" ref-type="bibr">35</xref>] 2021</sc></bold>
</td><td align="center" rowspan="4" colspan="1">NPC2</td><td align="center" rowspan="4" colspan="1">Smear microscopy, culture (nonspecified) and/or Xpert</td><td align="center" rowspan="4" colspan="1">HHC (nonprogressors) (n=208)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1">ATB (progressors) (n=13)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1">Time to TB ≤3 months</td><td align="center" rowspan="1" colspan="1">92.3 (64–99.8)</td><td align="center" rowspan="1" colspan="1">75 (68.5–80.7)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">ATB (progressors) (n=34)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1">Time to TB 4–6 months</td><td align="center" rowspan="1" colspan="1">50 (32.4–67.6)</td><td align="center" rowspan="1" colspan="1">75 (68.5–80.7)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">ATB (progressors) (n=19)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1">Time to TB 7–12 months</td><td align="center" rowspan="1" colspan="1">63.2 (38.4–83.7)</td><td align="center" rowspan="1" colspan="1">75 (68.5–80.7)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">ATB (progressors) (n=32)<break/> No group characteristics reported</td><td align="center" rowspan="1" colspan="1">Time to TB 13–18 months</td><td align="center" rowspan="1" colspan="1">34.4 (18.6–53.2)</td><td align="center" rowspan="1" colspan="1">75 (68.5–80.7)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="4" colspan="1">
<bold>R<sc>oe [<xref rid="C30" ref-type="bibr">30</xref>] 2020</sc></bold>
</td><td align="center" rowspan="4" colspan="1">Three-gene SVM model (BATF2, GBP5, SCARF1)
BATF2</td><td align="center" rowspan="4" colspan="1">Culture (nonspecified), chest radiography</td><td align="center" rowspan="2" colspan="1">Progressors (n=12)</td><td align="center" rowspan="4" colspan="1">Nonprogressors (n=48)</td><td align="center" rowspan="1" colspan="1">Time to TB within 90 days</td><td align="center" rowspan="1" colspan="1">0.83 (0.52–0.98)</td><td align="center" rowspan="1" colspan="1">0.96 (0.83–0.99)</td><td align="center" rowspan="1" colspan="1">23 (NR)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.96 (0.92–1)</td></tr><tr><td align="center" rowspan="1" colspan="1">Time to TB within 90 days</td><td align="center" rowspan="1" colspan="1">0.83 (0.52–0.98)</td><td align="center" rowspan="1" colspan="1">0.92 (0.83–0.99)</td><td align="center" rowspan="1" colspan="1">13</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.93 (0.86–1)</td></tr><tr><td rowspan="1" colspan="1">Progressors (n=29)</td><td align="center" rowspan="1" colspan="1">Time to TB 91–360 days</td><td align="center" rowspan="1" colspan="1">0.52 (0.36–0.74)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">Progressors (n=25)</td><td align="center" rowspan="1" colspan="1">Time to TB >360 days</td><td align="center" rowspan="1" colspan="1">0.24 (0.12–0.49)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="4" colspan="1">
<bold>P<sc>enn-</sc>N<sc>icholson [<xref rid="C31" ref-type="bibr">31</xref>] 2020</sc></bold>
</td><td align="center" rowspan="4" colspan="1">RISK6</td><td align="center" rowspan="4" colspan="1">NR</td><td align="center" rowspan="2" colspan="1">Progressor (ACS cohort) (n=46)</td><td align="center" rowspan="2" colspan="1">Nonprogressors (ACS cohort) (n=284)</td><td align="center" rowspan="1" colspan="1">0–12 months before TB diagnosis</td><td align="center" rowspan="1" colspan="1">82.8 (NR)</td><td align="center" rowspan="1" colspan="1">82.8 (NR)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">87.6% (82.8–92.4)</td></tr><tr><td rowspan="1" colspan="1">12–24 months before TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">74.0% (66.0–82.0)</td></tr><tr><td rowspan="2" colspan="1">Progressors (GC6-74 cohort) (n=50)</td><td align="center" rowspan="2" colspan="1">Non-progressors (GC6-74 cohort) (n=372)</td><td align="center" rowspan="1" colspan="1">0–12 months before TB diagnosis</td><td align="center" rowspan="1" colspan="1">50.0 (NR)</td><td align="center" rowspan="1" colspan="1">56.0 (NR)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">70.6% (61.6–79.5)</td></tr><tr><td rowspan="1" colspan="1">12–24 months before TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">67.6% (58.2–76.9)</td></tr><tr><td rowspan="7" colspan="1">
<bold>L<sc>eong [<xref rid="C32" ref-type="bibr">32</xref>] 2020</sc></bold>
</td><td align="center" rowspan="1" colspan="1">ACS-CoR</td><td align="center" rowspan="7" colspan="1">Clinical signs, culture (nonspecified), chest radiography</td><td align="center" rowspan="7" colspan="1">Progressors (n=11)<break/> Nine of the 11 progressors were culture-proven TB<break/>Predominately female (11/16)<break/>Median age: 22</td><td align="center" rowspan="7" colspan="1">Nonprogressors (n=21)<break/> Age- and sex- matched</td><td align="center" rowspan="7" colspan="1">Up to 5 years: five were diagnosed within 2 years (early progressors) and six were diagnosed after 2 years (late progressors) of sample collection</td><td align="center" rowspan="1" colspan="1">51.5 (46.0–57.1)</td><td align="center" rowspan="1" colspan="1">77.4 (72.8–82.0)</td><td align="center" rowspan="1" colspan="1">10.6 (5.7–16.5)</td><td align="center" rowspan="1" colspan="1">96.8 (94.6–98.2)</td><td align="center" rowspan="1" colspan="1">0.670 (0.640–0.700)</td></tr><tr><td rowspan="1" colspan="1">Jacobsen3</td><td align="center" rowspan="1" colspan="1">55.3 (48.9–61.6)</td><td align="center" rowspan="1" colspan="1">63.3 (57.0–69.5)</td><td align="center" rowspan="1" colspan="1">7.2 (3.9–11.3)</td><td align="center" rowspan="1" colspan="1">96.4 (93.8–98.0)</td><td align="center" rowspan="1" colspan="1">0.575 (0.546–0.605)</td></tr><tr><td rowspan="1" colspan="1">Kaforou</td><td align="center" rowspan="1" colspan="1">66.4 (60.3–72.5)</td><td align="center" rowspan="1" colspan="1">56.7 (51.3–62.0)</td><td align="center" rowspan="1" colspan="1">7.3 (4.6–11.5)</td><td align="center" rowspan="1" colspan="1">97.0 (94.2–98.5)</td><td align="center" rowspan="1" colspan="1">0.629 (0.602–0.656)</td></tr><tr><td rowspan="1" colspan="1">PREDICT29</td><td align="center" rowspan="1" colspan="1">74.2 (70.4–78.0)</td><td align="center" rowspan="1" colspan="1">84.8 (81.6–88.0)</td><td align="center" rowspan="1" colspan="1">20.2 (13.1–29.4)</td><td align="center" rowspan="1" colspan="1">98.5 (96.9–99.3)</td><td align="center" rowspan="1" colspan="1">0.911 (0.894–0.928)</td></tr><tr><td rowspan="1" colspan="1">RISK4</td><td align="center" rowspan="1" colspan="1">41.3 (33.5–49.1)</td><td align="center" rowspan="1" colspan="1">66.5 (58.8–74.3)</td><td align="center" rowspan="1" colspan="1">6.1 (3.5–11.4)</td><td align="center" rowspan="1" colspan="1">95.6 (93.3–97.6)</td><td align="center" rowspan="1" colspan="1">0.461 (0.434–0.488)</td></tr><tr><td rowspan="1" colspan="1">Sambarey10</td><td align="center" rowspan="1" colspan="1">63.2 (57.1–69.3)</td><td align="center" rowspan="1" colspan="1">58.6 (52.5–64.7)</td><td align="center" rowspan="1" colspan="1">7.3 (4.5–11.6)</td><td align="center" rowspan="1" colspan="1">96.8 (94.4–98.5)</td><td align="center" rowspan="1" colspan="1">0.623 (0.595–0.651)</td></tr><tr><td rowspan="1" colspan="1">Sweeney3</td><td align="center" rowspan="1" colspan="1">42.7 (35.3–50.1)</td><td align="center" rowspan="1" colspan="1">74.6 (67.6–81.6)</td><td align="center" rowspan="1" colspan="1">8.0 (4.2–13.6)</td><td align="center" rowspan="1" colspan="1">96.1 (93.9–97.6)</td><td align="center" rowspan="1" colspan="1">0.590 (0.560–0.620)</td></tr><tr><td rowspan="5" colspan="1">
<bold>W<sc>arsinske [<xref rid="C33" ref-type="bibr">33</xref>] 2018</sc></bold>
</td><td align="center" rowspan="5" colspan="1">Three-gene TB score</td><td align="center" rowspan="5" colspan="1">Smear microscopy, culture (nonspecified)</td><td align="center" rowspan="5" colspan="1">Progressors (ATB) (n=43)<break/> Adolescents: 12–18 years old</td><td align="center" rowspan="5" colspan="1">Nonprogressors (LTBI) (n=101)<break/> Adolescents: 12–18 years old</td><td align="center" rowspan="1" colspan="1">0–7 days</td><td align="center" rowspan="1" colspan="1">89.47 (NR)</td><td align="center" rowspan="1" colspan="1">63.37 (NR)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">99.13<sup>§</sup> (NR)</td><td align="center" rowspan="1" colspan="1">0.86 (0.77–0.96)</td></tr><tr><td rowspan="1" colspan="1">8–180 days</td><td align="center" rowspan="1" colspan="1">86 (NR)</td><td align="center" rowspan="1" colspan="1">84 (NR)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">98.63 (NR)<sup>§</sup></td><td align="center" rowspan="1" colspan="1">0.86 (0.70–1.00)</td></tr><tr><td rowspan="1" colspan="1">181–360 days</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.62 (0.55–0.69)</td></tr><tr><td rowspan="1" colspan="1">361–540 days</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.86 (0.49–0.65)</td></tr><tr><td rowspan="1" colspan="1">541–720 days</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.86 (0.51–0.75)</td></tr><tr><td rowspan="2" colspan="1">
<bold>Z<sc>ak [<xref rid="C14" ref-type="bibr">14</xref>] 2016</sc></bold>
</td><td align="center" rowspan="2" colspan="1">Zak16</td><td align="center" rowspan="2" colspan="1">Culture (nonspecified), smear microscopy, clinical signs</td><td align="center" rowspan="1" colspan="1">Progressors (n=46)</td><td align="center" rowspan="1" colspan="1">Healthy controls (n=107)</td><td align="center" rowspan="2" colspan="1">1–360 days before TB</td><td align="center" rowspan="1" colspan="1">66.1 (63.2–68.9)</td><td align="center" rowspan="1" colspan="1">80.0 (78.6–81.4)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.779 (0.761–0.798)</td></tr><tr><td rowspan="1" colspan="1">Progressors (n=73)</td><td align="center" rowspan="1" colspan="1">Nonprogressors (n=301)</td><td align="center" rowspan="1" colspan="1">53.7 (42.6–64.3)</td><td align="center" rowspan="1" colspan="1">82.8 (78.7–86)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.718 (0.637–0.800)</td></tr><tr><td rowspan="5" colspan="1">
<bold>G<sc>upta [<xref rid="C34" ref-type="bibr">34</xref>] 2022</sc></bold>
</td><td align="center" rowspan="5" colspan="1">Nine-gene signature</td><td align="center" rowspan="5" colspan="1">NR</td><td align="center" rowspan="5" colspan="1">Progressor (n=NR)</td><td align="center" rowspan="5" colspan="1">Nonprogressor (n=NR)</td><td align="center" rowspan="1" colspan="1">Up to 1 year prior to sputum conversion</td><td align="center" rowspan="1" colspan="1">76 (NR)</td><td align="center" rowspan="1" colspan="1">83 (NR)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">8–180 days prior to sputum conversion</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.934 (0.841–1)</td></tr><tr><td rowspan="1" colspan="1">181–360 days prior to sputum conversion</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.841 (0.734–0.949)</td></tr><tr><td rowspan="1" colspan="1">361–540 days prior to sputum conversion</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.59 (0.418–0.761)</td></tr><tr><td rowspan="1" colspan="1">541–720 days prior to sputum conversion</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.44 (0.213–0.667)</td></tr></tbody></table><table-wrap-foot><p>ACS: Adolescent Cohort Study; AUC: area under the curve; BATF2: basic leucine zipper ATF-like transcription factor 2; BMI: body mass index; GBP5: guanylate binding protein 5; GC: Grand Challenges; HHC: household contacts; hsa-miR: <italic toggle="yes">Homo sapiens</italic> microRNA; LTBI: latent tuberculosis infection; NPC2: Niemann–Pick type C2; NPV: negative predictive value; NR: not reported; PPV: positive predictive value; SCARF1: scavenger receptor class F member 1; SVM: support vector machine; TB: tuberculosis. <sup>#</sup>: <italic toggle="yes">ADM, AIM2</italic>, <italic toggle="yes">ANKRD22</italic>, <italic toggle="yes">APOL6</italic>, <italic toggle="yes">BATF2</italic>, <italic toggle="yes">C1QB</italic>, <italic toggle="yes">CASP5</italic>, <italic toggle="yes">CD19</italic>, <italic toggle="yes">CD274</italic>, <italic toggle="yes">CD5</italic>, <italic toggle="yes">DUSP3</italic>, <italic toggle="yes">GBP1</italic>, <italic toggle="yes">GBP4</italic>, <italic toggle="yes">GBP5</italic>, <italic toggle="yes">FBXO6</italic>, <italic toggle="yes">FCGR1B</italic>, <italic toggle="yes">GK</italic>, <italic toggle="yes">HP</italic>, <italic toggle="yes">IFIT2</italic>, <italic toggle="yes">IFIT3</italic>, <italic toggle="yes">JAK2</italic>, <italic toggle="yes">KCNJ15</italic>, <italic toggle="yes">LAP3</italic>, <italic toggle="yes">LHFPL2</italic>, <italic toggle="yes">LMNB1</italic>, <italic toggle="yes">LRRK2</italic>, <italic toggle="yes">LY96</italic>, <italic toggle="yes">MAPK14</italic>, <italic toggle="yes">NELL2</italic>, <italic toggle="yes">P2RY14</italic>, <italic toggle="yes">PSTPIP2</italic>, <italic toggle="yes">RSAD2</italic>, <italic toggle="yes">RTP4</italic>, <italic toggle="yes">SAMD9L</italic>, <italic toggle="yes">SERPING1</italic>, <italic toggle="yes">SLC6A12</italic>, <italic toggle="yes">SPOCK2</italic>, <italic toggle="yes">STAT1</italic>, <italic toggle="yes">TIMM10</italic>, <italic toggle="yes">TLR5</italic>, <italic toggle="yes">VAMP5</italic> and <italic toggle="yes">ZNF43.</italic>
<sup>¶</sup>: <italic toggle="yes">ADM</italic>, <italic toggle="yes">ANKRD22</italic>, <italic toggle="yes">APOL6</italic>, <italic toggle="yes">BATF2</italic>, <italic toggle="yes">CD274</italic>, <italic toggle="yes">CD5</italic>, <italic toggle="yes">DUSP3</italic>, <italic toggle="yes">GBP4</italic>, <italic toggle="yes">GBP5</italic>, <italic toggle="yes">FBXO6</italic>, <italic toggle="yes">FCGR1B</italic>, <italic toggle="yes">LMNB1</italic>, <italic toggle="yes">LRRK2</italic>, <italic toggle="yes">IFIT2</italic>, <italic toggle="yes">NELL2</italic>, <italic toggle="yes">SPOCK2</italic>, <italic toggle="yes">VAMP5</italic> and <italic toggle="yes">ZNF438.</italic>
<sup>+</sup>: <italic toggle="yes">CD274</italic>, <italic toggle="yes">KIF1B</italic>, <italic toggle="yes">IL15</italic>, <italic toggle="yes">TLR1</italic>, <italic toggle="yes">TLR5</italic>, <italic toggle="yes">FCGR1A</italic>, <italic toggle="yes">GBP1</italic>, <italic toggle="yes">NOD2</italic>, <italic toggle="yes">GBP2</italic> and <italic toggle="yes">EGF.</italic>
<sup>§</sup>: At 4% prevalence.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="TB3" orientation="portrait"><label>TABLE 3</label><caption><p>Summary of diagnostic performance of host protein biomarkers for predicting progression to active tuberculosis (ATB) (n=4 studies)</p></caption><table frame="hsides" rules="groups"><colgroup span="1"><col align="left" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/></colgroup><thead><tr><th align="left" rowspan="1" colspan="1">First author [ref.], year</th><th align="center" rowspan="1" colspan="1">Index test name</th><th align="center" rowspan="1" colspan="1">Reference standard used</th><th align="center" rowspan="1" colspan="1">Group 1</th><th align="center" rowspan="1" colspan="1">Group 2</th><th align="center" rowspan="1" colspan="1">Timing</th><th align="center" rowspan="1" colspan="1">Sensitivity % (95% CI)</th><th align="center" rowspan="1" colspan="1">Specificity % (95% CI)</th><th align="center" rowspan="1" colspan="1">PPV % (95% CI)</th><th align="center" rowspan="1" colspan="1">NPV % (95% CI)</th><th align="center" rowspan="1" colspan="1">AUC (95% CI)</th></tr></thead><tbody><tr><td rowspan="2" colspan="1">
<bold>B<sc>urel [<xref rid="C36" ref-type="bibr">36</xref>] 2024</sc></bold>
</td><td align="center" rowspan="1" colspan="1">IgG glycosylation profiling</td><td align="center" rowspan="2" colspan="1">No reference standard used</td><td align="center" rowspan="1" colspan="1">LTBI-Risk (as determined by Zak16) (n=10)</td><td align="center" rowspan="1" colspan="1">LTBI-Other (n=46)</td><td align="center" rowspan="1" colspan="1">Pre-treatment</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.71</td></tr><tr><td rowspan="1" colspan="1">AbGlyc score</td><td align="center" rowspan="1" colspan="1">ATB (n=19)<break/> Mean age 33 years<break/> 79% male</td><td align="center" rowspan="1" colspan="1">LTBI-R (n=10)<break/> Mean age 29 years<break/> 54% male</td><td align="center" rowspan="1" colspan="1">Pre-treatment</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.82</td></tr><tr><td rowspan="2" colspan="1">
<bold>A<sc>raujo [<xref rid="C37" ref-type="bibr">37</xref>] 2018</sc></bold>
</td><td align="center" rowspan="1" colspan="1">p-29878/MMP9</td><td align="center" rowspan="2" colspan="1">Chest radiography</td><td align="center" rowspan="2" colspan="1">Incipient TB (n=28)</td><td align="center" rowspan="2" colspan="1">Healthy controls (n=14)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">61.54 (31.25–91.83)</td><td align="center" rowspan="1" colspan="1">50.00 (14.01–85.99)</td><td align="center" rowspan="1" colspan="1">61.55 (31.25–91.83)</td><td align="center" rowspan="1" colspan="1">50.00 (14.01–85.99)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="1" colspan="1">p-29878/uPAR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">46.43 (26.17–66.69)</td><td align="center" rowspan="1" colspan="1">86.67 (66.13–100.0)</td><td align="center" rowspan="1" colspan="1">86.67 (66.13–100.0)</td><td align="center" rowspan="1" colspan="1">46.43 (26.17–66.69)</td><td align="center" rowspan="1" colspan="1">NR</td></tr><tr><td rowspan="8" colspan="1">
<bold>S<sc>inger [<xref rid="C39" ref-type="bibr">39</xref>] 2023</sc></bold>
</td><td align="center" rowspan="1" colspan="1">Eight-protein signature (comprising CD14, CD166, DSG2, LAMP1, LRP1, NCAM2, R4RL2, VASN)</td><td align="center" rowspan="4" colspan="1">NR</td><td align="center" rowspan="4" colspan="1">ATB (n=24)<break/> Mean age 38 years<break/> USA cohort</td><td align="center" rowspan="4" colspan="1">Non-TB (n=48)<break/> Mean age 39 years</td><td align="center" rowspan="1" colspan="1">0–6 months prior to TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.74</td></tr><tr><td rowspan="1" colspan="1">12-protein signature (comprising AMPN, CA2D1, CBPQ, CNTN1, COL11, DPP4, HYOU1, ICAM1, LAMP1, LUM, PCOC1, PIGR)</td><td align="center" rowspan="1" colspan="1">18–24 months prior to TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.8</td></tr><tr><td rowspan="1" colspan="1">Seven-protein signature (comprising APOA1, APOA4, GP1BA, LUM, NID1, PLSL, SCTM1)</td><td align="center" rowspan="1" colspan="1">6–12 months prior to TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.76</td></tr><tr><td rowspan="1" colspan="1">Seven-protein signature (comprising APOA4, CATA, COL11, LUM, MEGF8, PLSL, SCTM1)</td><td align="center" rowspan="1" colspan="1">12–18 months prior to TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.84</td></tr><tr><td rowspan="1" colspan="1">Five-protein signature (comprising CD248, HYOU1, NID1, PON1, VCAM1)</td><td align="center" rowspan="4" colspan="1">Culture (nonspecified), chest radiography, smear microscopy</td><td align="center" rowspan="4" colspan="1">ATB (n=30)<break/> Females 73%<break/> South African cohort</td><td align="center" rowspan="4" colspan="1">Non-TB (n=62)<break/> Female 76%</td><td align="center" rowspan="1" colspan="1">0–6 months prior to TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.93</td></tr><tr><td rowspan="1" colspan="1">Six-protein signature (comprising CD248, HYOU1, NID1, PON1, VCAM1, APOA1)</td><td align="center" rowspan="1" colspan="1">6–12 months prior to TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.86</td></tr><tr><td rowspan="1" colspan="1">Eight-protein signature (comprising APOA4, CBPN, CD14, CPN2, LCAT, LUM, LYAM1, PNPH)</td><td align="center" rowspan="1" colspan="1">12–18 months before diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.75</td></tr><tr><td rowspan="1" colspan="1">11 protein signature (comprising CD248, CSTN1, FUCO2, HYOU1, MINP1, MMP2, MYOC, NID1, PON1, VASN, VCAM1)</td><td align="center" rowspan="1" colspan="1">18–24 months prior to TB diagnosis</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.87</td></tr><tr><td rowspan="20" colspan="1">
<bold>P<sc>enn-</sc>N<sc>icholson [<xref rid="C38" ref-type="bibr">38</xref>] 2019</sc></bold>
</td><td align="center" rowspan="8" colspan="1">3PR</td><td align="center" rowspan="4" colspan="1">Smear microscopy, culture (nonspecified)</td><td align="center" rowspan="4" colspan="1">Progressor (ACS training+test cohort) (n=37)</td><td align="center" rowspan="4" colspan="1">Nonprogressor (ACS training+test cohort) (n=106)</td><td align="center" rowspan="1" colspan="1">Time to TB 0–12 months</td><td align="center" rowspan="1" colspan="1">75 (61.05–85.97)</td><td align="center" rowspan="1" colspan="1">70.69 (65.09–75.87)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.80 (0.74–0.86)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 0–6 months</td><td align="center" rowspan="1" colspan="1">90.91 (70.84–98.88)</td><td align="center" rowspan="1" colspan="1">70.69 (65.09–75.87)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.89 (0.84–0.95)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 7–12 months</td><td align="center" rowspan="1" colspan="1">63.33 (43.86–80.07)</td><td align="center" rowspan="1" colspan="1">70.69 (65.09–75.87)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.72 (0.64–0.81)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 13–24 months</td><td align="center" rowspan="1" colspan="1">57.14 (37.18–75.54)</td><td align="center" rowspan="1" colspan="1">70.69 (65.09–75.87)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.71 (0.63–0.80)</td></tr><tr><td rowspan="4" colspan="1">Culture (nonspecified), smear microscopy, chest radiography, clinical signs</td><td align="center" rowspan="4" colspan="1">Progressor (GC6 validation cohort) (n=34)</td><td align="center" rowspan="4" colspan="1">Nonprogressor (GC6 validation cohort) (n=115)</td><td align="center" rowspan="1" colspan="1">Time to TB 0–12 months</td><td align="center" rowspan="1" colspan="1">46.34 (30.66–62.58)</td><td align="center" rowspan="1" colspan="1">75.0 (68.2–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.65 (0.55–0.75)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 0–6 months</td><td align="center" rowspan="1" colspan="1">47.83 (26.82–69.41)</td><td align="center" rowspan="1" colspan="1">75.0 (68.26–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.64 (0.50–0.78)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 7–12 months</td><td align="center" rowspan="1" colspan="1">44.44 (21.53–69.24)</td><td align="center" rowspan="1" colspan="1">75.0 (68.26–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.67 (0.55–0.79)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 13–24 months</td><td align="center" rowspan="1" colspan="1">42.11 (20.25–66.5)</td><td align="center" rowspan="1" colspan="1">75.0 (68.26–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.62 (0.48–0.76)</td></tr><tr><td rowspan="12" colspan="1">TRM5</td><td align="center" rowspan="4" colspan="1">Smear microscopy, culture (nonspecified)</td><td align="center" rowspan="4" colspan="1">Progressor (ACS training cohort) (n=24)</td><td align="center" rowspan="4" colspan="1">Nonprogressor (ACS training cohort) (n=70)</td><td align="center" rowspan="1" colspan="1">Time to TB 0–12 months</td><td align="center" rowspan="1" colspan="1">75.76 (57.74–88.91)</td><td align="center" rowspan="1" colspan="1">75.0 (67.22–81.75)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.84 (0.75–0.92)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 0–6 months</td><td align="center" rowspan="1" colspan="1">100 (73.54–100)</td><td align="center" rowspan="1" colspan="1">75.0 (67.22–81.75)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.96 (0.93–0.99)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 7–12 months</td><td align="center" rowspan="1" colspan="1">61.9 (38.44–81.89)</td><td align="center" rowspan="1" colspan="1">75.0 (67.22–81.75)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.76 (0.65–0.87)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 13–24 months</td><td align="center" rowspan="1" colspan="1">37.5 (18.8–59.41)</td><td align="center" rowspan="1" colspan="1">70.27 (62.21–77.5)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.55 (0.41–0.69)</td></tr><tr><td rowspan="4" colspan="1">Smear microscopy, culture (nonspecified)</td><td align="center" rowspan="4" colspan="1">Progressor (ACS test cohort) (n=13)</td><td align="center" rowspan="4" colspan="1">Nonprogressor (ACS test cohort) (n=36)</td><td align="center" rowspan="1" colspan="1">Time to TB 0–12 months</td><td align="center" rowspan="1" colspan="1">78.95 (54.43–93.95)</td><td align="center" rowspan="1" colspan="1">75.27 (65.24–83.63)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.80 (0.70–0.89)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 0–6 months</td><td align="center" rowspan="1" colspan="1">90 (55.5–99.75)</td><td align="center" rowspan="1" colspan="1">75.27 (65.24–83.63)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.85 (0.75–0.96)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 7–12 months</td><td align="center" rowspan="1" colspan="1">66.67 (29.93–92.5)</td><td align="center" rowspan="1" colspan="1">75.27 (65.24–83.63)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.73 (0.62–0.85)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 13–24 months</td><td align="center" rowspan="1" colspan="1">37.5 (8.52–75.51)</td><td align="center" rowspan="1" colspan="1">75.27 (65.24–83.63)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.69 (0.53–0.84)</td></tr><tr><td rowspan="4" colspan="1">Culture (nonspecified), smear microscopy, chest radiography, clinical signs</td><td align="center" rowspan="4" colspan="1">Progressor (GC6 validation cohort) (n=34)</td><td align="center" rowspan="4" colspan="1">Nonprogressor (GC6 validation cohort) (n=115)</td><td align="center" rowspan="1" colspan="1">Time to TB 0–12 months</td><td align="center" rowspan="1" colspan="1">48.78 (32.88–64.87)</td><td align="center" rowspan="1" colspan="1">75.0 (68.26–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.66 (0.56–0.75)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 0–6 months</td><td align="center" rowspan="1" colspan="1">60.87 (38.54–80.29)</td><td align="center" rowspan="1" colspan="1">75.0 (68.26–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.69 (0.57–0.82)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 7–12 months</td><td align="center" rowspan="1" colspan="1">33.33 (13.34–59.01)</td><td align="center" rowspan="1" colspan="1">75.0 (68.26–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.61 (0.48–0.74)</td></tr><tr><td rowspan="1" colspan="1">Time to TB 13–24 months</td><td align="center" rowspan="1" colspan="1">36.84 (16.29–61.64)</td><td align="center" rowspan="1" colspan="1">75.0 (68.26–80.96)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.61 (0.47–0.75)</td></tr></tbody></table><table-wrap-foot><p>3PR: 3-protein pair-ratio; AbGlyc: antibody glycosylation; ACS: Adolescent Cohort Study; AMPN: aminopeptidase N; APOA1: apolipoprotein A-I; APOA4: apolipoprotein A-IV; AUC: area under the curve; CA2D1: calcium voltage-gated channel subunit alpha-2/delta-1; CATA: catalase; CBPN: carboxypeptidase N catalytic chain; CBPQ: carboxypeptidase Q; CD: cluster of differentiation; CNTN1: contactin 1; COL11: collectin-II; CoR: Correlation of Risk; CPN2: carboxypeptidase N subunit 2; CSTN1: calsyntenin 1; DSG2: desmoglein-2; DPP4: dipeptidyl peptidase-4; FUCO2: plasma alpha-L-fucosidase; GC6: Grand Challenges 6; GP1BA: platelet glycoprotein Ib alpha chain; HHC: household contacts; HYOU1: hypoxia upregulated protein 1; ICAM1: intercellular adhesion molecule 1; LAMP1: lysosomal associated membrane protein 1; LCAT: lecithin-cholesterol acyltransferase; LRP1: prolow-density lipoprotein receptor related protein 1; LTBI: latent tuberculosis infection; LUM: lumican; LYAM1: L-selectin; MEGF8: multiple epidermal growth factor like domains protein 8; MINP1: multiple inositol polyphosphate phosphotase 1; MMP2: matrix metalloproteinase-2; MYOC: myocilin; NCAM2: neural cell adhesion molecule 2; NID1: nidogen 1; NPV: negative predictive value; NR: not reported; PCOC1: procollagen C-endopeptidase enhancer 1; PIGR: polymeric immunoglobulin receptor; PLSL: plastin-2; PNPH: purine nucleoside phosphorylase; PON1: serum paraoxonase/arylesterase 1; PPV: positive predictive value; R4RL2: reticulon-4 receptor-like 2; SCTM1: secreted and transmembrane protein 1; TB: tuberculosis; TRM5: TB Risk Model 5; uPAR: urokinase-type plasminogen activator receptor; VASN: vasorin; VCAM1: vascular cell adhesion molecule 1.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="TB4" orientation="portrait"><label>TABLE 4</label><caption><p>Summary of diagnostic performance of pathogen cytokines or chemokines biomarkers for predicting progression to active tuberculosis (ATB) (n=3 studies)</p></caption><table frame="hsides" rules="groups"><colgroup span="1"><col align="left" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/><col align="center" span="1"/></colgroup><thead><tr><th align="left" rowspan="1" colspan="1">First author [ref.], year</th><th align="center" rowspan="1" colspan="1">Index test name</th><th align="center" rowspan="1" colspan="1">Reference standard used</th><th align="center" rowspan="1" colspan="1">Group 1</th><th align="center" rowspan="1" colspan="1">Group 2</th><th align="center" rowspan="1" colspan="1">Timing</th><th align="center" rowspan="1" colspan="1">Sensitivity % (95% CI)</th><th align="center" rowspan="1" colspan="1">Specificity % (95% CI)</th><th align="center" rowspan="1" colspan="1">PPV % (95% CI)</th><th align="center" rowspan="1" colspan="1">NPV % (95% CI)</th><th align="center" rowspan="1" colspan="1">AUC (95% CI)</th></tr></thead><tbody><tr><td rowspan="17" colspan="1">
<bold>D<sc>aniel [<xref rid="C40" ref-type="bibr">40</xref>] 2023</sc></bold>
</td><td align="center" rowspan="1" colspan="1">CCL11</td><td align="center" rowspan="16" colspan="1">Chest radiography, liquid culture, smear microscopy, TST, IGRA</td><td align="center" rowspan="17" colspan="1">HHC (progressors to ATB) (n=14)</td><td align="center" rowspan="17" colspan="1">HHC (nonprogressors) (n=20)</td><td align="center" rowspan="15" colspan="1">2-year follow-up</td><td align="center" rowspan="1" colspan="1">100 (85.8–100)</td><td align="center" rowspan="1" colspan="1">75 (50.9–91.3)</td><td align="center" rowspan="1" colspan="1">82.8 (64.2–94.2)</td><td align="center" rowspan="1" colspan="1">100 (78.2–100)</td><td align="center" rowspan="1" colspan="1">88.6 (75.4–96.2) cutoff<sup>#</sup> >204.20</td></tr><tr><td rowspan="1" colspan="1">CCL19</td><td align="center" rowspan="1" colspan="1">100 (85.8–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (85.8–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (92–100) cutoff<sup>#</sup> <628.98</td></tr><tr><td rowspan="1" colspan="1">CCL3</td><td align="center" rowspan="1" colspan="1">91.7 (73–99)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (84.6–100)</td><td align="center" rowspan="1" colspan="1">90.9 (70.8–98.9)</td><td align="center" rowspan="1" colspan="1">95.5 (84.5–99.4)<break/> cutoff<sup>#</sup> >234.20</td></tr><tr><td rowspan="1" colspan="1">CCL4</td><td align="center" rowspan="1" colspan="1">66.7 (44.7–84.4)</td><td align="center" rowspan="1" colspan="1">80 (56.3–94.3)</td><td align="center" rowspan="1" colspan="1">80 (56.3–94.3)</td><td align="center" rowspan="1" colspan="1">66.7 (44.7–84.4)</td><td align="center" rowspan="1" colspan="1">72.7 (57.2–85)<break/> cutoff<sup>#</sup> >862.18</td></tr><tr><td rowspan="1" colspan="1">GM-CSF</td><td align="center" rowspan="1" colspan="1">91.7 (73–99)</td><td align="center" rowspan="1" colspan="1">90 (68.3–98.8)</td><td align="center" rowspan="1" colspan="1">91.7 (73–99)</td><td align="center" rowspan="1" colspan="1">90 (68.3–98.8)</td><td align="center" rowspan="1" colspan="1">90.9 (78.3–97.5)<break/> cutoff<sup>#</sup> >128.73</td></tr><tr><td rowspan="1" colspan="1">IFN-α</td><td align="center" rowspan="1" colspan="1">79.2 (57.8–92.9)</td><td align="center" rowspan="1" colspan="1">80 (56.3–94.3)</td><td align="center" rowspan="1" colspan="1">82.6 (61.2–95)</td><td align="center" rowspan="1" colspan="1">76.2 (52.8–91.8)</td><td align="center" rowspan="1" colspan="1">79.5 (64.7–90.2)<break/> cutoff<sup>#</sup> >326.45</td></tr><tr><td rowspan="1" colspan="1">IFN-β</td><td align="center" rowspan="1" colspan="1">79.2 (57.8–92.9)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (82.4–100)</td><td align="center" rowspan="1" colspan="1">80 (59.3–93.2)</td><td align="center" rowspan="1" colspan="1">88.6 (75.4–96.2)<break/> cutoff<sup>#</sup> >36.76</td></tr><tr><td rowspan="1" colspan="1">IFN-γ</td><td align="center" rowspan="1" colspan="1">100 (85.8–100)</td><td align="center" rowspan="1" colspan="1">90 (68.3–98.8)</td><td align="center" rowspan="1" colspan="1">92.3 (74.9–99.1)</td><td align="center" rowspan="1" colspan="1">100 (81.5–100)</td><td align="center" rowspan="1" colspan="1">95.5 (84.5–99.4)<break/> cutoff<sup>#</sup> <491.20</td></tr><tr><td rowspan="1" colspan="1">IL-1ra</td><td align="center" rowspan="1" colspan="1">100 (85.8–100)</td><td align="center" rowspan="1" colspan="1">85 (62.1–96.8)</td><td align="center" rowspan="1" colspan="1">88.9 (70.8–97.6)</td><td align="center" rowspan="1" colspan="1">100 (80.5–100)</td><td align="center" rowspan="1" colspan="1">93.2 (81.3–98.6)<break/> cutoff<sup>#</sup> <2405.48</td></tr><tr><td rowspan="1" colspan="1">IL-1α</td><td align="center" rowspan="1" colspan="1">79.2 (57.8–92.9)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (82.4–100)</td><td align="center" rowspan="1" colspan="1">80 (59.3–93.2)</td><td align="center" rowspan="1" colspan="1">88.6 (75.4–96.2)<break/> cutoff<sup>#</sup> <305.80</td></tr><tr><td rowspan="1" colspan="1">IL-2</td><td align="center" rowspan="1" colspan="1">62.5 (40.6–81.2)</td><td align="center" rowspan="1" colspan="1">95 (75.1–99.9)</td><td align="center" rowspan="1" colspan="1">93.8 (69.8–99.8)</td><td align="center" rowspan="1" colspan="1">67.9 (47.6–84.1)</td><td align="center" rowspan="1" colspan="1">77.3 (62.2–88.5)<break/> cutoff<sup>#</sup> <180.94</td></tr><tr><td rowspan="1" colspan="1">IL-4</td><td align="center" rowspan="1" colspan="1">79.2 (57.8–92.9)</td><td align="center" rowspan="1" colspan="1">80 (56.3–94.3)</td><td align="center" rowspan="1" colspan="1">82.6 (61.2–95)</td><td align="center" rowspan="1" colspan="1">76.2 (52.8–91.8)</td><td align="center" rowspan="1" colspan="1">79.5 (64.7–90.2)<break/> cutoff<sup>#</sup> >141.14</td></tr><tr><td rowspan="1" colspan="1">IL-5</td><td align="center" rowspan="1" colspan="1">83.3 (62.6–95.3)</td><td align="center" rowspan="1" colspan="1">80 (56.3–94.3)</td><td align="center" rowspan="1" colspan="1">83.3 (62.6–95.3)</td><td align="center" rowspan="1" colspan="1">80 (56.3–94.3)</td><td align="center" rowspan="1" colspan="1">81.8 (67.3–91.8)<break/> cutoff<sup>#</sup> >75.29</td></tr><tr><td rowspan="1" colspan="1">IL-6</td><td align="center" rowspan="1" colspan="1">50 (29.1–70.9)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (73.5–100)</td><td align="center" rowspan="1" colspan="1">62.5 (43.7–78.9)</td><td align="center" rowspan="1" colspan="1">72.7 (57.2–85)<break/> cutoff<sup>#</sup> >227.00</td></tr><tr><td rowspan="1" colspan="1">IP-10</td><td align="center" rowspan="1" colspan="1">100 (85.8–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (85.8–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (92–100)<break/> cutoff<sup>#</sup> >226.06</td></tr><tr><td rowspan="1" colspan="1">QuantiFERON supernatants (IP10/CCL19)</td><td align="center" rowspan="1" colspan="1"><6 months</td><td align="center" rowspan="1" colspan="1">100 (73.5–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (73.5–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (89.1–100)<break/> cutoff >0.24</td></tr><tr><td rowspan="1" colspan="1">QuantiFERON supernatants (IP10/CCL19)</td><td align="center" rowspan="1" colspan="1">TST, IGRA, liquid culture, smear microscopy, chest radiography</td><td align="center" rowspan="1" colspan="1">>6 months</td><td align="center" rowspan="1" colspan="1">100 (73.5–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (73.5–100)</td><td align="center" rowspan="1" colspan="1">100 (83.2–100)</td><td align="center" rowspan="1" colspan="1">100 (89.1–100)<break/> cutoff >0.24</td></tr><tr><td rowspan="2" colspan="1">
<bold>R<sc>obison [<xref rid="C41" ref-type="bibr">41</xref>] 2021</sc></bold>
</td><td align="center" rowspan="1" colspan="1">Threshold-based reduced random forest feature analysis for the high-risk clinical designation – 14 features</td><td align="center" rowspan="2" colspan="1">IGRA, other</td><td align="center" rowspan="2" colspan="1">High risk of TB reactivation (n=24)<break/> Mean±<sc>sd</sc> age 45.7±18.6<break/> 54.2% females</td><td align="center" rowspan="2" colspan="1">Unexposed control group (n=23)<break/> mean±<sc>sd</sc> age 58.5±16.0<break/> 80% females<break/> Includes patients under treatment for LTBI</td><td align="center" rowspan="1" colspan="1">Baseline</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.855</td></tr><tr><td rowspan="1" colspan="1">Full random plot analysis <break/>143 features</td><td align="center" rowspan="1" colspan="1">Baseline</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.715</td></tr><tr><td rowspan="16" colspan="1">
<bold>Z<sc>hang [<xref rid="C42" ref-type="bibr">42</xref>] 2020</sc></bold>
</td><td align="center" rowspan="1" colspan="1">Basic FGF</td><td align="center" rowspan="16" colspan="1">Smear microscopy, Xpert, culture (nonspecified)</td><td align="center" rowspan="16" colspan="1">ATB cases (untreated) (n=9)</td><td align="center" rowspan="16" colspan="1">Non-TB cases (untreated) <italic toggle="yes">i.e.</italic> QTF positive (n=18)<break/> Age and gender-matched to ATB cases</td><td align="center" rowspan="16" colspan="1">To further evaluate the performance on predicting active TB, the AUCs were calculated among untreated participants at T0 (baseline, 1 week before starting treatment)</td><td align="center" rowspan="1" colspan="1">88.89 (51.75–99.72)</td><td align="center" rowspan="1" colspan="1">94.44 (72.71–99.86)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.96 (0.88–1.02)<break/> cutoff 17.49 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">Eotaxin</td><td align="center" rowspan="1" colspan="1">88.89 (51.75–99.72)</td><td align="center" rowspan="1" colspan="1">94.44 (72.71–99.86)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.95 (0.87–1.03)<break/> cutoff 50.54 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">G-CSF</td><td align="center" rowspan="1" colspan="1">88.89 (51.75–99.72)</td><td align="center" rowspan="1" colspan="1">88.89 (65.29–98.62)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.89 (0.76–1.02)<break/> cutoff 58.89 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">IFN-γ</td><td align="center" rowspan="1" colspan="1">88.89 (51.75–99.72)</td><td align="center" rowspan="1" colspan="1">94.44 (72.71–99.86)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.95 (0.87–1.03)<break/> cutoff 17.41 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">IL-17A</td><td align="center" rowspan="1" colspan="1">100.00 (66.37–100.0)</td><td align="center" rowspan="1" colspan="1">72.22 (46.52–90.31)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.80 (0.63–0.98)<break/> cutoff 3.48 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">IL-1ra</td><td align="center" rowspan="1" colspan="1">100.00 (66.37–100.0)</td><td align="center" rowspan="1" colspan="1">72.22 (46.52–90.31)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.85 (0.70–1.00)<break/> cutoff 114.70 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">IL-1α</td><td align="center" rowspan="1" colspan="1">88.89 (51.75–99.72)</td><td align="center" rowspan="1" colspan="1">88.89 (65.29–98.62)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.92 (0.82–1.02)<break/> cutoff 8.06 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">IL-4</td><td align="center" rowspan="1" colspan="1">100.00 (66.37–100.0)</td><td align="center" rowspan="1" colspan="1">100.00 (81.47–100.0)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">1.00 (1.00–1.00) cutoff 1.16 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">IL-8</td><td align="center" rowspan="1" colspan="1">77.78 (39.99–97.19)</td><td align="center" rowspan="1" colspan="1">83.33 (58.58–96.42)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.86 (0.73–1.00) cutoff 3.44 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">IL-9</td><td align="center" rowspan="1" colspan="1">88.89 (51.75–99.72)</td><td align="center" rowspan="1" colspan="1">94.44 (72.71–99.86)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.90 (0.77–1.04) cutoff 38.85 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">MCP-1 (MCAF)</td><td align="center" rowspan="1" colspan="1">66.67 (29.93–92.51)</td><td align="center" rowspan="1" colspan="1">83.33 (58.58–96.42)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.79 (0.59–0.99) cutoff 10.34 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">MIP-1β</td><td align="center" rowspan="1" colspan="1">55.56 (21.20–86.30)</td><td align="center" rowspan="1" colspan="1">94.44 (72.71–99.86)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.75 (0.55–0.94)<break/> cutoff 57.65 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">SDF-1α</td><td align="center" rowspan="1" colspan="1">100.00 (66.37–100.0)</td><td align="center" rowspan="1" colspan="1">50.00 (26.02–73.98)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.79 (0.61–0.97) cutoff 218.00 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">TNF-α</td><td align="center" rowspan="1" colspan="1">100.00 (66.37–100.0)</td><td align="center" rowspan="1" colspan="1">61.11 (35.75–82.70)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.82 (0.66–0.97) cutoff 10.48 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">TRAIL</td><td align="center" rowspan="1" colspan="1">77.78 (39.99–97.19)</td><td align="center" rowspan="1" colspan="1">88.89 (65.29–98.62)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.88 (0.74–1.01) cutoff 34.51 pg·mL<sup>−1</sup></td></tr><tr><td rowspan="1" colspan="1">VEGF-A</td><td align="center" rowspan="1" colspan="1">88.89 (51.75–99.72)</td><td align="center" rowspan="1" colspan="1">72.22 (46.52–90.31)</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">NR</td><td align="center" rowspan="1" colspan="1">0.83 (0.66–1.00) cutoff 14.31 pg·mL<sup>−1</sup></td></tr></tbody></table><table-wrap-foot><p>AUC: area under the curve; CCL11: C-C motif chemokine ligand 11 (eotaxin-1); FGF: fibroblast growth factor; G-CSF: granulocyte colony-stimulating factor; GM-CSF: granulocyte macrophage–colony-stimulating factor; HHC: household contacts; IFN: interferon; IGRA: interferon gamma release assay; IL: interleukin; IP10: interferon gamma-induced protein 10; LTBI: latent tuberculosis infection; MCP-1 (MCAF): monocyte chemoattractant protein-1 (monocyte chemotactic and activating factor, CCL2); MIP-1β: macrophage inflammatory protein-1 beta (CCL4); NPV: negative predictive value; NR: not reported; PPV: positive predictive value; QFT: QuantiFERON-TB test; SDF-1α: stromal cell derived factor-1 alpha (CXCL12); TB: tuberculosis; TNF-α: tumour necrosis factor alpha; TRAIL: TNF-related apoptosis inducing ligand; TST: tuberculin skin test; VEGF-A: vascular endothelial growth factor A. <sup>#</sup>: As determined by classification and regression tree (CART) model.</p></table-wrap-foot></table-wrap><sec id="s3d1"><title>Transcriptomic signatures</title><p>Our systematic review identified several studies that investigated the potential of transcriptomic profiling as a biomarker approach predicting progression from TBI to active TB (10 studies) [<xref rid="C14" ref-type="bibr">14</xref>, <xref rid="C27" ref-type="bibr">27</xref>–<xref rid="C35" ref-type="bibr">35</xref>]. Across these 10 studies, 23 unique gene signatures were identified, with data on sensitivity, specificity and predictive values at multiple time points prior to disease onset (<xref rid="TB2" ref-type="table">table 2</xref>). Although many signatures demonstrated promising performance, significant heterogeneity in population characteristics, reference standards and reporting practices limited interpretability and comparability.</p><sec id="s3d1a"><title>Short-term prediction</title><p>10 studies investigated gene signatures designed to predict progression to active TB within different time windows (<xref rid="TB2" ref-type="table">table 2</xref>). For example, V<sc>argas</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C27" ref-type="bibr">27</xref>] assessed a 42-gene signature and reported a sensitivity of 87.3% (95% CI 81.5–93.1) and specificity of 91.7% (95% CI 89.9–93.5) for predicting progression within three months of TB diagnosis by conventional methods. Performance declined over extended intervals (>18 months), suggesting that some transcriptomic markers may primarily support short-term prediction.</p><p>P<sc>enn-</sc>N<sc>icholson</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C31" ref-type="bibr">31</xref>] evaluated RISK6, reporting an AUC of 87.6% (95% CI 82.8–92.4) for predictions within 0–12 months. Sensitivity decreased to 50% and specificity remained 82.8% (95% CI not reported) for predictions 12–24 months prior to disease onset. The study authors provided minimal details on the study participants’ TB exposure or TB infection status definitions, adding uncertainty about the test's applicability in broader cohorts.</p></sec><sec id="s3d1b"><title>Longer-term prediction</title><p>L<sc>eong</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C32" ref-type="bibr">32</xref>] assessed multiple previously reported gene signatures up to 5 years prior to disease progression, finding AUCs ranging from 0.62 (95% CI 0.59–0.65) to 0.91 (95% CI 0.89–93). Among most studies included in our review, reporting was limited for several important characteristics, including how reference standards were applied, and lacked detailed demographic data, complicating the applicability of results. A lack of reporting on comorbidities or sociodemographic factors in the study population limits the extrapolation of these findings to diverse settings.</p></sec><sec id="s3d1c"><title>Micro-RNA signatures</title><p>X<sc>in</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C28" ref-type="bibr">28</xref>] evaluated microRNA markers (<italic toggle="yes">e.g.</italic> hsa-miR-16-5p and hsa-miR-451a), reporting a sensitivity of 80.8% (95% CI 70.3–88.2) and specificity of 79.2% (95% CI 68.9–86.8) over 5 years when adjusted for covariables (<xref rid="TB2" ref-type="table">table 2</xref>). However, inconsistent reference standard definitions across studies hinder direct comparisons and underscore the need for harmonised methods and criteria when assessing microRNA-based markers.</p></sec></sec><sec id="s3d2"><title>Proteomic signatures</title><p>Four studies evaluated proteomic signatures for predicting TB progression [<xref rid="C36" ref-type="bibr">36</xref>–<xref rid="C39" ref-type="bibr">39</xref>] (<xref rid="TB3" ref-type="table">table 3</xref>). Performance varied across markers and panels.</p><sec id="s3d2a"><title>IgG glycosylation profiling</title><p>B<sc>urel</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C36" ref-type="bibr">36</xref>] evaluated IgG glycosylation features to differentiate higher-risk individuals (based on the Zak16 signature) from lower-risk TBI populations, reporting an AUC of 0.71 (95% CI not reported) without using a reference standard. An AbGlyc score distinguished active TB from higher-risk TBI with an AUC of 0.82 (95% CI not reported), indicating moderate but improved discrimination potential. Like studies focused on transcriptomic profiles, there was significant inter-study heterogeneity, which limits the interpretability and comparability of results.</p></sec><sec id="s3d2b"><title>Protein panels</title><p>S<sc>inger</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C39" ref-type="bibr">39</xref>] examined multiple protein panels at varying time intervals before active TB diagnosis (<xref rid="TB3" ref-type="table">table 3</xref>), identifying a five-protein panel (cluster of differentiation 248, hypoxia upregulated protein 1, nidogen 1, paraoxonase 1 and vascular cell adhesion molecule 1) that performed best within 0–6 months of disease onset (AUC 0.93, 95% CI not reported) while a six-protein panel measured 6–12 months before diagnosis achieved an AUC of 0.86 (95% CI not reported). Additional panels assessed in this study demonstrated moderate performance, with AUCs ranging from 0.75 to 0.87 depending on the timing of sample collection (<xref rid="TB2" ref-type="table">table 2</xref>). In P<sc>enn-</sc>N<sc>icholson</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C38" ref-type="bibr">38</xref>], the tRNA (guanine-N¹)-methyltransferase 5 protein showed the strongest discriminatory ability, with an AUC of 0.96 (95% CI 0.93–0.99), 100% sensitivity (95% CI 73.5–100.0) and 75% specificity (95% CI 67.2–81.7) for predicting TB progression within 6 months in their training cohort, although performance declined substantially for longer prediction intervals (AUCs 0.55–0.76 up to 24 months). While proteomic markers demonstrate some promising short-term predictive potential, heterogeneity in study design, populations and reference standards limits direct comparability across studies and biomarkers.</p></sec></sec><sec id="s3d3"><title>Cytokines/chemokines</title><p>Our review identified three studies [<xref rid="C40" ref-type="bibr">40</xref>–<xref rid="C42" ref-type="bibr">42</xref>] that investigated pathogen-induced cytokine and chemokine biomarkers for predicting progression to active TB. The findings highlight varying diagnostic performance across different biomarkers (<xref rid="TB4" ref-type="table">table 4</xref>).</p><p>D<sc>aniel</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C40" ref-type="bibr">40</xref>] evaluated multiple cytokines, including C-C motif chemokine ligand (CCL) 19 and interferon gamma-induced protein 10, both showing 100% sensitivity and specificity (AUC 1.00, 95% CI 0.92–1.00) for prediction within 2 years. However, confidence intervals were fairly wide (83.2–100.0%) and sample sizes small (<20 per group) (<xref rid="TB4" ref-type="table">table 4</xref>). Notably, the CCL19 marker outperformed others, with a perfect sensitivity, specificity, PPV and NPV at optimal cutoff values, suggesting its high potential as a predictive biomarker. Other cytokines such as CCL11 and granulocyte macrophage–colony-stimulating factor also showed promising results with sensitivities and specificities over 90% but slightly lower AUC values.</p><p>Z<sc>hang</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C42" ref-type="bibr">42</xref>] examined various cytokines, including interferon (IFN)-γ, interleukin (IL)-17A and IL-4, in exposed individuals. IL-4 achieved the highest AUC of 1.00, indicating excellent diagnostic capacity, while IFN-γ and IL-1α had AUC values between 0.95 (95% CI 0.87–1.03) and 0.92 (95% CI 0.82–1.02), with sensitivities and specificities close to 90%. In contrast, some markers, such as macrophage inflammatory protein-1β and monocyte chemoattractant protein-1, displayed moderate AUCs (0.75 and 0.79) with lower sensitivity values, indicating more limited predictive utility for TB progression.</p><p>R<sc>obison</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C41" ref-type="bibr">41</xref>] applied a random forest model to cytokine data among individuals at high risk of reactivation, achieving an AUC of 0.855 (95% CI not reported).</p></sec></sec></sec><sec sec-type="discussion" id="s4"><title>Discussion</title><p>Our systematic review highlights the rapidly evolving landscape of biomarker-based diagnostics for predicting progression from TBI to active disease. Although multiple promising host- and pathogen-derived signatures have emerged since 2016, no single test has yet demonstrated sufficient accuracy, reproducibility or generalisability to support clinical implementation. This gap is particularly pressing given the global push toward TB elimination, the emerging interest in fighting asymptomatic TB [<xref rid="C43" ref-type="bibr">43</xref>] and the limited predictive utility of existing tests such as TSTs and IGRAs [<xref rid="C4" ref-type="bibr">4</xref>].</p><p>Among the various biomarkers evaluated, transcriptomic signatures have received the most attention, with several studies investigating whole-blood gene expression profiles. While some signatures, such as the 16-gene correlate developed by Z<sc>ak</sc>
<italic toggle="yes">et al</italic>. [<xref rid="C14" ref-type="bibr">14</xref>], have shown encouraging predictive performance, their accuracy tends to decline as the prediction window increases. For instance, signatures that perform well within a 6-month prediction interval often fall below acceptable thresholds (as defined by WHO target product profiles (TPPs)) at 12 or 24 months [<xref rid="C38" ref-type="bibr">38</xref>]. Biomarker performance possibly declines as the prediction window increases because early host responses may wane over time, biological signals become less specific and greater variability in progression timing introduces noise [<xref rid="C38" ref-type="bibr">38</xref>, <xref rid="C44" ref-type="bibr">44</xref>]. Predictive performance is also shaped by pre-test probability; biomarkers tend to perform better in high-risk populations, while their positive predictive value declines in lower-risk settings, even if sensitivity and specificity remain unchanged. Thus, performance is context-dependent, with factors such as concurrent infections, genetic diversity and geographic variation potentially affecting test utility.</p><p>While several biomarkers evaluated in this review demonstrated sensitivity and specificity that meet or exceed WHO-recommended TPP thresholds, particularly within short prediction windows, none have progressed to late-stage clinical trials or implementation studies. This disconnect between promising accuracy estimates and limited translational advancement likely reflects a combination of factors. First, many studies were conducted under idealised conditions with highly selected populations and performance may not generalise to routine clinical settings. Second, studies often lacked critical information on test reproducibility, cost and operational feasibility, which are key considerations for developers and public health programmes. Third, incomplete reporting and methodological inconsistencies, including nonstandardised reference standards, limit regulatory confidence and hinder investment. Lastly, few biomarker platforms have been integrated into user-friendly diagnostic formats suitable for point-of-care use, further delaying the transition to clinical trial evaluation.</p><p>Protein-based signatures and cytokine/chemokine profiles offer potential advantages, such as lower cost and easier scalability for decentralised applications. In theory, these tests could support decentralised screening in community or primary care settings, especially in high-burden countries where laboratory infrastructure is limited. However, several barriers remain. Our review identified very few studies that evaluated test performance in primary care/decentralised settings or provided evidence on operational feasibility. Additionally, limited data exist on turnaround time, cost-per-test or integration into existing diagnostic workflows. Most importantly, the trade-offs between sensitivity and specificity observed across studies raise concerns about the risk of overtreatment or missed cases if these tools are implemented prematurely.</p><p>Multianalyte approaches that combine transcriptomic, proteomic and immunological markers may enhance predictive power. However, such strategies increase assay complexity, technical requirements and cost. Without advances in assay simplification and automation, the feasibility of implementing these platforms at the point-of-care level will likely remain aspirational.</p><p>The studies included in our review were highly heterogeneous in terms of design, population and reference standards. This variation precluded meta-analytic synthesis and underscores the need for harmonised approaches to study design and reporting. While transcriptomic, proteomic and immunological biomarkers offer theoretical and early empirical promise, their clinical utility and applicability hinge on validation across diverse settings and populations, particularly among high-risk groups such as people living with HIV, young children and close contacts of individuals with active TB.</p><p>Moreover, methodological limitations persist. Many studies exhibited unclear risk of bias due to incomplete reporting on sampling methods, reference standards or participant follow-up. This lack of transparency is not uncommon in diagnostic research [<xref rid="C16" ref-type="bibr">16</xref>] and compromises the translational potential of findings related to biomarker-based tests [<xref rid="C22" ref-type="bibr">22</xref>].</p><p>A major strength of this review is the notably sensitive search strategy used and its comprehensive analysis of biomarker tests across multiple domains, highlighting the potential of various biomarker-based approaches for predicting TB progression. However, several important limitations were identified, particularly concerning the transparency of reporting within the included studies. Many studies received “unclear” ratings for risk of bias and applicability, primarily due to insufficient details regarding patient selection, recruitment methods and the flow of participants through the study. This lack of transparency not only complicates the assessment of methodological quality but also limits the external validity of the findings, making it challenging to generalise results to broader populations. Furthermore, the observed heterogeneity in study designs and outcomes underscores the pressing need for standardised reporting practices in TB biomarker research, which would enhance the interpretability and applicability of findings.</p><p>Future research of promising biomarkers must prioritise large-scale, prospective validation studies using standardised protocols to evaluate biomarkers based on WHO-endorsed TPPs or similar criteria intended to determine the usefulness of such tests. Biomarker development should align with implementation science frameworks to assess feasibility, cost-effectiveness and scalability in high-burden settings [<xref rid="C15" ref-type="bibr">15</xref>]. Without such methodological and operational rigor, the potential for biomarkers to guide targeted TB preventive therapy and reduce disease incidence will remain unrealised.</p><p>Ultimately, our findings show that while the concept of disease progression is scientifically sound, the path to clinical adoption requires more robust evidence. Cross-disciplinary collaboration, global investment and the use of standardised evaluation frameworks will be essential to translate the promise of TB biomarkers into practice.</p><boxed-text position="anchor" orientation="portrait"><sec id="s6"><title>Points for clinical practice</title><list list-type="bullet"><list-item><p>Current biomarker-based tests for predicting progression from TBI to active disease show promise but are not yet ready for clinical implementation.</p></list-item><list-item><p>Most evaluated tests demonstrate only moderate predictive accuracy and lack validation in diverse, real-world populations.</p></list-item><list-item><p>Existing immunological tests (<italic toggle="yes">e.g.</italic> TSTs and IGRAs) have low PPV, underscoring the need for more accurate tools to guide targeted preventive therapy.</p></list-item><list-item><p>Clinicians should remain cautious in interpreting biomarker test results for TB progression until standardised, validated and cost-effective tools become available.</p></list-item><list-item><p>Integration of biomarker-based tests into TB prevention programmes will require careful consideration of feasibility, cost and operational logistics, especially in high-burden settings.</p></list-item></list></sec></boxed-text><boxed-text position="anchor" orientation="portrait"><sec id="s7"><title>Questions for future research</title><list list-type="bullet"><list-item><p>Which biomarker signatures (<italic toggle="yes">e.g.</italic> transcriptomic, proteomic, cytokine-based) offer the best balance of sensitivity, specificity and feasibility for predicting TB progression?</p></list-item><list-item><p>How can future studies improve methodological rigor and transparency to enhance reproducibility and comparability of biomarker performance?</p></list-item><list-item><p>What are the optimal time windows for prediction and how does predictive accuracy vary across different populations and epidemiological contexts?</p></list-item><list-item><p>Can multianalyte or combined biomarker approaches improve predictive performance without compromising scalability or affordability?</p></list-item><list-item><p>What are the cost-effectiveness and implementation challenges of deploying biomarker-based tests in resource-limited settings?</p></list-item></list></sec></boxed-text></sec></body><back><ack><title>Acknowledgements</title><p>We thank Kaitryn Campbell, MLIS, MSc for peer review of the Medline search strategy. We are also very grateful to Dennis Falzon (Global Tuberculosis Programme, World Health Organization) for providing us with important insights that helped data synthesis and interpretation. We acknowledge the use of ChatGPT version 4o (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://chat.openai.com/" ext-link-type="uri">https://chat.openai.com/</ext-link>) to refine the academic language of our work and support the formatting of the included results tables. The tool was not used to generate any content but solely for formatting and proofreading purposes.</p></ack><fn-group><fn fn-type="other"><p>Provenance: Submitted article, peer reviewed.</p></fn><fn fn-type="other"><p>The systematic review protocol was registered with PROSPERO (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://www.crd.york.ac.uk/prospero/" ext-link-type="uri">https://www.crd.york.ac.uk/prospero/</ext-link>) with identifier: CRD42024554184.</p></fn><fn fn-type="COI-statement"><p>Conflict of interest: K. Tingley has nothing to disclose. A. Li reports support for the present study from University of Ottawa Faculty of Medicine Medical Student Summer Research Programme. E.L. MacLean reports grants from Canadian Institutes of Health Research. B. Skidmore reports support for the present study from the project lead (independent consulting services). N. Maredia has nothing to disclose. R. Ravikumar has nothing to disclose. G.G. Alvarez reports grants from Canadian Institutes of Health Research. A. Matteelli has nothing to disclose. G. Sulis reports support for the present study from the New Diagnostics Working Group (NDWG; NDWG is a multi-sector group that aims to progress TB diagnostics, it receives funding from the Stop TB Partnership and United States Agency for International Development (USAID)), and grants from Canadian Institutes of Health Research, Canadian Immunization Research Network, J. P. Bickell Foundation and Canada Research Chairs Programme.</p></fn><fn fn-type="financial-disclosure"><p>Support statement: The current project was initiated by the Tuberculosis Infection and Test of Progression Task Force of the New Diagnostics Working Group (NDWG). NDWG is a multi-sector group that aims to progress TB diagnostics; NDWG receives funding from the Stop TB Partnership and United States Agency for International Development (USAID). G. Sulis holds a Tier 2 Canada Research Chair in Communicable Disease Epidemiology (https://www.chairs-chaires.gc.ca/home-accueil-eng.aspx). E.L. MacLean is funded by a CIHR Fellowship (472823). The funders had no role in the conceptualisation, design, conduct, decision to publish or preparation of this manuscript. Funding information for this article has been deposited with the <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://www.crossref.org/services/funder-registry/" ext-link-type="uri">Open Funder Registry</ext-link>.</p></fn></fn-group><sec sec-type="supplementary-material" id="s8"><title>Supplementary material</title><supplementary-material id="supplementary" position="float" content-type="local-data" orientation="portrait"><object-id pub-id-type="doi">10.1183/16000617.0151-2025.Supp1</object-id><p><bold>Please note:</bold> supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.</p><p>Supplementary material <fig id="SS1" position="anchor" orientation="portrait"><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="ERR-0151-2025.SUPPLEMENT.pdf" position="float" orientation="portrait"/></fig>ERR-0151-2025.SUPPLEMENT</p></supplementary-material></sec><ref-list><title>References</title><ref id="C1"><label>1</label><mixed-citation publication-type="book"><person-group person-group-type="author"><collab>World Health Organization</collab></person-group>. <source>Global Tuberculosis Report 2025</source>. <publisher-loc>Geneva</publisher-loc>, <publisher-name>World Health Organization</publisher-name>, <year>2025</year>.</mixed-citation></ref><ref id="C2"><label>2</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Houben</surname>
<given-names>RMGJ</given-names></string-name>, <string-name name-style="western"><surname>Dodd</surname>
<given-names>PJ</given-names></string-name></person-group>. <article-title>The global burden of latent tuberculosis infection: a re-estimation using mathematical modelling</article-title>. <source>PLoS Med</source>
<year>2016</year>; <volume>13</volume>: <fpage>e1002152</fpage>. doi:<pub-id pub-id-type="doi">10.1371/journal.pmed.1002152</pub-id><pub-id pub-id-type="pmid">27780211</pub-id>
<pub-id pub-id-type="pmcid">PMC5079585</pub-id></mixed-citation></ref><ref id="C3"><label>3</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Vynnycky</surname>
<given-names>E</given-names></string-name>, <string-name name-style="western"><surname>Fine</surname>
<given-names>PEM</given-names></string-name></person-group>. <article-title>The natural history of tuberculosis: the implications of age-dependent risks of disease and the role of reinfection</article-title>. <source>Epidemiol Infect</source>
<year>1997</year>; <volume>119</volume>: <fpage>183</fpage>–<lpage>201</lpage>. doi:<pub-id pub-id-type="doi">10.1017/S0950268897007917</pub-id><pub-id pub-id-type="pmid">9363017</pub-id>
<pub-id pub-id-type="pmcid">PMC2808840</pub-id></mixed-citation></ref><ref id="C4"><label>4</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Pai</surname>
<given-names>M</given-names></string-name>, <string-name name-style="western"><surname>Behr</surname>
<given-names>MA</given-names></string-name>, <string-name name-style="western"><surname>Dowdy</surname>
<given-names>D</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Tuberculosis</article-title>. <source>Nat Rev Dis Primers</source>
<year>2016</year>; <volume>2</volume>: <fpage>16076</fpage>. doi:<pub-id pub-id-type="doi">10.1038/nrdp.2016.76</pub-id><pub-id pub-id-type="pmid">27784885</pub-id>
</mixed-citation></ref><ref id="C5"><label>5</label><mixed-citation publication-type="book"><person-group person-group-type="author"><collab>World Health Organization</collab></person-group>. <article-title>Framework for implementing the “End TB Strategy” in the African Region 2016–2020</article-title>. <publisher-loc>Geneva</publisher-loc>, <publisher-name>World Health Organization</publisher-name>, <year>2017</year>.</mixed-citation></ref><ref id="C6"><label>6</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Uplekar</surname>
<given-names>M</given-names></string-name>, <string-name name-style="western"><surname>Weil</surname>
<given-names>D</given-names></string-name>, <string-name name-style="western"><surname>Lonnroth</surname>
<given-names>K</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>WHO's new end TB strategy</article-title>. <source>Lancet</source>
<year>2015</year>; <volume>385</volume>: <fpage>1799</fpage>–<lpage>1801</lpage>. doi:<pub-id pub-id-type="doi">10.1016/S0140-6736(15)60570-0</pub-id><pub-id pub-id-type="pmid">25814376</pub-id>
</mixed-citation></ref><ref id="C7"><label>7</label><mixed-citation publication-type="book"><person-group person-group-type="author"><collab>World Health Organization</collab></person-group>. <source>WHO Operational Handbook on Tuberculosis. Module 1: Prevention: Tuberculosis Preventive Treatment</source>. <edition>2nd Edn</edition>. <publisher-loc>Geneva</publisher-loc>, <publisher-name>World Health Organization</publisher-name>, <year>2024</year>.</mixed-citation></ref><ref id="C8"><label>8</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Zellweger</surname>
<given-names>JP</given-names></string-name>, <string-name name-style="western"><surname>Sotgiu</surname>
<given-names>G</given-names></string-name>, <string-name name-style="western"><surname>Corradi</surname>
<given-names>M</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>The diagnosis of latent tuberculosis infection (LTBI): currently available tests, future developments, and perspectives to eliminate tuberculosis (TB)</article-title>. <source>Med Lav</source>
<year>2020</year>; <volume>111</volume>: <fpage>170</fpage>–<lpage>183</lpage>. doi:<pub-id pub-id-type="doi">10.23749/mdl.v111i3.9983</pub-id><pub-id pub-id-type="pmid">32624559</pub-id>
<pub-id pub-id-type="pmcid">PMC7809945</pub-id></mixed-citation></ref><ref id="C9"><label>9</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Melnychuk</surname>
<given-names>L</given-names></string-name>, <string-name name-style="western"><surname>Perlman-Arrow</surname>
<given-names>S</given-names></string-name>, <string-name name-style="western"><surname>Lisboa Bastos</surname>
<given-names>M</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>A systematic review and meta-analysis of tuberculous preventative therapy adverse events</article-title>. <source>Clin Infect Dis</source>
<year>2023</year>; <volume>77</volume>: <fpage>287</fpage>–<lpage>294</lpage>. doi:<pub-id pub-id-type="doi">10.1093/cid/ciad246</pub-id><pub-id pub-id-type="pmid">37125482</pub-id>
</mixed-citation></ref><ref id="C10"><label>10</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Matteelli</surname>
<given-names>A</given-names></string-name>, <string-name name-style="western"><surname>Lovatti</surname>
<given-names>S</given-names></string-name>, <string-name name-style="western"><surname>Sforza</surname>
<given-names>A</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Programmatic management of tuberculosis preventive therapy: past, present, future</article-title>. <source>Int J Infect Dis</source>
<year>2023</year>; <volume>130</volume>: Suppl. 1, <fpage>S43</fpage>–<lpage>S46</lpage>. doi:<pub-id pub-id-type="doi">10.1016/j.ijid.2023.02.016</pub-id><pub-id pub-id-type="pmid">36889577</pub-id>
<pub-id pub-id-type="pmcid">PMC10186937</pub-id></mixed-citation></ref><ref id="C11"><label>11</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>James</surname>
<given-names>R</given-names></string-name>, <string-name name-style="western"><surname>Theron</surname>
<given-names>G</given-names></string-name>, <string-name name-style="western"><surname>Cobelens</surname>
<given-names>F</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Framing the detection of incipient tuberculosis infection: a qualitative study of political prioritisation</article-title>. <source>Trop Med Int Health</source>
<year>2022</year>; <volume>27</volume>: <fpage>445</fpage>–<lpage>453</lpage>. doi:<pub-id pub-id-type="doi">10.1111/tmi.13734</pub-id><pub-id pub-id-type="pmid">35156273</pub-id>
<pub-id pub-id-type="pmcid">PMC9306665</pub-id></mixed-citation></ref><ref id="C12"><label>12</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Bodaghi</surname>
<given-names>A</given-names></string-name>, <string-name name-style="western"><surname>Fattahi</surname>
<given-names>N</given-names></string-name>, <string-name name-style="western"><surname>Ramazani</surname>
<given-names>A</given-names></string-name></person-group>. <article-title>Biomarkers: promising and valuable tools towards diagnosis, prognosis and treatment of COVID-19 and other diseases</article-title>. <source>Heliyon</source>
<year>2023</year>; <volume>9</volume>: <fpage>e13323</fpage>. doi:<pub-id pub-id-type="doi">10.1016/j.heliyon.2023.e13323</pub-id><pub-id pub-id-type="pmid">36744065</pub-id>
<pub-id pub-id-type="pmcid">PMC9884646</pub-id></mixed-citation></ref><ref id="C13"><label>13</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Nogueira</surname>
<given-names>BMF</given-names></string-name>, <string-name name-style="western"><surname>Krishnan</surname>
<given-names>S</given-names></string-name>, <string-name name-style="western"><surname>Barreto-Duarte</surname>
<given-names>B</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Diagnostic biomarkers for active tuberculosis: progress and challenges</article-title>. <source>EMBO Mol Med</source>
<year>2022</year>; <volume>14</volume>: <fpage>e14088</fpage>. doi:<pub-id pub-id-type="doi">10.15252/emmm.202114088</pub-id><pub-id pub-id-type="pmid">36314872</pub-id>
<pub-id pub-id-type="pmcid">PMC9728055</pub-id></mixed-citation></ref><ref id="C14"><label>14</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Zak</surname>
<given-names>DE</given-names></string-name>, <string-name name-style="western"><surname>Penn-Nicholson</surname>
<given-names>A</given-names></string-name>, <string-name name-style="western"><surname>Scriba</surname>
<given-names>TJ</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>A blood RNA signature for tuberculosis disease risk: a prospective cohort study</article-title>. <source>Lancet</source>
<year>2016</year>; <volume>387</volume>: <fpage>2312</fpage>–<lpage>2322</lpage>. doi:<pub-id pub-id-type="doi">10.1016/S0140-6736(15)01316-1</pub-id><pub-id pub-id-type="pmid">27017310</pub-id>
<pub-id pub-id-type="pmcid">PMC5392204</pub-id></mixed-citation></ref><ref id="C15"><label>15</label><mixed-citation publication-type="book"><person-group person-group-type="author"><collab>World Health Organization</collab></person-group>. <source>Consensus Meeting Report: Development of a Target Product Profile (TPP) and a Framework for Evaluation for a Test for Predicting Progression from Tuberculosis Infection to Active Disease</source>. <publisher-loc>Geneva</publisher-loc>, <publisher-name>World Health Organization</publisher-name>, <year>2017</year>.</mixed-citation></ref><ref id="C16"><label>16</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>MacLean</surname>
<given-names>E</given-names></string-name>, <string-name name-style="western"><surname>Broger</surname>
<given-names>T</given-names></string-name>, <string-name name-style="western"><surname>Yerlikaya</surname>
<given-names>S</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>A systematic review of biomarkers to detect active tuberculosis</article-title>. <source>Nat Microbiol</source>
<year>2019</year>; <volume>4</volume>: <fpage>748</fpage>–<lpage>758</lpage>. doi:<pub-id pub-id-type="doi">10.1038/s41564-019-0380-2</pub-id><pub-id pub-id-type="pmid">30804546</pub-id>
</mixed-citation></ref><ref id="C17"><label>17</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Mulenga</surname>
<given-names>H</given-names></string-name>, <string-name name-style="western"><surname>Zauchenberger</surname>
<given-names>CZ</given-names></string-name>, <string-name name-style="western"><surname>Bunyasi</surname>
<given-names>EW</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Performance of diagnostic and predictive host blood transcriptomic signatures for tuberculosis disease: a systematic review and meta-analysis</article-title>. <source>PLoS One</source>
<year>2020</year>; <volume>15</volume>: <fpage>e0237574</fpage>. doi:<pub-id pub-id-type="doi">10.1371/journal.pone.0237574</pub-id><pub-id pub-id-type="pmid">32822359</pub-id>
<pub-id pub-id-type="pmcid">PMC7442252</pub-id></mixed-citation></ref><ref id="C18"><label>18</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Page</surname>
<given-names>MJ</given-names></string-name>, <string-name name-style="western"><surname>McKenzie</surname>
<given-names>JE</given-names></string-name>, <string-name name-style="western"><surname>Bossuyt</surname>
<given-names>PM</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>The PRISMA 2020 statement: an updated guideline for reporting systematic reviews</article-title>. <source>BMJ</source>
<year>2021</year>; <volume>372</volume>: <fpage>n71</fpage>. doi:<pub-id pub-id-type="doi">10.1136/bmj.n71</pub-id><pub-id pub-id-type="pmid">33782057</pub-id>
<pub-id pub-id-type="pmcid">PMC8005924</pub-id></mixed-citation></ref><ref id="C19"><label>19</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>McGowan</surname>
<given-names>J</given-names></string-name>, <string-name name-style="western"><surname>Sampson</surname>
<given-names>M</given-names></string-name>, <string-name name-style="western"><surname>Salzwedel</surname>
<given-names>DM</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>PRESS peer review of electronic search strategies: 2015 guideline statement</article-title>. <source>J Clin Epidemiol</source>
<year>2016</year>; <volume>75</volume>: <fpage>40</fpage>–<lpage>46</lpage>. doi:<pub-id pub-id-type="doi">10.1016/j.jclinepi.2016.01.021</pub-id><pub-id pub-id-type="pmid">27005575</pub-id>
</mixed-citation></ref><ref id="C20"><label>20</label><mixed-citation publication-type="standard"><person-group person-group-type="author"><collab>CDA-AMC</collab></person-group>. <article-title>Grey matters: a practical tool for searching health-related grey literature</article-title>. <comment>Date last accessed 21 May 2024. Date last updated: April 2019</comment>. <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="www.cadth.ca/node/88098" ext-link-type="uri">www.cadth.ca/node/88098</ext-link></mixed-citation></ref><ref id="C21"><label>21</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Khangura</surname>
<given-names>S</given-names></string-name>, <string-name name-style="western"><surname>Konnyu</surname>
<given-names>K</given-names></string-name>, <string-name name-style="western"><surname>Cushman</surname>
<given-names>R</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Evidence summaries: the evolution of a rapid review approach</article-title>. <source>Syst Rev</source>
<year>2012</year>; <volume>1</volume>: <fpage>10</fpage>. doi:<pub-id pub-id-type="doi">10.1186/2046-4053-1-10</pub-id><pub-id pub-id-type="pmid">22587960</pub-id>
<pub-id pub-id-type="pmcid">PMC3351736</pub-id></mixed-citation></ref><ref id="C22"><label>22</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Whiting</surname>
<given-names>PF</given-names></string-name>, <string-name name-style="western"><surname>Rutjes</surname>
<given-names>AWS</given-names></string-name>, <string-name name-style="western"><surname>Westwood</surname>
<given-names>ME</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies</article-title>. <source>Ann Intern Med</source>
<year>2011</year>; <volume>155</volume>: <fpage>529</fpage>–<lpage>536</lpage>. doi:<pub-id pub-id-type="doi">10.7326/0003-4819-155-8-201110180-00009</pub-id><pub-id pub-id-type="pmid">22007046</pub-id>
</mixed-citation></ref><ref id="C23"><label>23</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Land</surname>
<given-names>KJ</given-names></string-name>, <string-name name-style="western"><surname>Boeras</surname>
<given-names>DI</given-names></string-name>, <string-name name-style="western"><surname>Chen</surname>
<given-names>XS</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>REASSURED diagnostics to inform disease control strategies, strengthen health systems and improve patient outcomes</article-title>. <source>Nat Microbiol</source>
<year>2019</year>; <volume>4</volume>: <fpage>46</fpage>–<lpage>54</lpage>. doi:<pub-id pub-id-type="doi">10.1038/s41564-018-0295-3</pub-id><pub-id pub-id-type="pmid">30546093</pub-id>
<pub-id pub-id-type="pmcid">PMC7097043</pub-id></mixed-citation></ref><ref id="C24"><label>24</label><mixed-citation publication-type="book"><person-group person-group-type="author"><collab>World Health Organization</collab></person-group>. <source>High-priority Target Product Profiles for New Tuberculosis Diagnostics: Report of a Consensus Meeting</source>. <publisher-loc>Geneva</publisher-loc>, <publisher-name>World Health Organization</publisher-name>, <year>2014</year>.</mixed-citation></ref><ref id="C25"><label>25</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Wang</surname>
<given-names>X</given-names></string-name>, <string-name name-style="western"><surname>VanValkenberg</surname>
<given-names>A</given-names></string-name>, <string-name name-style="western"><surname>Odom</surname>
<given-names>AR</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Comparison of gene set scoring methods for reproducible evaluation of tuberculosis gene signatures</article-title>. <source>BMC Infect Dis</source>
<year>2024</year>; <volume>24</volume>: <fpage>610</fpage>. doi:<pub-id pub-id-type="doi">10.1186/s12879-024-09457-z</pub-id><pub-id pub-id-type="pmid">38902649</pub-id>
<pub-id pub-id-type="pmcid">PMC11191245</pub-id></mixed-citation></ref><ref id="C26"><label>26</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Warsinske</surname>
<given-names>H</given-names></string-name>, <string-name name-style="western"><surname>Vashisht</surname>
<given-names>R</given-names></string-name>, <string-name name-style="western"><surname>Khatri</surname>
<given-names>P</given-names></string-name></person-group>. <article-title>Host-response-based gene signatures for tuberculosis diagnosis: a systematic comparison of 16 signatures</article-title>. <source>PLoS Med</source>
<year>2019</year>; <volume>16</volume>: <fpage>e1002786</fpage>. doi:<pub-id pub-id-type="doi">10.1371/journal.pmed.1002786</pub-id><pub-id pub-id-type="pmid">31013272</pub-id>
<pub-id pub-id-type="pmcid">PMC6478271</pub-id></mixed-citation></ref><ref id="C27"><label>27</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Vargas</surname>
<given-names>R</given-names></string-name>, <string-name name-style="western"><surname>Abbott</surname>
<given-names>L</given-names></string-name>, <string-name name-style="western"><surname>Bower</surname>
<given-names>D</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Gene signature discovery and systematic validation across diverse clinical cohorts for TB prognosis and response to treatment</article-title>. <source>PLoS Comput Biol</source>
<year>2023</year>; <volume>19</volume>: <fpage>e1010770</fpage>. doi:<pub-id pub-id-type="doi">10.1371/journal.pcbi.1010770</pub-id><pub-id pub-id-type="pmid">37471455</pub-id>
<pub-id pub-id-type="pmcid">PMC10393163</pub-id></mixed-citation></ref><ref id="C28"><label>28</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Xin</surname>
<given-names>H</given-names></string-name>, <string-name name-style="western"><surname>Cao</surname>
<given-names>X</given-names></string-name>, <string-name name-style="western"><surname>Du</surname>
<given-names>Y</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>The association between circulating microRNAs and the risk of active disease development from latent tuberculosis infection: a nested case-control study</article-title>. <source>Microbiol Spectr</source>
<year>2022</year>; <volume>10</volume>: <fpage>e0262521</fpage>. doi:<pub-id pub-id-type="doi">10.1128/spectrum.02625-21</pub-id><pub-id pub-id-type="pmid">35435753</pub-id>
<pub-id pub-id-type="pmcid">PMC9241859</pub-id></mixed-citation></ref><ref id="C29"><label>29</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Long</surname>
<given-names>NP</given-names></string-name>, <string-name name-style="western"><surname>Phat</surname>
<given-names>NK</given-names></string-name>, <string-name name-style="western"><surname>Yen</surname>
<given-names>NTH</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>A 10-gene biosignature of tuberculosis treatment monitoring and treatment outcome prediction</article-title>. <source>Tuberculosis</source>
<year>2021</year>; <volume>131</volume>: <fpage>102138</fpage>. doi:<pub-id pub-id-type="doi">10.1016/j.tube.2021.102138</pub-id><pub-id pub-id-type="pmid">34801869</pub-id>
</mixed-citation></ref><ref id="C30"><label>30</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Roe</surname>
<given-names>J</given-names></string-name>, <string-name name-style="western"><surname>Venturini</surname>
<given-names>C</given-names></string-name>, <string-name name-style="western"><surname>Gupta</surname>
<given-names>RK</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Blood transcriptomic stratification of short-term risk in contacts of tuberculosis</article-title>. <source>Clin Infect Dis</source>
<year>2020</year>; <volume>70</volume>: <fpage>731</fpage>–<lpage>737</lpage>. doi:<pub-id pub-id-type="doi">10.1093/cid/ciz252</pub-id><pub-id pub-id-type="pmid">30919880</pub-id>
<pub-id pub-id-type="pmcid">PMC7617764</pub-id></mixed-citation></ref><ref id="C31"><label>31</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Penn-Nicholson</surname>
<given-names>A</given-names></string-name>, <string-name name-style="western"><surname>Mbandi</surname>
<given-names>SK</given-names></string-name>, <string-name name-style="western"><surname>Thompson</surname>
<given-names>E</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>RISK6, a 6-gene transcriptomic signature of TB disease risk, diagnosis and treatment response</article-title>. <source>Sci Rep</source>
<year>2020</year>; <volume>10</volume>: <fpage>8629</fpage>. doi:<pub-id pub-id-type="doi">10.1038/s41598-020-65043-8</pub-id><pub-id pub-id-type="pmid">32451443</pub-id>
<pub-id pub-id-type="pmcid">PMC7248089</pub-id></mixed-citation></ref><ref id="C32"><label>32</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Leong</surname>
<given-names>S</given-names></string-name>, <string-name name-style="western"><surname>Zhao</surname>
<given-names>Y</given-names></string-name>, <string-name name-style="western"><surname>Ribeiro-Rodrigues</surname>
<given-names>R</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Cross-validation of existing signatures and derivation of a novel 29-gene transcriptomic signature predictive of progression to TB in a Brazilian cohort of household contacts of pulmonary TB</article-title>. <source>Tuberculosis</source>
<year>2020</year>; <volume>120</volume>: <fpage>101898</fpage>. doi:<pub-id pub-id-type="doi">10.1016/j.tube.2020.101898</pub-id><pub-id pub-id-type="pmid">32090859</pub-id>
<pub-id pub-id-type="pmcid">PMC7066850</pub-id></mixed-citation></ref><ref id="C33"><label>33</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Warsinske</surname>
<given-names>HC</given-names></string-name>, <string-name name-style="western"><surname>Rao</surname>
<given-names>AM</given-names></string-name>, <string-name name-style="western"><surname>Moreira</surname>
<given-names>FMF</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Assessment of validity of a blood-based 3-gene signature score for progression and diagnosis of tuberculosis, disease severity, and treatment response</article-title>. <source>JAMA Netw Open</source>
<year>2018</year>; <volume>1</volume>: <fpage>e183779</fpage>. doi:<pub-id pub-id-type="doi">10.1001/jamanetworkopen.2018.3779</pub-id><pub-id pub-id-type="pmid">30646264</pub-id>
<pub-id pub-id-type="pmcid">PMC6324428</pub-id></mixed-citation></ref><ref id="C34"><label>34</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Gupta</surname>
<given-names>S</given-names></string-name>, <string-name name-style="western"><surname>Rao</surname>
<given-names>A</given-names></string-name>, <string-name name-style="western"><surname>Scott</surname>
<given-names>M</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>A nine-gene blood-based signature meets the World Health Organization target product profiles for diagnosis of active tuberculosis and predicting progression from latent to active disease</article-title>. <source>Open Forum Infect Dis</source>
<year>2022</year>; <volume>9</volume>: <fpage>S56</fpage>–<lpage>S56</lpage>. doi:<pub-id pub-id-type="doi">10.1093/ofid/ofac492.094</pub-id></mixed-citation></ref><ref id="C35"><label>35</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>de Araujo</surname>
<given-names>LS</given-names></string-name>, <string-name name-style="western"><surname>Ribeiro-alves</surname>
<given-names>M</given-names></string-name>, <string-name name-style="western"><surname>Wipperman</surname>
<given-names>MF</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Transcriptomic biomarkers for tuberculosis: validation of npc2 as a single mRNA biomarker to diagnose tb, predict disease progression, and monitor treatment response</article-title>. <source>Cells</source>
<year>2021</year>; <volume>10</volume>: <fpage>2704</fpage>. doi:<pub-id pub-id-type="doi">10.3390/cells10102704</pub-id><pub-id pub-id-type="pmid">34685683</pub-id>
<pub-id pub-id-type="pmcid">PMC8534371</pub-id></mixed-citation></ref><ref id="C36"><label>36</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Burel</surname>
<given-names>JG</given-names></string-name>, <string-name name-style="western"><surname>Wang</surname>
<given-names>W</given-names></string-name>, <string-name name-style="western"><surname>Wuhrer</surname>
<given-names>M</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>IgG glycosylation associates with risk of progression from latent to active tuberculosis</article-title>. <source>J Infect</source>
<year>2024</year>; <volume>88</volume>: <fpage>106115</fpage>. doi:<pub-id pub-id-type="doi">10.1016/j.jinf.2024.01.014</pub-id><pub-id pub-id-type="pmid">38309308</pub-id>
</mixed-citation></ref><ref id="C37"><label>37</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Araujo</surname>
<given-names>Z</given-names></string-name>, <string-name name-style="western"><surname>Macias-Segura</surname>
<given-names>N</given-names></string-name>, <string-name name-style="western"><surname>López-Ramos</surname>
<given-names>JE</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Diagnostic accuracy of combinations of serological biomarkers for identifying clinical tuberculosis</article-title>. <source>J Infect Dev Ctries</source>
<year>2018</year>; <volume>12</volume>: <fpage>429</fpage>–<lpage>441</lpage>. doi:<pub-id pub-id-type="doi">10.3855/jidc.9554</pub-id><pub-id pub-id-type="pmid">31940294</pub-id>
</mixed-citation></ref><ref id="C38"><label>38</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Penn-Nicholson</surname>
<given-names>A</given-names></string-name>, <string-name name-style="western"><surname>Hraha</surname>
<given-names>T</given-names></string-name>, <string-name name-style="western"><surname>Thompson</surname>
<given-names>EG</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Discovery and validation of a prognostic proteomic signature for tuberculosis progression: a prospective cohort study</article-title>. <source>PLoS Med</source>
<year>2019</year>; <volume>16</volume>: <fpage>e1002781</fpage>. doi:<pub-id pub-id-type="doi">10.1371/journal.pmed.1002781</pub-id><pub-id pub-id-type="pmid">30990820</pub-id>
<pub-id pub-id-type="pmcid">PMC6467365</pub-id></mixed-citation></ref><ref id="C39"><label>39</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Singer</surname>
<given-names>SN</given-names></string-name>, <string-name name-style="western"><surname>Ndumnego</surname>
<given-names>OC</given-names></string-name>, <string-name name-style="western"><surname>Kim</surname>
<given-names>RS</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Plasma host protein biomarkers correlating with increasing <italic toggle="yes">Mycobacterium tuberculosis</italic> infection activity prior to tuberculosis diagnosis in people living with HIV</article-title>. <source>EBioMedicine</source>
<year>2022</year>; <volume>75</volume>: <fpage>103787</fpage>. doi:<pub-id pub-id-type="doi">10.1016/j.ebiom.2021.103787</pub-id><pub-id pub-id-type="pmid">34968761</pub-id>
<pub-id pub-id-type="pmcid">PMC8718743</pub-id></mixed-citation></ref><ref id="C40"><label>40</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Daniel</surname>
<given-names>EA</given-names></string-name>, <string-name name-style="western"><surname>Thiruvengadam</surname>
<given-names>K</given-names></string-name>, <string-name name-style="western"><surname>Rajamanickam</surname>
<given-names>A</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>QuantiFERON supernatant-based host biomarkers predicting progression to active tuberculosis disease among household contacts of tuberculosis patients</article-title>. <source>Clin Infect Dis</source>
<year>2023</year>; <volume>76</volume>: <fpage>1802</fpage>–<lpage>1813</lpage>. doi:<pub-id pub-id-type="doi">10.1093/cid/ciac979</pub-id><pub-id pub-id-type="pmid">36582115</pub-id>
</mixed-citation></ref><ref id="C41"><label>41</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Robison</surname>
<given-names>HM</given-names></string-name>, <string-name name-style="western"><surname>Chapman</surname>
<given-names>CA</given-names></string-name>, <string-name name-style="western"><surname>Zhou</surname>
<given-names>H</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Risk assessment of latent tuberculosis infection through a multiplexed cytokine biosensor assay and machine learning feature selection</article-title>. <source>Sci Rep</source>
<year>2021</year>; <volume>11</volume>: <fpage>20544</fpage>. doi:<pub-id pub-id-type="doi">10.1038/s41598-021-99754-3</pub-id><pub-id pub-id-type="pmid">34654869</pub-id>
<pub-id pub-id-type="pmcid">PMC8520014</pub-id></mixed-citation></ref><ref id="C42"><label>42</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Zhang</surname>
<given-names>H</given-names></string-name>, <string-name name-style="western"><surname>Cao</surname>
<given-names>X</given-names></string-name>, <string-name name-style="western"><surname>Xin</surname>
<given-names>H</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Serum level of IL-1ra was associated with the treatment of latent tuberculosis infection in a Chinese population</article-title>. <source>BMC Infect Dis</source>
<year>2020</year>; <volume>20</volume>: <fpage>330</fpage>. doi:<pub-id pub-id-type="doi">10.1186/s12879-020-05047-x</pub-id><pub-id pub-id-type="pmid">32384874</pub-id>
<pub-id pub-id-type="pmcid">PMC7206663</pub-id></mixed-citation></ref><ref id="C43"><label>43</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Esmail</surname>
<given-names>H</given-names></string-name>, <string-name name-style="western"><surname>Miller</surname>
<given-names>C</given-names></string-name>, <string-name name-style="western"><surname>Falzon</surname>
<given-names>D</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Scaling-up symptom-agnostic, community-wide screening toward global tuberculosis elimination: opportunities, challenges, and lessons from history</article-title>. <source>Int J Infect Dis</source>
<year>2025</year>; <volume>155</volume>: <fpage>107875</fpage>. doi:<pub-id pub-id-type="doi">10.1016/j.ijid.2025.107875</pub-id><pub-id pub-id-type="pmid">40068708</pub-id>
</mixed-citation></ref><ref id="C44"><label>44</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><surname>Suliman</surname>
<given-names>S</given-names></string-name>, <string-name name-style="western"><surname>Thompson</surname>
<given-names>EG</given-names></string-name>, <string-name name-style="western"><surname>Sutherland</surname>
<given-names>J</given-names></string-name>, <etal>et al.</etal></person-group>
<article-title>Four-gene pan-African blood signature predicts progression to tuberculosis</article-title>. <source>Am J Respir Crit Care Med</source>
<year>2018</year>; <volume>197</volume>: <fpage>1198</fpage>–<lpage>1208</lpage>. doi:<pub-id pub-id-type="doi">10.1164/rccm.201711-2340OC</pub-id><pub-id pub-id-type="pmid">29624071</pub-id>
<pub-id pub-id-type="pmcid">PMC6019933</pub-id></mixed-citation></ref></ref-list></back></article><article xml:lang="en" article-type="research-article" dtd-version="1.4"><processing-meta base-tagset="archiving" mathml-version="3.0" table-model="xhtml" tagset-family="jats"><restricted-by>pmc</restricted-by></processing-meta><front><journal-meta><journal-id journal-id-type="nlm-ta">JMIR Res Protoc</journal-id><journal-id journal-id-type="iso-abbrev">JMIR Res Protoc</journal-id><journal-id journal-id-type="pmc-domain-id">2066</journal-id><journal-id journal-id-type="pmc-domain">resprot</journal-id><journal-id journal-id-type="publisher-id">ResProt</journal-id><journal-title-group><journal-title>JMIR Research Protocols</journal-title></journal-title-group><issn pub-type="epub">1929-0748</issn><publisher><publisher-name>JMIR Publications Inc.</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="pmcid">PMC13016435</article-id><article-id pub-id-type="pmcid-ver">PMC13016435.1</article-id><article-id pub-id-type="pmcaid">13016435</article-id><article-id pub-id-type="pmcaiid">13016435</article-id><article-id pub-id-type="pmid">41880605</article-id><article-id pub-id-type="doi">10.2196/78589</article-id><article-id pub-id-type="publisher-id">78589</article-id><article-version article-version-type="pmc-version">1</article-version><article-categories><subj-group subj-group-type="primary-section"><subject>Non-Randomized Study Protocols and Methods (Non-eHealth)</subject></subj-group><subj-group subj-group-type="secondary-section"><subject>Breast Cancer</subject></subj-group><subj-group subj-group-type="secondary-section"><subject>Chemotherapy and Side Effects Management</subject></subj-group><subj-group subj-group-type="secondary-section"><subject>Adverse Drug Events Detection, Pharmacovigilance and Surveillance</subject></subj-group><subj-group subj-group-type="secondary-section"><subject>Cardiac Risk and Cardiac Risk Calculators</subject></subj-group><subj-group subj-group-type="secondary-section"><subject>Clinical Cancer Research</subject></subj-group><subj-group subj-group-type="heading"><subject>Protocol</subject></subj-group></article-categories><title-group><article-title>Exploring the Impact of Initiating Endocrine Therapy on Metabolic Health in Early Breast Cancer: Protocol for the Prospective Follow-Up EMETA-Study</article-title></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid" authenticated="false">https://orcid.org/0009-0005-5276-7195</contrib-id><name name-style="western"><surname>Nissen</surname><given-names initials="L">Lærke</given-names></name><degrees>BSc</degrees><xref rid="aff1" ref-type="aff">1</xref><xref rid="aff2" ref-type="aff">2</xref><xref rid="cor1" ref-type="corresp"/></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">https://orcid.org/0009-0009-6004-0104</contrib-id><name name-style="western"><surname>Busk Holm</surname><given-names initials="J">Jonas</given-names></name><degrees>MD</degrees><xref rid="aff1" ref-type="aff">1</xref><xref rid="aff2" ref-type="aff">2</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">https://orcid.org/0000-0001-7938-8893</contrib-id><name name-style="western"><surname>Borgquist</surname><given-names initials="S">Signe</given-names></name><degrees>MD, PhD</degrees><xref rid="aff1" ref-type="aff">1</xref><xref rid="aff2" ref-type="aff">2</xref></contrib><aff id="aff1"><label>1</label><institution content-type="department">Department of Clinical Medicine</institution>, <institution>Aarhus University</institution>, <addr-line>Palle Juul-Jensens Boulevard 11</addr-line>, <addr-line content-type="city">Aarhus N</addr-line>, <addr-line>8200</addr-line>, <country>Denmark</country>, <phone>45 91167294</phone></aff><aff id="aff2"><label>2</label><institution content-type="department">Department of Oncology</institution>, <institution>Aarhus University Hospital</institution>, <addr-line content-type="city">Aarhus N</addr-line>, <country>Denmark</country></aff></contrib-group><contrib-group><contrib contrib-type="editor"><name name-style="western"><surname>Schwartz</surname><given-names initials="A">Amy</given-names></name></contrib></contrib-group><author-notes><corresp id="cor1">Lærke Nissen, BSc, Department of Clinical Medicine, Aarhus University, Palle Juul-Jensens Boulevard 11, Aarhus N, 8200, Denmark, <phone>45 91167294</phone>; <email xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="laenis@rm.dk">laenis@rm.dk</email></corresp></author-notes><pub-date pub-type="collection"><year>2026</year></pub-date><pub-date pub-type="epub"><day>25</day><month>3</month><year>2026</year></pub-date><volume>15</volume><issue-id pub-id-type="pmc-issue-id">503669</issue-id><elocation-id>e78589</elocation-id><history><date date-type="received"><day>16</day><month>7</month><year>2025</year></date><date date-type="rev-recd"><day>14</day><month>1</month><year>2026</year></date><date date-type="accepted"><day>28</day><month>1</month><year>2026</year></date></history><pub-history><event event-type="pmc-release"><date><day>25</day><month>03</month><year>2026</year></date></event><event event-type="pmc-live"><date><day>26</day><month>03</month><year>2026</year></date></event><event event-type="pmc-last-change"><date iso-8601-date="2026-03-26 10:25:14.117"><day>26</day><month>03</month><year>2026</year></date></event></pub-history><permissions><copyright-statement>Copyright © Lærke Nissen, Jonas Busk Holm, Signe Borgquist. Originally published in JMIR Research Protocols (https://www.researchprotocols.org)</copyright-statement><copyright-year>2026</copyright-year><license><ali:license_ref xmlns:ali="http://www.niso.org/schemas/ali/1.0/" specific-use="textmining" content-type="ccbylicense">https://creativecommons.org/licenses/by/4.0/</ali:license_ref><license-p>This is an open-access article distributed under the terms of the Creative Commons Attribution License (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">https://creativecommons.org/licenses/by/4.0/</ext-link>), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://www.researchprotocols.org">https://www.researchprotocols.org</ext-link>, as well as this copyright and license information must be included.</license-p></license></permissions><self-uri xmlns:xlink="http://www.w3.org/1999/xlink" content-type="pmc-pdf" xlink:href="resprot-v15-e78589.pdf"/><self-uri xmlns:xlink="http://www.w3.org/1999/xlink" xlink:title="pdf" xlink:href="resprot-v15-e78589.pdf"/><abstract><title>Abstract</title><sec><title>Background</title><p>Adjuvant endocrine therapy is a cornerstone in managing estrogen receptor–positive early breast cancer but may adversely affect metabolic health, including weight gain, insulin resistance, and dyslipidemia. These changes increase the risk of cardiovascular disease and may influence breast cancer outcomes. However, the timing and magnitude of early metabolic changes following endocrine therapy initiation remain poorly characterized. Conventional definitions such as metabolic syndrome rely on dichotomous thresholds and may lack sensitivity to detect early treatment-related metabolic changes, highlighting the need for refined assessment approaches.</p></sec><sec><title>Objective</title><p>This prospective follow-up study aims to investigate early metabolic effects of initiating adjuvant endocrine therapy in women with estrogen receptor–positive early breast cancer and to compare conventional and expanded approaches to metabolic health classification.</p></sec><sec><title>Methods</title><p>This single-center, prospective observational study was conducted at Aarhus University Hospital, Denmark. Women aged≥18 years with early-stage estrogen receptor–positive breast cancer initiating adjuvant endocrine therapy and without pre-existing diabetes were eligible. Metabolic health was assessed at baseline and after 3 months using biometric measurements (weight, waist and hip circumference, waist-to-hip ratio, and blood pressure) and non-fasting blood samples (plasma glucose; hemoglobin A<sub>1c,</sub> (HbA<sub>1c</sub>); lipid profile; and estradiol). The 3-month follow-up was selected to capture early metabolic changes while aligning with routine clinical care to minimize additional visits and reduce selection bias. Metabolic health will be evaluated using two conventional measures and two extended, exploratory measures. Conventional measures are metabolic syndrome (MetS), defined as meeting ≥3 of 5 established criteria (blood pressure ≥130/85 mmHg, triglycerides >2 mmol/l, high-density lipoprotein cholesterol <1.295 mmol/l, waist circumference >88 cm, and plasma glucose >7.8 mmol/l), and the Metabolic Syndrome z score (MetS-Z), a continuous standardized composite of the MetS components. Additional extended measures are exploratory: the extended MetS, which expands the standard MetS definition by incorporating low-density lipoprotein cholesterol (>3 mmol/l), body mass index (≥30 kg/m²), waist-to-hip ratio (>0.85), and HbA<sub>1c</sub> (≥42 mmol/mol), and the EMETA score, a standardized composite of the extended MetS components calculated using the same approach as the MetS-<italic toggle="yes">z</italic> score.</p></sec><sec><title>Results</title><p>The study was funded in July 2024. Recruitment occurred between November 2024 and April 2025, and follow-up was completed in September 2025. Statistical analyses are planned for February 2026, with results expected to be published in summer 2026.</p></sec><sec><title>Conclusions</title><p>This study is expected to provide insights into early metabolic changes following initiation of adjuvant endocrine therapy and evaluate different approaches to classifying metabolic health. The aim to inform future research by helping to identify patients at increased risk of cardiometabolic complications and adverse breast cancer outcomes, warranting confirmation and validation of expanded metabolic measures in longer-term, larger cohorts.</p></sec></abstract><kwd-group kwd-group-type="author-keywords"><title>Keywords</title><kwd>metabolic syndrome</kwd><kwd>breast neoplasms</kwd><kwd>antineoplastic agents</kwd><kwd>cardiometabolic risk factors</kwd><kwd>weight gain</kwd><kwd>hormonal</kwd></kwd-group><custom-meta-group><custom-meta><meta-name>pmc-status-qastatus</meta-name><meta-value>0</meta-value></custom-meta><custom-meta><meta-name>pmc-status-live</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-status-embargo</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-status-released</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-open-access</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-olf</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-manuscript</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-legally-suppressed</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-pdf</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-supplement</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-pdf-only</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-suppress-copyright</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-real-version</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-scanned-article</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-preprint</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-in-epmc</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-license-ref</meta-name><meta-value>CC BY</meta-value></custom-meta></custom-meta-group></article-meta><notes notes-type="disclaimer"><p>The authors declare the use of generative AI in the research and writing process. According to the GAIDeT taxonomy (2025), the following tasks were delegated to GAI tools under full human supervision: Idea generation & Proofreading and editing. The GAI tool used was: ChatGPT 5. Responsibility for the final manuscript lies entirely with the authors. GAI tools are not listed as authors and do not bear responsibility for the final outcomes.</p></notes></front><body><sec sec-type="intro" id="s1"><title>Introduction</title><p>Breast cancer is both the most frequent type of cancer and the leading cause of cancer-related mortality among women [<xref rid="R1" ref-type="bibr">1</xref>]. As survival rates improve, currently up to 90% in developed countries, there is increasing attention to the long-term health of breast cancer survivors [<xref rid="R1" ref-type="bibr">1</xref><xref rid="R2" ref-type="bibr">2</xref>]. Estrogen receptor-positive (ER+) early (stage I-III) breast cancer (EBC) represents the most common subtype, and adjuvant endocrine therapy (ET) is a cornerstone of its treatment [<xref rid="R3" ref-type="bibr">3</xref><xref rid="R4" ref-type="bibr">4</xref>]. While ET has significantly reduced recurrence and mortality, its metabolic side effects have emerged as a growing concern in survivorship care [<xref rid="R4" ref-type="bibr">4-6</xref>].</p><p>ET may influence metabolic health through estrogen deprivation [<xref rid="R5" ref-type="bibr">5</xref><xref rid="R7" ref-type="bibr">7</xref>]. Estrogen plays an important role in metabolic homeostasis by promoting insulin sensitivity, regulating lipid metabolism, and supporting vascular function [<xref rid="R8" ref-type="bibr">8</xref><xref rid="R9" ref-type="bibr">9</xref>]. By blocking estrogen receptors (tamoxifen) or suppressing aromatase activity (aromatase inhibitors, AI), ET may disrupt normal metabolic regulation, promoting weight gain, insulin resistance, and dyslipidemia [<xref rid="R5" ref-type="bibr">5</xref><xref rid="R10" ref-type="bibr">10-13</xref><xref rid="R11" ref-type="bibr">undefined</xref><xref rid="R12" ref-type="bibr">undefined</xref><xref rid="R13" ref-type="bibr">undefined</xref>]. These changes are risk factors for cardiovascular disease (CVD), which represents a major noncancer cause of morbidity and mortality among breast cancer survivors [<xref rid="R14" ref-type="bibr">14</xref>]. Furthermore, obesity and metabolic syndrome (MetS) at diagnosis have been associated with increased risk of breast cancer recurrence and mortality [<xref rid="R15" ref-type="bibr">15-18</xref>].</p><p>Although ET is typically administered for several years, routine clinical follow-up is often concentrated around treatment initiation. This limits systematic long-term monitoring of metabolic health in routine care, which highlights the importance of characterizing early metabolic changes. A short-term follow-up allows evaluation of initial changes in weight, lipid metabolism, and glycemic markers that may signal early metabolic vulnerability. A 3-month timeframe balances biological relevance with clinical feasibility, aligns with standard follow-up practice at the study site, and supports broad inclusion while minimizing selection bias.</p><p>Metabolic health is traditionally assessed using MetS, a dichotomous classification that identifies individuals as metabolically unhealthy when at least three of five criteria are met: hypertension (blood pressure ≥130/85), hypertriglyceridemia (triglycerides >2 mmol/l), low HDL (HDL <1.295 mmol/l), central obesity (waist circumference >88 cm), or hyperglycemia (plasma glucose >7.8 mmol/l) [<xref rid="R19" ref-type="bibr">19</xref>]. However, this binary approach may overlook subtle or early metabolic changes that do not meet clinical thresholds. [<xref rid="R19" ref-type="bibr">19</xref><xref rid="R20" ref-type="bibr">20</xref>]. An established alternative is the Metabolic Syndrome <italic toggle="yes">z</italic> score (MetS-<italic toggle="yes">z</italic> score), a continuous composite measure of standardized MetS components and widely used in research to assess metabolic health along a continuum [<xref rid="R21" ref-type="bibr">21-24</xref>]. However, both MetS and MetS-Z omit several cardiometabolic indicators relevant to breast cancer survivorship, including low-density lipoprotein (LDL), hemoglobin A<sub>1c</sub> (HbA<sub>1c</sub>), BMI, and waist-to-hip ratio (WHR) [<xref rid="R13" ref-type="bibr">13</xref><xref rid="R25" ref-type="bibr">25</xref><xref rid="R26" ref-type="bibr">26</xref>]. To address this limitation, this study introduces two expanded measures of metabolic health, for exploratory and hypothesis-generating purposes: An extended MetS definition and the EMETA score, a dichotomous classification and a composite <italic toggle="yes">z</italic> score using the same approach as the MetS-<italic toggle="yes">z</italic> score, both incorporating the extended metabolic variables.</p><p>The aim of this study is to evaluate the early impact of initiating adjuvant endocrine therapy on metabolic health in women with ER+ EBC and to compare conventional and expanded approaches to classify metabolic health.</p></sec><sec sec-type="methods" id="s2"><title>Methods</title><sec id="s2-1"><title>Study Population</title><p>This prospective, single-center observational study enrolled 112 patients with ER+ EBC who initiated adjuvant ET. Over a six-month period from November 2024, participants were recruited from the Department of Oncology, Aarhus University Hospital (AUH), Denmark. Patients were invited to participate during the consultation ET was prescribed. Participants received usual care throughout the study.</p></sec><sec id="s2-2"><title>Eligibility Criteria</title><p>Patients were eligible if they met the following criteria: age ≥18 years, diagnosis of invasive ER+ EBC, and initiation of adjuvant ET. Up to three days of ET treatment were allowed, to allow for the timeframe of informed consent.</p><p>Patients were excluded for enrollment in case of pregnancy or lactation, pre-existing diabetes (type I or II), more than three dispenses of ET, or any psychological, familial, sociological, or geographical conditions potentially hampering compliance with the study protocol and follow-up schedule. These conditions were discussed with the patient before trial registration. Patients with pre-existing diabetes were excluded because their chronic metabolic dysregulation and glucose-lowering treatments could confound assessment of early metabolic changes. Patients were excluded if metastatic breast cancer was diagnosed in the period between study inclusion and the 3-month follow-up. As the study did not involve any intervention administered by clinical staff beyond routine care, there were no additional eligibility criteria for sites or individuals delivering interventions.</p></sec><sec id="s2-3"><title>Study Design</title><p>Participants underwent two metabolic health screenings: one at baseline and one at 3-month follow-up (<xref rid="F1" ref-type="fig">Figure 1</xref>). The baseline screening was performed on the same day as study inclusion. The 3-month follow-up screening was performed on the same day as the 3-month ET follow-up. This design minimized additional study-related visits and supported participant retention.</p><fig position="float" id="F1" fig-type="figure" orientation="portrait"><label>Figure 1.</label><caption><title> EMETA study design and participation created in BioRender [<xref rid="R27" ref-type="bibr">27</xref>] is licensed under CC BY 4.0 [<xref rid="R28" ref-type="bibr">28</xref>]. AUH: Aarhus University Hospital.</title></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="resprot-v15-e78589-g001.jpg"/></fig></sec><sec id="s2-4"><title>Exposure</title><p>The exposure of interest was initiation of ET over a 3-month period. ET consisted of either tamoxifen (20 mg daily) or an AI (letrozole 2.5 mg daily or exemestane 25 mg daily), prescribed according to standard clinical practice.</p></sec><sec id="s2-5"><title>Outcomes</title><p>The primary outcomes were selected to address whether initiation of ET is associated with early changes in overall metabolic health. Given the short-term follow-up and the aim to capture early, potentially subtle metabolic alterations, the primary outcomes include an anthropometric and a composite metabolic measure.</p><p>The primary outcomes are:</p><list list-type="order" list-content="ol"><list-item><p>Change in mean BMI from baseline to 3-month follow-up</p></list-item><list-item><p>Change in mean MetS-Z from baseline to 3-month follow-up</p></list-item></list><p>BMI was chosen because weight changes are a well-known sideeffect of ET and may occur early following ET initiation, representing a clinically meaningful marker of metabolic health. Although short-term changes may be modest, even small increases may indicate emerging metabolic vulnerability. As a continuous composite measure, the widely used MetS-<italic toggle="yes">z</italic> score was selected to capture multidimensional metabolic changes that may not be reflected by individual metabolic parameters alone [<xref rid="R21" ref-type="bibr">21-24</xref><xref rid="R22" ref-type="bibr">29</xref><xref rid="R23" ref-type="bibr">undefined</xref><xref rid="R24" ref-type="bibr">undefined</xref><xref rid="R29" ref-type="bibr">undefined</xref>].</p><p>Secondary outcomes were selected to further characterize early metabolic changes and to compare expanded and conventional approaches to metabolic health classification.</p><p>The secondary outcomes include:</p><list list-type="order" list-content="ol"><list-item><p>Change in the prevalence of MetS from baseline to 3-month follow-up</p></list-item><list-item><p>Comparison of the prevalence of MetS and extended MetS at 3-month follow-up</p></list-item><list-item><p>Change in mean EMETA score from baseline to 3-month follow-up</p></list-item><list-item><p>Change in mean values of individual metabolic parameters from baseline to 3-month follow-up, including waist circumference, hip circumference, WHR, blood pressure, and circulating levels of HbA<sub>1c</sub>, total cholesterol, triglycerides, LDL cholesterol, HDL cholesterol, and estradiol</p></list-item></list><p>The extended MetS and EMETA score are included as exploratory outcomes to assess whether broader composite measures capture early metabolic changes not identified by conventional definitions. Analysis of individual metabolic parameters provides insight into which metabolic domains are most affected shortly after ET initiation.</p></sec><sec id="s2-6"><title>Measurements</title><p>The screening procedures included biometric measurements and blood sample collections. The biometric measurements were height, weight, hip and waist circumference, and blood pressure. Non-fasting venous blood samples were analyzed for HbA<sub>1c</sub>, plasma glucose, hemoglobin, total cholesterol, triglycerides, LDL, HDL, and circulating estradiol.</p></sec><sec id="s2-7"><title>Biometric Measurements</title><p>The biometric measurements were measured through the following protocols. Weight was measured using a standardized scale with participants wearing light clothing and no shoes. Height was recorded with participants standing straight against a wall-mounted measuring tape without shoes. Waist circumference was measured 5 cm above the center of the navel and hip circumference at the widest part of the hip, both using a stretch-resistant tape, ensuring the tape was parallel to the floor and snug but not compressing the skin. Blood pressure was measured using a calibrated cuff on the upper arm after participants had rested for 5 minutes, with three consecutive readings taken with the participant seated both feet on the floor and both arms on their legs.</p></sec><sec id="s2-8"><title>Blood Samples</title><p>Nonfasting venous blood samples were analyzed by the Department of Clinical Biochemistry, AUH for levels of HbA<sub>1c</sub>, glucose, total cholesterol, triglycerides, LDL, HDL, estradiol, and sex-hormone binding globulin. The blood samples were destroyed immediately after analysis. Results outside standard values were reviewed by consultant physicians and patients, and their general practitioners were notified with follow-up recommendations when necessary.</p></sec><sec id="s2-9"><title>Data Collection and Management</title><p>Information on patient demographics, treatment modalities, and cancer-specific characteristics were collected through a manual search of electronic medical journals, including pathological records. All data were securely stored in a RedCap database hosted by Aarhus University.</p><p>To promote participant retention and ensure complete follow-up, the 3-month appointments were booked by hospital secretaries as part of routine care. If a participant was not scheduled for a follow-up appointment, trial personnel proactively contacted the participant to arrange the study visit. If participants discontinued ET, outcome data were still be collected at the 3-month follow-up visit, unless consent was withdrawn.</p></sec><sec id="s2-10"><title>Cohort Size</title><p>The study aimed to include 112 participants over the six-month enrollment period. The target was based on an annual referral rate of approximately 360 patients to the Department of Oncology, AUH, for adjuvant systemic therapy, of which approximately 90% were eligible for ET due to ER+ EBC, and assuming a 70% participation rate among eligible patients.</p><p>Statistical power calculations supported the sample size. Based on a comparable study with a similar population [<xref rid="R30" ref-type="bibr">30</xref>], where the standard deviation of the change in BMI over 3‐5 years after diagnosis was 3, a cohort of 112 participants provides >80% power to detect a true change in BMI of 0.8 kg/m<sup>2</sup> at a 5% significance level. Furthermore, other studies have found larger BMI changes over shorter periods. For instance, Heideman et al observed a mean BMI change of 2.0 kg/m<sup>2</sup> (± 4.9) after one year, with 26% of patients gaining ≥5 kg following adjuvant treatment [<xref rid="R31" ref-type="bibr">31</xref>].</p></sec><sec id="s2-11"><title>Statistical Analysis</title><sec id="s2-11-1"><title>Definition of Analytic Values</title><p>BMI will be calculated as weight (kg)/height<sup>2</sup> (m). WHR will be calculated as waist circumference divided by hip circumference. Blood pressure will be calculated as the average of three measurements. Middle arterial pressure will be calculated as 2 times diastolic blood pressure added systolic blood pressure divided by 3.</p><p>MetS will be defined as meeting at least three of five criteria: hypertension (blood pressure ≥130/85), hypertriglyceridemia (triglycerides >2 mmol/l), low HDL (HDL <1.295 mmol/l), central obesity (waist circumference >88 cm), or hyperglycemia (plasma glucose >7.8 mmol/l) (<xref rid="F2" ref-type="fig">Figure 2</xref>).</p><p>Extended MetS will be defined as meeting at least 3 of the five criteria: hypertension (blood pressure ≥130/85), hypertriglycemia (triglycerides >2 mmol/l or LDL >3 mmol/l), low HDL (HDL <1.295 mmol/l), adiposity (waist circumference >88 cm, BMI ≥30, or WHR >0.85), or hyperglycemia (plasma glucose >7.8 mmol/l or HbA<sub>1c</sub> ≥42 mmol/mol).</p><fig position="float" id="F2" fig-type="figure" orientation="portrait"><label>Figure 2.</label><caption><title>Comparison of MetS, extended MetS, MetS-Z, and EMETA scores. Red text: Variables specific to the extended MetS and EMETA score. HDL: High-density lipoprotein, LDL: Low-density lipoprotein. HbA<sub>1c</sub>: Hemoglobin A<sub>1c</sub>. Created in BioRender. Busk Holm, J. (2025) <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://BioRender.com/" ext-link-type="uri">https://BioRender.com/</ext-link></title></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="resprot-v15-e78589-g002.jpg"/></fig></sec><sec id="s2-11-2"><title><italic toggle="yes">z</italic> Score Calculations</title><p>The MetS-<italic toggle="yes">z</italic> score will be calculated as a standardized composite score of the five MetS components. Standardization will be achieved by subtracting ATP III criteria from each individual’s value of waist circumference (WC), triglycerides (TG), middle arterial pressure (MAP), plasma glucose (PG), and inverted HDL, dividing by the baseline cohort standard deviation (σ).</p><disp-formula id="eqn1">
<label>(1)</label>
<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1" overflow="scroll"><mml:mstyle displaystyle="true" scriptlevel="0"><mml:mrow><mml:mstyle displaystyle="true" scriptlevel="0"><mml:mrow><mml:mi mathvariant="normal">M</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">e</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">t</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">S</mml:mi></mml:mrow><mml:mtext>-</mml:mtext><mml:mi>Z</mml:mi><mml:mo>=</mml:mo><mml:mfrac><mml:mrow><mml:mrow><mml:mi mathvariant="normal">W</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">C</mml:mi></mml:mrow><mml:mo>−</mml:mo><mml:mn>88</mml:mn></mml:mrow><mml:msub><mml:mrow><mml:mi>σ</mml:mi></mml:mrow><mml:mrow><mml:mrow><mml:mi mathvariant="normal">w</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">c</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mrow><mml:mi mathvariant="normal">T</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">G</mml:mi></mml:mrow><mml:mo>−</mml:mo><mml:mn>2</mml:mn></mml:mrow><mml:msub><mml:mrow><mml:mi>σ</mml:mi></mml:mrow><mml:mrow><mml:mrow><mml:mi mathvariant="normal">T</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">G</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mrow><mml:mi mathvariant="normal">M</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">A</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mo>−</mml:mo><mml:mn>100</mml:mn></mml:mrow><mml:msub><mml:mrow><mml:mi>σ</mml:mi></mml:mrow><mml:mrow><mml:mrow><mml:mi mathvariant="normal">M</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">A</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">P</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mrow><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">G</mml:mi></mml:mrow><mml:mo>−</mml:mo><mml:mn>7.8</mml:mn></mml:mrow><mml:msub><mml:mrow><mml:mi>σ</mml:mi></mml:mrow><mml:mrow><mml:mrow><mml:mi mathvariant="normal">P</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">G</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:mfrac><mml:mo>−</mml:mo><mml:mfrac><mml:mrow><mml:mrow><mml:mi mathvariant="normal">H</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">D</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">L</mml:mi></mml:mrow><mml:mo>−</mml:mo><mml:mn>1.295</mml:mn></mml:mrow><mml:msub><mml:mrow><mml:mi>σ</mml:mi></mml:mrow><mml:mrow><mml:mrow><mml:mi mathvariant="normal">H</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">D</mml:mi></mml:mrow><mml:mrow><mml:mi mathvariant="normal">L</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:mfrac></mml:mstyle></mml:mrow></mml:mstyle></mml:math>
</disp-formula><p>The EMETA score is a exploratory composite <italic toggle="yes">z</italic> score that will be calculated using the same approach as MetS-<italic toggle="yes">z</italic> score, based on the extended MetS components.</p><disp-formula id="eqn2">
<label>(2)</label>
<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2" overflow="scroll"><mml:mstyle displaystyle="true" scriptlevel="0"><mml:mrow><mml:mstyle displaystyle="true" scriptlevel="0"><mml:mtable columnalign="right left right left right left right left right left right left" rowspacing="3pt" columnspacing="0em 2em 0em 2em 0em 2em 0em 2em 0em 2em 0em" displaystyle="true"><mml:mtr><mml:mtd><mml:mtext>EMETA</mml:mtext><mml:mo>=</mml:mo></mml:mtd><mml:mtd><mml:mfrac><mml:mrow><mml:mi>W</mml:mi><mml:mi>C</mml:mi><mml:mo>−</mml:mo><mml:mn>88</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>W</mml:mi><mml:mi>C</mml:mi></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mi>B</mml:mi><mml:mi>M</mml:mi><mml:mi>I</mml:mi><mml:mo>−</mml:mo><mml:mn>30</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>B</mml:mi><mml:mi>M</mml:mi><mml:mi>I</mml:mi></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mi>W</mml:mi><mml:mi>H</mml:mi><mml:mi>R</mml:mi><mml:mo>−</mml:mo><mml:mn>0.85</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>W</mml:mi><mml:mi>H</mml:mi><mml:mi>R</mml:mi></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mi>T</mml:mi><mml:mi>G</mml:mi><mml:mo>−</mml:mo><mml:mn>2</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>T</mml:mi><mml:mi>G</mml:mi></mml:mrow></mml:msub></mml:mfrac></mml:mtd></mml:mtr><mml:mtr><mml:mtd><mml:mo>−</mml:mo><mml:mfrac><mml:mrow><mml:mi>H</mml:mi><mml:mi>D</mml:mi><mml:mi>L</mml:mi><mml:mo>−</mml:mo><mml:mn>1.295</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>H</mml:mi><mml:mi>D</mml:mi><mml:mi>L</mml:mi></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mi>L</mml:mi><mml:mi>D</mml:mi><mml:mi>L</mml:mi><mml:mo>−</mml:mo><mml:mn>3</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>L</mml:mi><mml:mi>D</mml:mi><mml:mi>L</mml:mi></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mi>M</mml:mi><mml:mi>A</mml:mi><mml:mi>P</mml:mi><mml:mo>−</mml:mo><mml:mn>100</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>M</mml:mi><mml:mi>A</mml:mi><mml:mi>P</mml:mi></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mi>P</mml:mi><mml:mi>G</mml:mi><mml:mo>−</mml:mo><mml:mn>7.8</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>P</mml:mi><mml:mi>G</mml:mi></mml:mrow></mml:msub></mml:mfrac><mml:mo>+</mml:mo><mml:mfrac><mml:mrow><mml:mi>H</mml:mi><mml:mi>b</mml:mi><mml:mi>A</mml:mi><mml:mn>1</mml:mn><mml:mi>c</mml:mi><mml:mo>−</mml:mo><mml:mn>42</mml:mn></mml:mrow><mml:msub><mml:mi>σ</mml:mi><mml:mrow><mml:mi>H</mml:mi><mml:mi>b</mml:mi><mml:mi>A</mml:mi><mml:mn>1</mml:mn><mml:mi>c</mml:mi></mml:mrow></mml:msub></mml:mfrac></mml:mtd></mml:mtr></mml:mtable></mml:mstyle></mml:mrow></mml:mstyle></mml:math>
</disp-formula></sec><sec id="s2-11-3"><title>Covariables</title><p>Information on the following co-variables has been collected from medical records and will be included in adjusted analyses. These include patient demographics: age, weight, height, Carlson Comorbidity score (0, 1‐2, 3‐4, ≥5), current comorbidities (none, cardiopulmonary disease, neurological, dyslipidemia, musculoskeletal, other cancers, abdominal, urogenital, endocrine, dermatological, others), smoking status (never smoker, former smoker, active smoker), civil status (living alone, living with a partner), and area of living (rural/city). Cancer-specific characteristics: date of breast cancer diagnosis, disease stage (T+N), human epidermal growth factor receptor 2 status, histological type, grade of malignancy, and previous breast cancer history. Treatment modalities included duration and type of neo-adjuvant and adjuvant therapy (including supportive prednisolone treatment), date of breast cancer surgery, and radiotherapy.</p></sec></sec><sec id="s2-12"><title>Analytical Approach</title><p>Statistical analyses will be conducted using Stata (version 18.5), with a 5% significance level for hypothesis testing. Changes in continuous primary outcomes (BMI and MetS-<italic toggle="yes">z</italic> score) will be analyzed using linear regression or linear mixed-effects models, as appropriate, to estimate mean within-person changes from baseline to 3-month follow-up. Changes in secondary continuous outcomes, including the EMETA score and individual metabolic parameters, will be analyzed using similar regression-based approaches. Changes in categorical outcomes (MetS and Extended MetS prevalence) will be analyzed using logistic regression models.</p><p>To account for exposure heterogeneity and treatment-related confounding, type of ET (tamoxifen vs aromatase inhibitor) and prior cancer treatments with potential metabolic impact, including neoadjuvant or adjuvant chemotherapy and supportive corticosteroid use, will be included as covariates in all adjusted analyses.</p><p>To address the secondary research question of whether certain patients are more vulnerable to early metabolic health deterioration following the initiation of ET, exploratory subgroup analyses will be conducted across both primary and secondary outcomes. These analyses will examine whether metabolic changes differ according to baseline and treatment-related characteristics, including age, BMI category, baseline metabolic status, menopausal status, type of ET, prior chemotherapy, smoking status, and comorbidity score. These subgroup analyses will be considered hypothesis-generating. Missing data will be assessed for randomness and extent. If data are missing at random, appropriate methods such as multiple imputation or mixed-effects models will be applied. Sensitivity analyses will be conducted to evaluate the potential impact of missing data on the study’s primary and secondary outcomes.</p></sec><sec id="s2-13"><title>Ethical Considerations</title><p>This study was conducted in agreement with the Declaration of Helsinki and the laws and regulations of Denmark, whichever provided the highest level of patient protection. The study protocol was approved by the Scientific Ethics Committee for the Central Denmark Region (case nr: 1-10-72-122-24). Any important modifications to the protocol (eg, changes to eligibility criteria, outcomes, or analysis methods) were communicated to the relevant ethics committee and trial registry (ClinicalTrials.gov ID <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="pmc:clinical-trial" xlink:href="NCT06623903">NCT06623903</ext-link>).</p><sec id="s2-13-1"><title>Informed Consent</title><p>Written informed consent was obtained from all participants before study entry, and any study-specific procedures were performed. This process was conducted in accordance with national and local regulatory requirements. Each consent form had to be signed and personally dated by the participant.</p></sec><sec id="s2-13-2"><title>Risks and Patient Inconvenience</title><p>The risk of adverse events associated with study participation was minimal. Rarely, minor bleeding or inflammation could occur at the blood sample puncture site. No long-term adverse events were anticipated. Participation required two additional blood samples and an approximate extension of 30 minutes during two oncology department visits (at study inclusion and the three-mo follow-up). Participants were covered under the standard patient compensation system in Denmark for any unexpected harm related to study procedures.</p></sec><sec id="s2-13-3"><title>Data Protection and Privacy</title><p>Personal information collected during the study was securely stored in the RedCap database hosted at Aarhus University. Data were pseudonymized to enable linkage with clinical records, with access to identifiable information restricted to authorized study personnel. All data were stored on secure, password-protected servers with two-factor authentication. All data were processed in accordance with the Danish Data Protection Act and the European General Data Protection Regulation (GDPR). The project was registered in the internal directory of the Central Denmark Region.</p></sec><sec id="s2-13-4"><title>Dissemination of Results</title><p>Study findings will be disseminated in international peer-reviewed journals and shared at local journal clubs, scientific meetings, and at least one international congress. Study participants who consented at study inclusion to receive study results will be provided with a lay summary of the findings, supporting transparency and engagement in survivorship care. Patients or members of the public were not involved in the design, conduct, or reporting of this study.</p></sec><sec id="s2-13-5"><title>Monitoring</title><p>A formal Data Monitoring Committee was not established for this study, as this was a noninterventional, observational trial involving minimal risk. Instead, the study was overseen by the study personnel (the authors) who were responsible for the daily conduct and integrity of the study. The study personnel completed and monitored all data collection to ensure protocol adherence, data quality, and participant safety. As the study sponsor (AUH) did not influence design, data analysis, or reporting, oversight by the study team was considered both appropriate and independent in practice. No interim analyses or formal stopping guidelines were planned, given the short follow-up period and non-interventional design. All data will be analyzed following the completion of the follow-up period.</p></sec></sec></sec><sec sec-type="results" id="s3"><title>Results</title><p>The study was funded in July 2024. Participant recruitment occurred from November 2024 through April 2025, during which 112 of 166 eligible patients were enrolled. Among the 54 eligible patients who were not enrolled, 18 declined participation, 26 were not informed of the study by their treating practitioner, and 10 elected not to initiate adjuvant ET. All enrolled participants completed baseline assessments, and follow-up data collection was completed in September 2025. No deviations from the approved study protocol occurred, and no protocol amendments were made following ethical approval. Data cleaning and statistical analyses are scheduled to commence in February 2026, with dissemination of the final study results anticipated in July 2026.</p></sec><sec sec-type="discussion" id="s4"><title>Discussion</title><p>This study is anticipated to indicate that initiation of ET in women with ER+ EBC will be associated with modest early metabolic changes over the first three months of treatment. These changes may include small increases in BMI and selected lipid-related parameters, while markers with longer biological response times, such as HbA<sub>1c</sub>, may show minimal or no short-term change. Extended and continuous measures of metabolic health may capture early alterations not identified by conventional MetS definitions. Importantly, these analyses are exploratory and intended to evaluate sensitivity and feasibility rather than to establish clinical thresholds or predictive value.</p><p>Previous studies have linked ET with weight gain, insulin resistance, and dyslipidemia, all of which increase cardiovascular risk and may impact breast cancer outcomes [<xref rid="R5" ref-type="bibr">5</xref><xref rid="R10" ref-type="bibr">10-12</xref><xref rid="R11" ref-type="bibr">14</xref><xref rid="R12" ref-type="bibr">32</xref><xref rid="R14" ref-type="bibr">undefined</xref><xref rid="R32" ref-type="bibr">undefined</xref>]. However, most prior work has focused on longer-term outcomes or relied on dichotomous classifications. By evaluating both conventional and expanded measures within a short-term framework, this study contributes to understanding how early metabolic changes may be characterized following ET initiation.</p><p>Key strengths of this study include its prospective design with paired baseline and 3-month follow-up data, allowing within-subject comparison of early metabolic changes. The use of both conventional and expanded metabolic measures enables comparison of different classification approaches and addresses known limitations of binary definitions. The short follow-up period and alignment with routine clinical care minimize additional participant burden, supporting broad inclusion and reducing the risk of selection bias related to study participation.</p><p>The study is conducted at a single center, which may limit generalizability beyond similar healthcare settings. Exclusion of patients with pre-existing diabetes, while necessary to reduce baseline metabolic confounding, restricts applicability to a population that commonly receives endocrine therapy. In addition, although efforts were made to minimize attrition, informative dropout cannot be excluded, particularly if early metabolic or treatment-related symptoms influenced follow-up participation. The modest sample size limits power for subgroup analyses, which are therefore considered exploratory.</p><p>Overall, this study aims to inform future research and clinical practice by clarifying early metabolic patterns following ET initiation and by evaluating alternative approaches to metabolic health assessment. Larger and longer-term studies are needed to determine whether early metabolic changes are associated with subsequent cardiovascular risk or breast cancer outcomes, and to validate expanded metabolic measures before any potential clinical application.</p></sec></body><back><ack><title>Acknowledgments</title><p>The authors would like to express their gratitude to and acknowledge the patients who participated in the study, and the scientific and medical staff from the Department of Oncology, AUH, who participated in conducting the study. The sponsor of this study is AUH (Contact: AarhusUniversitetshospital@auh.rm.dk).</p></ack><fn-group><fn fn-type="financial-disclosure"><p><bold>Funding:</bold> This work was supported by The Danish Cancer Society grant number R382-A22956, Fru Astrid Thaysens Legat for Lægevidenskabelig Grundforskning grant number ATL 24/06, and Department of Oncology Research Fund grant number 09-2271.</p></fn><fn fn-type="other"><p><bold>Authors’ Contributions:</bold> Conceptualiztion: LN (lead), SB (equal), JBH (equal), and SB research group (supporting).</p><p>Data curation, funding acqusition, investigation, methology, project administration, visualization: LN (lead), SB (supporting), JBH (supporting).</p><p>Supervision: SB (Lead), JBH (supporting).</p><p>Writing-original draft, review & editing: LN (lead), JBH (Supporting), SB (supporting).</p></fn><fn fn-type="other"><p><bold>Data Availability:</bold> The datasets generated or analyzed during this study are not publicly available due to patient privacy and confidentiality concerns but are available from the corresponding author on reasonable request.</p></fn><fn fn-type="COI-statement"><p><bold>Conflicts of Interest:</bold> None declared.</p></fn></fn-group><notes><def-list><title>Abbreviations</title><def-item><term>AI</term><def><p>aromatase inhibitor</p></def></def-item><def-item><term>AUH</term><def><p>Aarhus University Hospital</p></def></def-item><def-item><term>CVD</term><def><p>cardiovascular disease</p></def></def-item><def-item><term>EBC</term><def><p>early breast cancer</p></def></def-item><def-item><term>ER+</term><def><p>estrogen receptor positive</p></def></def-item><def-item><term>ET</term><def><p>adjuvant endocrine therapy</p></def></def-item><def-item><term>HbA<sub>1c</sub></term><def><p>hemoglobin A<sub>1c</sub></p></def></def-item><def-item><term>HDL</term><def><p>high-density lipoprotein</p></def></def-item><def-item><term>LDL</term><def><p>low-density lipoprotein</p></def></def-item><def-item><term>MetS</term><def><p>metabolic syndrome</p></def></def-item><def-item><term>MetS-Z score</term><def><p>Metabolic Syndrome z score</p></def></def-item><def-item><term>WHR</term><def><p>waist-to-hip ratio</p></def></def-item></def-list></notes><ref-list><title>References</title><ref id="R1"><label>1.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Miller</surname><given-names>KD</given-names></name>
<name name-style="western"><surname>Nogueira</surname><given-names>L</given-names></name>
<name name-style="western"><surname>Devasia</surname><given-names>T</given-names></name>
<etal>et al</etal>
</person-group><article-title>Cancer treatment and survivorship statistics, 2022</article-title><source>CA Cancer J Clin</source><month>Sep</month><year>2022</year><volume>72</volume><issue>5</issue><fpage>409</fpage><lpage>436</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.3322/caac.21731</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">35736631</pub-id></element-citation></ref><ref id="R2"><label>2.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Sung</surname><given-names>H</given-names></name>
<name name-style="western"><surname>Ferlay</surname><given-names>J</given-names></name>
<name name-style="western"><surname>Siegel</surname><given-names>RL</given-names></name>
<etal>et al</etal>
</person-group><article-title>Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries</article-title><source>CA Cancer J Clin</source><month>05</month><year>2021</year><volume>71</volume><issue>3</issue><fpage>209</fpage><lpage>249</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.3322/caac.21660</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">33538338</pub-id></element-citation></ref><ref id="R3"><label>3.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Galuia</surname><given-names>M</given-names></name>
<name name-style="western"><surname>Fedorova</surname><given-names>J</given-names></name>
<name name-style="western"><surname>McHayleh</surname><given-names>W</given-names></name>
<name name-style="western"><surname>Mamounas</surname><given-names>E</given-names></name>
<name name-style="western"><surname>Ahmad</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Pavri</surname><given-names>S</given-names></name>
</person-group><article-title>Perioperative drug management of systemic therapies in breast cancer: a literature review and treatment recommendations</article-title><source>Curr Oncol</source><month>Mar</month><day>9</day><year>2025</year><volume>32</volume><issue>3</issue><elocation-id>154</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.3390/curroncol32030154</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">40136358</pub-id><pub-id pub-id-type="pmcid">PMC11941735</pub-id></element-citation></ref><ref id="R4"><label>4.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Franzoi</surname><given-names>MA</given-names></name>
<name name-style="western"><surname>Agostinetto</surname><given-names>E</given-names></name>
<name name-style="western"><surname>Perachino</surname><given-names>M</given-names></name>
<etal>et al</etal>
</person-group><article-title>Evidence-based approaches for the management of side-effects of adjuvant endocrine therapy in patients with breast cancer</article-title><source>Lancet Oncol</source><month>07</month><year>2021</year><volume>22</volume><issue>7</issue><fpage>e303</fpage><lpage>e313</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1016/S1470-2045(20)30666-5</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">33891888</pub-id></element-citation></ref><ref id="R5"><label>5.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Thomas</surname><given-names>NS</given-names></name>
<name name-style="western"><surname>Scalzo</surname><given-names>RL</given-names></name>
<name name-style="western"><surname>Wellberg</surname><given-names>EA</given-names></name>
</person-group><article-title>Diabetes mellitus in breast cancer survivors: metabolic effects of endocrine therapy</article-title><source>Nat Rev Endocrinol</source><month>01</month><year>2024</year><volume>20</volume><issue>1</issue><fpage>16</fpage><lpage>26</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1038/s41574-023-00899-0</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">37783846</pub-id><pub-id pub-id-type="pmcid">PMC11487546</pub-id></element-citation></ref><ref id="R6"><label>6.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Nardin</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Mora</surname><given-names>E</given-names></name>
<name name-style="western"><surname>Varughese</surname><given-names>FM</given-names></name>
<etal>et al</etal>
</person-group><article-title>Breast cancer survivorship, quality of life, and late toxicities</article-title><source>Front Oncol</source><year>2020</year><volume>10</volume><elocation-id>864</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.3389/fonc.2020.00864</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">32612947</pub-id><pub-id pub-id-type="pmcid">PMC7308500</pub-id></element-citation></ref><ref id="R7"><label>7.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Li</surname><given-names>Y</given-names></name>
<name name-style="western"><surname>Deng</surname><given-names>Z</given-names></name>
<name name-style="western"><surname>Wang</surname><given-names>Y</given-names></name>
<name name-style="western"><surname>Shen</surname><given-names>S</given-names></name>
</person-group><article-title>Lipid changes during endocrine therapy in early-stage breast cancer patients: a real-world study</article-title><source>Lipids Health Dis</source><month>01</month><day>8</day><year>2024</year><volume>23</volume><issue>1</issue><elocation-id>9</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.1186/s12944-024-02002-6</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">38191454</pub-id><pub-id pub-id-type="pmcid">PMC10773127</pub-id></element-citation></ref><ref id="R8"><label>8.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Tramunt</surname><given-names>B</given-names></name>
<name name-style="western"><surname>Smati</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Grandgeorge</surname><given-names>N</given-names></name>
<etal>et al</etal>
</person-group><article-title>Sex differences in metabolic regulation and diabetes susceptibility</article-title><source>Diabetologia</source><month>Mar</month><year>2020</year><volume>63</volume><issue>3</issue><fpage>453</fpage><lpage>461</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1007/s00125-019-05040-3</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">31754750</pub-id><pub-id pub-id-type="pmcid">PMC6997275</pub-id></element-citation></ref><ref id="R9"><label>9.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Sánchez-García</surname><given-names>M</given-names></name>
<name name-style="western"><surname>León-Wu</surname><given-names>K</given-names></name>
<name name-style="western"><surname>de Miguel-Ibáñez</surname><given-names>R</given-names></name>
<etal>et al</etal>
</person-group><article-title>Metabolic changes in patients with premature ovarian insufficiency: adipose tissue focus-a narrative review</article-title><source>Metabolites</source><month>Apr</month><day>2</day><year>2025</year><volume>15</volume><issue>4</issue><elocation-id>242</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.3390/metabo15040242</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">40278371</pub-id><pub-id pub-id-type="pmcid">PMC12029191</pub-id></element-citation></ref><ref id="R10"><label>10.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Kwan</surname><given-names>ML</given-names></name>
<name name-style="western"><surname>Cheng</surname><given-names>RK</given-names></name>
<name name-style="western"><surname>Iribarren</surname><given-names>C</given-names></name>
<etal>et al</etal>
</person-group><article-title>Risk of cardiometabolic risk factors in women with and without a history of breast cancer: The Pathways Heart Study</article-title><source>J Clin Oncol</source><month>05</month><day>20</day><year>2022</year><volume>40</volume><issue>15</issue><fpage>1635</fpage><lpage>1646</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1200/JCO.21.01738</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">35025627</pub-id><pub-id pub-id-type="pmcid">PMC9113213</pub-id></element-citation></ref><ref id="R11"><label>11.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Rillamas-Sun</surname><given-names>E</given-names></name>
<name name-style="western"><surname>Kwan</surname><given-names>ML</given-names></name>
<name name-style="western"><surname>Iribarren</surname><given-names>C</given-names></name>
<etal>et al</etal>
</person-group><article-title>Development of cardiometabolic risk factors following endocrine therapy in women with breast cancer</article-title><source>Breast Cancer Res Treat</source><month>Aug</month><year>2023</year><volume>201</volume><issue>1</issue><fpage>117</fpage><lpage>126</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1007/s10549-023-06997-x</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">37326764</pub-id><pub-id pub-id-type="pmcid">PMC10498727</pub-id></element-citation></ref><ref id="R12"><label>12.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Franchi</surname><given-names>M</given-names></name>
<name name-style="western"><surname>Tritto</surname><given-names>R</given-names></name>
<name name-style="western"><surname>Tarantini</surname><given-names>L</given-names></name>
<name name-style="western"><surname>Navazio</surname><given-names>A</given-names></name>
<name name-style="western"><surname>Corrao</surname><given-names>G</given-names></name>
</person-group><article-title>Adjuvant hormonotherapy and cardiovascular risk in post-menopausal women with breast cancer: a large population-based cohort study</article-title><source>Cancers (Basel)</source><month>05</month><day>8</day><year>2021</year><volume>13</volume><issue>9</issue><elocation-id>2254</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.3390/cancers13092254</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">34066685</pub-id><pub-id pub-id-type="pmcid">PMC8125834</pub-id></element-citation></ref><ref id="R13"><label>13.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Mehta</surname><given-names>LS</given-names></name>
<name name-style="western"><surname>Watson</surname><given-names>KE</given-names></name>
<name name-style="western"><surname>Barac</surname><given-names>A</given-names></name>
<etal>et al</etal>
</person-group><article-title>Cardiovascular disease and breast cancer: where these entities intersect: a scientific statement from the American Heart Association</article-title><source>Circulation</source><month>Feb</month><day>20</day><year>2018</year><volume>137</volume><issue>8</issue><fpage>e30</fpage><lpage>e66</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1161/CIR.0000000000000556</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">29437116</pub-id><pub-id pub-id-type="pmcid">PMC6722327</pub-id></element-citation></ref><ref id="R14"><label>14.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Bradshaw</surname><given-names>PT</given-names></name>
<name name-style="western"><surname>Stevens</surname><given-names>J</given-names></name>
<name name-style="western"><surname>Khankari</surname><given-names>N</given-names></name>
<name name-style="western"><surname>Teitelbaum</surname><given-names>SL</given-names></name>
<name name-style="western"><surname>Neugut</surname><given-names>AI</given-names></name>
<name name-style="western"><surname>Gammon</surname><given-names>MD</given-names></name>
</person-group><article-title>Cardiovascular disease mortality among breast cancer survivors</article-title><source>Epidemiology</source><month>01</month><year>2016</year><volume>27</volume><issue>1</issue><fpage>6</fpage><lpage>13</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1097/EDE.0000000000000394</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">26414938</pub-id><pub-id pub-id-type="pmcid">PMC4666721</pub-id></element-citation></ref><ref id="R15"><label>15.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Harborg</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Cronin-Fenton</surname><given-names>D</given-names></name>
<name name-style="western"><surname>Jensen</surname><given-names>MBR</given-names></name>
<name name-style="western"><surname>Ahern</surname><given-names>TP</given-names></name>
<name name-style="western"><surname>Ewertz</surname><given-names>M</given-names></name>
<name name-style="western"><surname>Borgquist</surname><given-names>S</given-names></name>
</person-group><article-title>Obesity and risk of recurrence in patients with breast cancer treated with aromatase inhibitors</article-title><source>JAMA Netw Open</source><month>Oct</month><day>2</day><year>2023</year><volume>6</volume><issue>10</issue><elocation-id>e2337780</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.1001/jamanetworkopen.2023.37780</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">37831449</pub-id><pub-id pub-id-type="pmcid">PMC10576219</pub-id></element-citation></ref><ref id="R16"><label>16.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Playdon</surname><given-names>MC</given-names></name>
<name name-style="western"><surname>Bracken</surname><given-names>MB</given-names></name>
<name name-style="western"><surname>Sanft</surname><given-names>TB</given-names></name>
<name name-style="western"><surname>Ligibel</surname><given-names>JA</given-names></name>
<name name-style="western"><surname>Harrigan</surname><given-names>M</given-names></name>
<name name-style="western"><surname>Irwin</surname><given-names>ML</given-names></name>
</person-group><article-title>Weight gain after breast cancer diagnosis and all-cause mortality: systematic review and meta-analysis</article-title><source>J Natl Cancer Inst</source><month>Dec</month><year>2015</year><volume>107</volume><issue>12</issue><elocation-id>djv275</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.1093/jnci/djv275</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">26424778</pub-id><pub-id pub-id-type="pmcid">PMC4715249</pub-id></element-citation></ref><ref id="R17"><label>17.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Wang</surname><given-names>X</given-names></name>
<name name-style="western"><surname>Hui</surname><given-names>TL</given-names></name>
<name name-style="western"><surname>Wang</surname><given-names>MQ</given-names></name>
<name name-style="western"><surname>Liu</surname><given-names>H</given-names></name>
<name name-style="western"><surname>Li</surname><given-names>RY</given-names></name>
<name name-style="western"><surname>Song</surname><given-names>ZC</given-names></name>
</person-group><article-title>Body mass index at diagnosis as a prognostic factor for early-stage invasive breast cancer after surgical resection</article-title><source>Oncol Res Treat</source><year>2019</year><volume>42</volume><issue>4</issue><fpage>195</fpage><lpage>201</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1159/000496548</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">30852575</pub-id></element-citation></ref><ref id="R18"><label>18.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Harborg</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Larsen</surname><given-names>HB</given-names></name>
<name name-style="western"><surname>Elsgaard</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Borgquist</surname><given-names>S</given-names></name>
</person-group><article-title>Metabolic syndrome is associated with breast cancer mortality: a systematic review and meta-analysis</article-title><source>J Intern Med</source><month>Mar</month><year>2025</year><volume>297</volume><issue>3</issue><fpage>262</fpage><lpage>275</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1111/joim.20052</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">39775978</pub-id><pub-id pub-id-type="pmcid">PMC11846077</pub-id></element-citation></ref><ref id="R19"><label>19.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Neeland</surname><given-names>IJ</given-names></name>
<name name-style="western"><surname>Lim</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Tchernof</surname><given-names>A</given-names></name>
<etal>et al</etal>
</person-group><article-title>Metabolic syndrome</article-title><source>Nat Rev Dis Primers</source><month>Oct</month><day>17</day><year>2024</year><volume>10</volume><issue>1</issue><fpage>77</fpage><comment>doi</comment><pub-id pub-id-type="doi">10.1038/s41572-024-00563-5</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">39420195</pub-id></element-citation></ref><ref id="R20"><label>20.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>DeBoer</surname><given-names>MD</given-names></name>
<name name-style="western"><surname>Filipp</surname><given-names>SL</given-names></name>
<name name-style="western"><surname>Gurka</surname><given-names>MJ</given-names></name>
</person-group><article-title>Use of a Metabolic syndrome severity <italic toggle="yes">z</italic> score to track risk during treatment of prediabetes: an analysis of the Diabetes Prevention Program</article-title><source>Diabetes Care</source><month>Nov</month><year>2018</year><volume>41</volume><issue>11</issue><fpage>2421</fpage><lpage>2430</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.2337/dc18-1079</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">30275282</pub-id><pub-id pub-id-type="pmcid">PMC6196828</pub-id></element-citation></ref><ref id="R21"><label>21.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Johnson</surname><given-names>JL</given-names></name>
<name name-style="western"><surname>Slentz</surname><given-names>CA</given-names></name>
<name name-style="western"><surname>Houmard</surname><given-names>JA</given-names></name>
<etal>et al</etal>
</person-group><article-title>Exercise training amount and intensity effects on metabolic syndrome (from Studies of a Targeted Risk Reduction Intervention through Defined Exercise)</article-title><source>Am J Cardiol</source><month>Dec</month><day>15</day><year>2007</year><volume>100</volume><issue>12</issue><fpage>1759</fpage><lpage>1766</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1016/j.amjcard.2007.07.027</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">18082522</pub-id><pub-id pub-id-type="pmcid">PMC2190779</pub-id></element-citation></ref><ref id="R22"><label>22.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Hettchen</surname><given-names>M</given-names></name>
<name name-style="western"><surname>von Stengel</surname><given-names>S</given-names></name>
<name name-style="western"><surname>Kohl</surname><given-names>M</given-names></name>
<etal>et al</etal>
</person-group><article-title>Changes in menopausal risk factors in early postmenopausal osteopenic women after 13 months of high-intensity exercise: the randomized controlled ACTLIFE-RCT</article-title><source>Clin Interv Aging</source><year>2021</year><volume>16</volume><fpage>83</fpage><lpage>96</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.2147/CIA.S283177</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">33469276</pub-id><pub-id pub-id-type="pmcid">PMC7810823</pub-id></element-citation></ref><ref id="R23"><label>23.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Morales-Palomo</surname><given-names>F</given-names></name>
<name name-style="western"><surname>Moreno-Cabañas</surname><given-names>A</given-names></name>
<name name-style="western"><surname>Alvarez-Jimenez</surname><given-names>L</given-names></name>
<name name-style="western"><surname>Mora-Gonzalez</surname><given-names>D</given-names></name>
<name name-style="western"><surname>Mora-Rodriguez</surname><given-names>R</given-names></name>
</person-group><article-title>Long-term effects of high-intensity aerobic training on metabolic syndrome: an 8-year follow-up randomized clinical trial</article-title><source>J Cachexia Sarcopenia Muscle</source><month>Apr</month><year>2025</year><volume>16</volume><issue>2</issue><elocation-id>e13780</elocation-id><comment>doi</comment><pub-id pub-id-type="doi">10.1002/jcsm.13780</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">40171886</pub-id><pub-id pub-id-type="pmcid">PMC11962644</pub-id></element-citation></ref><ref id="R24"><label>24.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Gonzalo-Encabo</surname><given-names>P</given-names></name>
<name name-style="western"><surname>Christopher</surname><given-names>CN</given-names></name>
<name name-style="western"><surname>Lee</surname><given-names>K</given-names></name>
<etal>et al</etal>
</person-group><article-title>High-intensity interval training improves metabolic syndrome in women with breast cancer receiving Anthracyclines</article-title><source>Scand J Med Sci Sports</source><month>Apr</month><year>2023</year><volume>33</volume><issue>4</issue><fpage>475</fpage><lpage>484</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1111/sms.14280</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">36427275</pub-id></element-citation></ref><ref id="R25"><label>25.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Grundy</surname><given-names>SM</given-names></name>
<name name-style="western"><surname>Cleeman</surname><given-names>JI</given-names></name>
<name name-style="western"><surname>Daniels</surname><given-names>SR</given-names></name>
<etal>et al</etal>
</person-group><article-title>Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement</article-title><source>Circulation</source><month>Oct</month><day>25</day><year>2005</year><volume>112</volume><issue>17</issue><fpage>2735</fpage><lpage>2752</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1161/CIRCULATIONAHA.105.169404</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">16157765</pub-id></element-citation></ref><ref id="R26"><label>26.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Sinning</surname><given-names>C</given-names></name>
<name name-style="western"><surname>Makarova</surname><given-names>N</given-names></name>
<name name-style="western"><surname>Völzke</surname><given-names>H</given-names></name>
<etal>et al</etal>
</person-group><article-title>Association of glycated hemoglobin A<sub>1c</sub> levels with cardiovascular outcomes in the general population: results from the BiomarCaRE (Biomarker for Cardiovascular Risk Assessment in Europe) consortium</article-title><source>Cardiovasc Diabetol</source><month>Nov</month><day>15</day><year>2021</year><volume>20</volume><issue>1</issue><fpage>223</fpage><comment>doi</comment><pub-id pub-id-type="doi">10.1186/s12933-021-01413-4</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">34781939</pub-id><pub-id pub-id-type="pmcid">PMC8594211</pub-id></element-citation></ref><ref id="R27"><label>27.</label><element-citation publication-type="webpage"><article-title>EMETA study design and participation</article-title><source>BioRender</source><comment>URL</comment><ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://BioRender.com/d5aglhl" ext-link-type="uri">https://BioRender.com/d5aglhl</ext-link><comment>Accessed</comment><date-in-citation>13-03-2026</date-in-citation></element-citation></ref><ref id="R28"><label>28.</label><element-citation publication-type="webpage"><article-title>Attribution 4.0 international (CC BY 4.0)</article-title><source>Creative Commons</source><comment>URL</comment><ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://creativecommons.org/licenses/by/4.0/" ext-link-type="uri">https://creativecommons.org/licenses/by/4.0/</ext-link><comment>Accessed</comment><date-in-citation>13-03-2026</date-in-citation></element-citation></ref><ref id="R29"><label>29.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Dieli-Conwright</surname><given-names>CM</given-names></name>
<name name-style="western"><surname>Courneya</surname><given-names>KS</given-names></name>
<name name-style="western"><surname>Demark-Wahnefried</surname><given-names>W</given-names></name>
<etal>et al</etal>
</person-group><article-title>Effects of aerobic and resistance exercise on metabolic syndrome, sarcopenic obesity, and circulating biomarkers in overweight or obese survivors of breast cancer: a randomized controlled trial</article-title><source>J Clin Oncol</source><month>Mar</month><day>20</day><year>2018</year><volume>36</volume><issue>9</issue><fpage>875</fpage><lpage>883</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1200/JCO.2017.75.7526</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">29356607</pub-id><pub-id pub-id-type="pmcid">PMC5858524</pub-id></element-citation></ref><ref id="R30"><label>30.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Goyal</surname><given-names>A</given-names></name>
<name name-style="western"><surname>Milner</surname><given-names>GE</given-names></name>
<name name-style="western"><surname>Cimino-Mathews</surname><given-names>A</given-names></name>
<etal>et al</etal>
</person-group><article-title>Weight gain after hormone receptor-positive breast cancer</article-title><source>Curr Oncol</source><month>Jun</month><day>6</day><year>2022</year><volume>29</volume><issue>6</issue><fpage>4090</fpage><lpage>4103</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.3390/curroncol29060326</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">35735435</pub-id><pub-id pub-id-type="pmcid">PMC9222132</pub-id></element-citation></ref><ref id="R31"><label>31.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Heideman</surname><given-names>WH</given-names></name>
<name name-style="western"><surname>Russell</surname><given-names>NS</given-names></name>
<name name-style="western"><surname>Gundy</surname><given-names>C</given-names></name>
<name name-style="western"><surname>Rookus</surname><given-names>MA</given-names></name>
<name name-style="western"><surname>Voskuil</surname><given-names>DW</given-names></name>
</person-group><article-title>The frequency, magnitude and timing of post-diagnosis body weight gain in Dutch breast cancer survivors</article-title><source>Eur J Cancer</source><month>01</month><year>2009</year><volume>45</volume><issue>1</issue><fpage>119</fpage><lpage>126</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1016/j.ejca.2008.09.003</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">18930387</pub-id></element-citation></ref><ref id="R32"><label>32.</label><element-citation publication-type="journal"><person-group person-group-type="author">
<name name-style="western"><surname>Bergman</surname><given-names>R</given-names></name>
<name name-style="western"><surname>Berko</surname><given-names>YA</given-names></name>
<name name-style="western"><surname>Sanchez</surname><given-names>V</given-names></name>
<etal>et al</etal>
</person-group><article-title>Obesity and metabolic syndrome are associated with short-term endocrine therapy resistance in early ER + breast cancer</article-title><source>Breast Cancer Res Treat</source><month>01</month><year>2023</year><volume>197</volume><issue>2</issue><fpage>307</fpage><lpage>317</lpage><comment>doi</comment><pub-id pub-id-type="doi">10.1007/s10549-022-06794-y</pub-id><comment>Medline</comment><pub-id pub-id-type="pmid">36396775</pub-id><pub-id pub-id-type="pmcid">PMC10603601</pub-id></element-citation></ref></ref-list></back></article><article xml:lang="en" article-type="research-article" dtd-version="1.4"><processing-meta base-tagset="archiving" mathml-version="3.0" table-model="xhtml" tagset-family="jats"><restricted-by>pmc</restricted-by></processing-meta><front><journal-meta><journal-id journal-id-type="nlm-ta">Sci Adv</journal-id><journal-id journal-id-type="iso-abbrev">Sci Adv</journal-id><journal-id journal-id-type="pmc-domain-id">2850</journal-id><journal-id journal-id-type="pmc-domain">sciadv</journal-id><journal-id journal-id-type="publisher-id">sciadv</journal-id><journal-title-group><journal-title>Science Advances</journal-title></journal-title-group><issn pub-type="epub">2375-2548</issn><publisher><publisher-name>American Association for the Advancement of Science</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="pmcid">PMC13015887</article-id><article-id pub-id-type="pmcid-ver">PMC13015887.1</article-id><article-id pub-id-type="pmcaid">13015887</article-id><article-id pub-id-type="pmcaiid">13015887</article-id><article-id pub-id-type="pmid">41880514</article-id><article-id pub-id-type="doi">10.1126/sciadv.aec9175</article-id><article-id pub-id-type="publisher-id">aec9175</article-id><article-version article-version-type="pmc-version">1</article-version><article-categories><subj-group subj-group-type="article-type"><subject>Research Article</subject></subj-group><subj-group subj-group-type="heading"><subject>Physical and Materials Sciences</subject></subj-group><subj-group subj-group-type="legacy-article-type"><subject>SciAdv r-articles</subject></subj-group><subj-group subj-group-type="field"><subject>Biochemistry</subject><subject>Chemistry</subject></subj-group><subj-group subj-group-type="overline"><subject>Biochemistry</subject></subj-group></article-categories><title-group><article-title>Machine learning–directed massively parallel programmable nucleic acid amplification</article-title><alt-title alt-title-type="short">Machine learning–guided programmable amplification</alt-title></title-group><contrib-group><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0000-0003-0136-1978</contrib-id><name name-style="western"><surname>Weng</surname><given-names initials="Z">Zhi</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Conceptualization" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-conceptualization/">Conceptualization</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Data curation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-data-curation/">Data curation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Formal analysis" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-formal-analysis/">Formal analysis</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Investigation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-investigation/">Investigation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Methodology" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-methodology/">Methodology</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Resources" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-resources/">Resources</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Software" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-software/">Software</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Validation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-validation/">Validation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Visualization" vocab-term-identifier="https://web.archive.org/web/20180313224017/http://dictionary.credit.niso.org/Contributor_Roles/Visualization">Visualization</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - original draft" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-original-draft/">Writing - original draft</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff1" ref-type="aff">
<sup>1</sup>
</xref><xref rid="aff2" ref-type="aff">
<sup>2</sup>
</xref></contrib><contrib contrib-type="author"><name name-style="western"><surname>Huang</surname><given-names initials="W">Wenle</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Data curation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-data-curation/">Data curation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Validation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-validation/">Validation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff1" ref-type="aff">
<sup>1</sup>
</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0009-0003-1951-495X</contrib-id><name name-style="western"><surname>Wu</surname><given-names initials="Y">Yi</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Data curation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-data-curation/">Data curation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Software" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-software/">Software</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff1" ref-type="aff">
<sup>1</sup>
</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0000-0003-0799-6683</contrib-id><name name-style="western"><surname>Xiu</surname><given-names initials="X">Xuehao</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Data curation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-data-curation/">Data curation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Software" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-software/">Software</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff1" ref-type="aff">
<sup>1</sup>
</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0009-0007-4900-9021</contrib-id><name name-style="western"><surname>Lv</surname><given-names initials="H">Hui</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff3" ref-type="aff">
<sup>3</sup>
</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0000-0003-3948-6899</contrib-id><name name-style="western"><surname>Wang</surname><given-names initials="F">Fei</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff4" ref-type="aff">
<sup>4</sup>
</xref></contrib><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0000-0001-7505-2727</contrib-id><name name-style="western"><surname>Zuo</surname><given-names initials="X">Xiaolei</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Formal analysis" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-formal-analysis/">Formal analysis</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff4" ref-type="aff">
<sup>4</sup>
</xref><xref rid="aff5" ref-type="aff">
<sup>5</sup>
</xref></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0000-0002-7171-7338</contrib-id><name name-style="western"><surname>Fan</surname><given-names initials="C">Chunhai</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Conceptualization" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-conceptualization/">Conceptualization</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Funding acquisition" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-funding-acquisition/">Funding acquisition</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Investigation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-investigation/">Investigation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Project administration" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-project-administration/">Project administration</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Resources" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-resources/">Resources</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Supervision" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-supervision/">Supervision</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff4" ref-type="aff">
<sup>4</sup>
</xref><xref rid="cor1" ref-type="corresp">*</xref></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid" authenticated="true">https://orcid.org/0000-0001-5947-4604</contrib-id><name name-style="western"><surname>Song</surname><given-names initials="P">Ping</given-names></name><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Conceptualization" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-conceptualization/">Conceptualization</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Data curation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-data-curation/">Data curation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Formal analysis" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-formal-analysis/">Formal analysis</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Funding acquisition" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-funding-acquisition/">Funding acquisition</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Investigation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-investigation/">Investigation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Methodology" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-methodology/">Methodology</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Project administration" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-project-administration/">Project administration</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Resources" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-resources/">Resources</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Software" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-software/">Software</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Supervision" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-supervision/">Supervision</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Validation" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-validation/">Validation</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Visualization" vocab-term-identifier="https://web.archive.org/web/20180313224017/http://dictionary.credit.niso.org/Contributor_Roles/Visualization">Visualization</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - original draft" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-original-draft/">Writing - original draft</role><role vocab="credit" vocab-identifier="https://credit.niso.org/term/contributor-roles/" vocab-term="Writing - review & editing" vocab-term-identifier="https://credit.niso.org/term/contributor-roles-writing-review-editing/">Writing - review & editing</role><xref rid="aff1" ref-type="aff">
<sup>1</sup>
</xref><xref rid="aff2" ref-type="aff">
<sup>2</sup>
</xref><xref rid="cor2" ref-type="corresp">*</xref></contrib><aff id="aff1"><label><sup>1</sup></label>International Peace Maternity and Child Health Hospital, School of Medicine, Shanghai Jiao Tong University, National Center for Translational Medicine, Shanghai 200030, China.</aff><aff id="aff2"><label><sup>2</sup></label>School of Biomedical Engineering, Zhangjiang Institute for Advanced Study and National Center for Translational Medicine, Shanghai Jiao Tong University, Shanghai 200240, China.</aff><aff id="aff3"><label><sup>3</sup></label>Institute of Materiobiology, College of Sciences, Shanghai University, Shanghai 200444, China.</aff><aff id="aff4"><label><sup>4</sup></label>State Key Laboratory of Synergistic Chem-Bio Synthesis, School of Chemistry and Chemical Engineering, New Cornerstone Science Laboratory, Frontiers Science Center for Transformative Molecules, Zhangjiang Institute for Advanced Study and National Center for Translational Medicine, Shanghai Jiao Tong University, Shanghai 200240, China.</aff><aff id="aff5"><label><sup>5</sup></label>Institute of Molecular Medicine, Shanghai Key Laboratory for Nucleic Acids Chemistry and Nanomedicine, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200127, China.</aff></contrib-group><author-notes><corresp id="cor1"><label>*</label>Corresponding author. Email: <email xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="fanchunhai@sjtu.edu.cn">fanchunhai@sjtu.edu.cn</email> (C.F.);</corresp><corresp id="cor2"><label>*</label>Corresponding author. Email: <email xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="songpingsjtu@sjtu.edu.cn">songpingsjtu@sjtu.edu.cn</email> (P.S.)</corresp></author-notes><pub-date pub-type="collection"><day>27</day><month>3</month><year>2026</year></pub-date><pub-date publication-format="electronic" date-type="pub"><day>25</day><month>3</month><year>2026</year></pub-date><volume>12</volume><issue>13</issue><issue-id pub-id-type="pmc-issue-id">510199</issue-id><elocation-id>eaec9175</elocation-id><history><date date-type="received"><day>08</day><month>10</month><year>2025</year></date><date date-type="accepted"><day>24</day><month>2</month><year>2026</year></date></history><pub-history><event event-type="pmc-release"><date><day>25</day><month>03</month><year>2026</year></date></event><event event-type="pmc-live"><date><day>26</day><month>03</month><year>2026</year></date></event><event event-type="pmc-last-change"><date iso-8601-date="2026-03-26 10:25:14.117"><day>26</day><month>03</month><year>2026</year></date></event></pub-history><permissions><copyright-statement>Copyright © 2026 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution License 4.0 (CC BY).</copyright-statement><copyright-year>2026</copyright-year><copyright-holder>The Authors</copyright-holder><license><ali:license_ref xmlns:ali="http://www.niso.org/schemas/ali/1.0/" specific-use="textmining" content-type="ccbylicense" start_date="2026-03-25">https://creativecommons.org/licenses/by/4.0/</ali:license_ref><license-p>This is an open-access article distributed under the terms of the <ext-link xmlns:xlink="http://www.w3.org/1999/xlink" ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">Creative Commons Attribution license</ext-link>, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</license-p></license></permissions><self-uri xmlns:xlink="http://www.w3.org/1999/xlink" content-type="pmc-pdf" xlink:href="sciadv.aec9175.pdf"/><self-uri xmlns:xlink="http://www.w3.org/1999/xlink" content-type="pdf" xlink:href="sciadv.aec9175.pdf"/><abstract><p>Dynamic regulation of amplification efficiency is pivotal yet challenging in molecular diagnostics and DNA data storage. Here, we develop a thermodynamics-based approach to achieve continuous and precise modulation of nucleic acid amplification efficiency. By decoupling sequence specificity from hybridization energy regulation via a primer-tag compensation strategy, we demonstrate programmed amplification with high resolution (33 versus 81%). Leveraging 2483 experimental data, we constructed a machine learning model that improved prediction accuracy from <italic toggle="yes">R</italic><sup>2</sup> = 0.62 to = 0.86. In DNA data storage, this amplification strategy increases the density for information preview by nearly one order of magnitude and robust file steganography via differential amplification. In clinical validation, our method outperformed uniform amplification in cervical cancer RNA variant analysis, detecting rare RNA fusions and improving detection sensitivity by 100-fold under 10<sup>4</sup> simulated sequencing depth. This programmable technique is anticipated to extend to single-cell sequencing and spatial transcriptomics, offering a powerful tool for molecular diagnostics and synthetic biology.</p></abstract><abstract abstract-type="teaser"><p>Machine learning–guided tunable DNA amplification enables intelligent DNA data storage and ultrasensitive molecular diagnostics.</p></abstract><funding-group specific-use="FundRef"><award-group id="award2263796"><funding-source>
<institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/501100001809</institution-id><institution>National Natural Science Foundation of China</institution></institution-wrap>
</funding-source><award-id>22574102, 22174094</award-id></award-group><award-group id="award2263800"><funding-source>
<institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/501100003395</institution-id><institution>Shanghai Municipal Education Commission</institution></institution-wrap>
</funding-source><award-id>ZXWH1082101</award-id></award-group><award-group id="award2359075"><funding-source>
<institution-wrap><institution-id institution-id-type="FundRef">http://dx.doi.org/10.13039/501100012226</institution-id><institution>Fundamental Research Funds for the Central Universities</institution></institution-wrap>
</funding-source><award-id>YG2023QNA33, YG2025ZD28</award-id></award-group><award-group id="award2263794"><funding-source>
<institution-wrap><institution-id institution-id-type="FundRef"/><institution>National key R&D program of China</institution></institution-wrap>
</funding-source><award-id>2022YFF1201800</award-id></award-group><award-group id="award2456985"><funding-source>
<institution-wrap><institution-id institution-id-type="FundRef"/><institution>Shanghai Science and Technology Committee</institution></institution-wrap>
</funding-source><award-id>24Y22800300</award-id></award-group><award-group id="award2456995"><funding-source>
<institution-wrap><institution-id institution-id-type="FundRef"/><institution>SJTU Trans-med Awards Research</institution></institution-wrap>
</funding-source><award-id>20240202</award-id></award-group><award-group id="award2456996"><funding-source>
<institution-wrap><institution-id institution-id-type="FundRef"/><institution>Shanghai Pilot Program for Basic Research-Shanghai Jiao Tong University</institution></institution-wrap>
</funding-source><award-id>21TQ1400222</award-id></award-group></funding-group><custom-meta-group><custom-meta><meta-name>pmc-status-qastatus</meta-name><meta-value>0</meta-value></custom-meta><custom-meta><meta-name>pmc-status-live</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-status-embargo</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-status-released</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-open-access</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-olf</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-manuscript</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-legally-suppressed</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-pdf</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-has-supplement</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-pdf-only</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-suppress-copyright</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-real-version</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-is-scanned-article</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-preprint</meta-name><meta-value>no</meta-value></custom-meta><custom-meta><meta-name>pmc-prop-in-epmc</meta-name><meta-value>yes</meta-value></custom-meta><custom-meta><meta-name>pmc-license-ref</meta-name><meta-value>CC BY</meta-value></custom-meta></custom-meta-group></article-meta></front><body><sec sec-type="introduction" disp-level="1"><title>INTRODUCTION</title><p>Nucleic acid amplification is indispensable to molecular diagnostics (<xref rid="R1" ref-type="bibr"><italic toggle="yes">1</italic></xref>–<xref rid="R3" ref-type="bibr"><italic toggle="yes">3</italic></xref>), genome sequencing (<xref rid="R4" ref-type="bibr"><italic toggle="yes">4</italic></xref>–<xref rid="R7" ref-type="bibr"><italic toggle="yes">7</italic></xref>), and emerging DNA-based data storage (<xref rid="R8" ref-type="bibr"><italic toggle="yes">8</italic></xref>–<xref rid="R11" ref-type="bibr"><italic toggle="yes">11</italic></xref>). However, its further advancement is bottlenecked by the limited flexibly and accurately with which amplification efficiency can be tuned across heterogeneous templates (<xref rid="R12" ref-type="bibr"><italic toggle="yes">12</italic></xref>, <xref rid="R13" ref-type="bibr"><italic toggle="yes">13</italic></xref>). Current high-throughput protocols still operate in a “one-size-fits-all” mode: applying identical primers, uniform thermal conditions, and fixed enzyme concentrations to every target, inevitably sacrificing sensitivity for specificity or vice versa (<xref rid="R14" ref-type="bibr"><italic toggle="yes">14</italic></xref>–<xref rid="R16" ref-type="bibr"><italic toggle="yes">16</italic></xref>). Conventional strategies that modulate efficiency by globally shortening or lengthening the same primer set result in switch-like responses: A single-base elongation or truncation can shift yields from undetectable to full saturation (<xref rid="R16" ref-type="bibr"><italic toggle="yes">16</italic></xref>). This binary behavior precludes programmable, template-specific tuning and cascades into poor detection limits, uneven coverage, and elevated error rates, particularly in massively parallel reactions.</p><p>The persistent binary nature of amplification becomes clearer when contrasted with nucleic acid hybridization. The same thermodynamic principles that transformed hybridization from an all-or-none process into a continuously tunable one could, in principle, be applied to polymerase chain reaction (PCR) (<xref rid="R17" ref-type="bibr"><italic toggle="yes">17</italic></xref>–<xref rid="R19" ref-type="bibr"><italic toggle="yes">19</italic></xref>). However, applying these principles to PCR remains challenging. First, amplification yield is governed by an extremely narrow parameter window: A single-base change in primer-template complementarity can shift yields from 0 to more than 90%, resulting in a sharp, nearly step-wise response curve. Second, the reaction couples primer-template hybridization with enzyme kinetics, deoxynucleoside triphosphate (dNTP) availability, and product inhibition, creating a high-dimensional parameter space that resists straightforward predictive models. Collectively, these factors have hindered the development of general framework for programmable and template-specific amplification.</p><p>To the best of our knowledge, no effective method for continuous and tunable control of amplification efficiency has been reported to date. Here, we introduce a machine learning–directed approach for massively parallel programmed nucleic acid amplification. We first developed a tag-primer architecture that incorporates an energy-compensating sequence to enable stepwise regression of amplification efficiency. This design mitigates the impact of initial sequence adjustments and substantially broadens the dynamic range of amplification, with hybridization reaction standard free energy change (Δ<italic toggle="yes">G</italic>°) expanded from 1.5 to 4.0 kcal mol<sup>−1</sup>. Using these tag primers as regulatory units, we constructed an ensemble learning model capable of accurately predicting and precisely controlling amplification yield. As a proof of concept, we applied our programmable amplification (PA) strategy to DNA data storage. By differentially modulating the amplification efficiency of primers during file random access, we achieved template-specific amplification, successfully implementing multilevel access control in DNA storage systems and advancing toward intelligent data retrieval. Furthermore, in clinical validation through cervical cancer RNA variant analysis, our method substantially outperformed uniform amplification, detecting rare RNA fusions while improving detection sensitivity by 100-fold.</p></sec><sec sec-type="results" disp-level="1"><title>RESULTS</title><sec disp-level="2"><title>PA overview and prediction model for amplification</title><p>On-demand multitemplate amplification is becoming increasingly important in complex sample analysis (<xref rid="R12" ref-type="bibr"><italic toggle="yes">12</italic></xref>, <xref rid="R20" ref-type="bibr"><italic toggle="yes">20</italic></xref>). Traditional amplification (TA) methods use uniform primers that produce consistent yields but offer limited flexibility in adjusting the relative proportions of different targets. This often leads to the underdetection of low-abundance species. In contrast, we developed PA to enable adjustable amplification efficiency for different targets, achieving on-demand amplification of multiple targets while maintaining precise analysis, particularly for low-abundance targets (<xref rid="F1" ref-type="fig">Fig. 1A</xref>). The process primarily consists of two stages: molecular recognition involving hybridization between nucleic acid primers and templates, which determines reaction specificity (<xref rid="F1" ref-type="fig">Fig. 1B</xref>). The hybridization standard free energy (Δ<italic toggle="yes">G</italic>) serves as an indicator of reaction efficiency and specificity (<xref rid="R16" ref-type="bibr"><italic toggle="yes">16</italic></xref>, <xref rid="R18" ref-type="bibr"><italic toggle="yes">18</italic></xref>, <xref rid="R21" ref-type="bibr"><italic toggle="yes">21</italic></xref>), influenced by multiple factors including sequence information, temperature, and salt concentration—all of which contribute to subsequent model training. The second stage involves enzymatic amplification after primer-template binding, which determines the enrichment yield for subsequent hybridization and amplification. These two processes are progressively amplified over multiple PCR cycles, resulting in gradually diverging amplification efficiencies (<xref rid="F1" ref-type="fig">Fig. 1, C and D</xref>). Of note, amplification efficiency denotes the fraction of template molecules duplicated per cycle with a theoretical maximum efficiency of 100%. The equivalent amplification efficiency represents the mean efficiency averaged across all cycles. Amplification yield is defined as the ratio of the actual product quantity to the theoretical maximum product quantity attainable after a given number of cycles.</p><fig position="float" id="F1" fig-type="image" specific-use="distribute" orientation="portrait"><label>Fig. 1.</label><caption><title>Overview of PA.</title><p>(<bold>A</bold>) In conventional multiplex amplification systems, uniform amplification often leads to loss of low-frequency templates. In contrast, PA enables template-specific amplification tuning, facilitating both enrichment of low-abundance templates and on-demand multitarget analysis. (<bold>B</bold>) Workflow of PA experiment, comprising two key steps: signal recognition (nucleic acid hybridization) and enrichment step (enzymatic amplification). (<bold>C</bold>) Relationship between equivalent amplification efficiency (abbreviated Equivalent ampl. eff.) and primer hybridization thermodynamics in conventional methods. The narrow ∆<italic toggle="yes">G</italic> tuning range results in abrupt efficiency changes upon single-base modifications. (<bold>D</bold>) Thermodynamic regulation in PA shows an expanded ∆<italic toggle="yes">G</italic> tuning range, allowing precise programmability of amplification efficiency. (<bold>E</bold>) Schematic of the machine learning–assisted efficiency prediction model for PA.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="sciadv.aec9175-f1.jpg"/></fig><p>Notably, conventional amplification systems show limited tunability when attempting to modulate efficiency through primer shortening, with single-base modifications often causing abrupt efficiency drops (<xref rid="F1" ref-type="fig">Fig. 1C</xref>). To overcome this limitation, our PA system incorporates a tag-mediated regulation mechanism that sustains high equivalent amplification efficiency even under suboptimal Δ<italic toggle="yes">G</italic> conditions, substantially broadening the available dynamic range (<xref rid="F1" ref-type="fig">Fig. 1D</xref>). Furthermore, to accurately predict equivalent amplification efficiency and meet the requirements for on-demand multitemplate amplification, we further developed a machine learning model incorporating multiple parameters of primer hybridization and amplification, such as ∆<italic toggle="yes">G</italic>, reaction time, temperature, and GC content, enabling accurate prediction of primer amplification yield (<xref rid="F1" ref-type="fig">Fig. 1E</xref>). This comprehensive approach not only solves the problem of low-abundance target detection but also establishes a framework for precision-controlled nucleic acid amplification.</p></sec><sec disp-level="2"><title>Theoretical modeling and validation of PA</title><p>To address the limited dynamic range and poor tunability of conventional primers, we developed a tag-primer design for PA. The key innovation involves incorporating a noncomplementary sequence at the 5′ terminus that provides progressive energetic compensation over amplification cycles, thereby enabling precise control of total amplification yield. This system gives rise to two template types: the initial short template (ST) and the extended long template (LT). Because of differential binding affinities, tag primers exhibit distinct amplification efficiencies for these templates. As cycles progress, LT proportion gradually increases, causing the overall amplification efficiency to asymptotically approach 100% (<xref rid="F2" ref-type="fig">Fig. 2A</xref>). This efficiency restoration phenomenon mitigates the impact of single-base variations, substantially expanding the reaction’s dynamic range for fine-tuned amplification control.</p><fig position="float" id="F2" fig-type="image" specific-use="distribute" orientation="portrait"><label>Fig. 2.</label><caption><title>Theoretical modeling of PA.</title><p>(<bold>A</bold>) Schematic of PA using tag primer. (<bold>B</bold>) Ratio variation between LT and ST and (<bold>C</bold>) corresponding equivalent amplification efficiency across PCR cycles, demonstrating results for primers with initial Δ<italic toggle="yes">G</italic>° = −6 kcal mol<sup>−1</sup> and tag Δ<italic toggle="yes">G</italic>° = −12 kcal mol<sup>−1</sup>. As cycle number increases, LT proportion progressively rises. Because of stronger binding between LT and primers, the per-cycle equivalent amplification efficiency gradually “rebounds” to normal levels. As a result, the tag-primer system is equivalent to a conventional system (both FP and RP) with 67% efficiency (black dashed line). This “efficiency rebound” mechanism mitigates the impact of initial Δ<italic toggle="yes">G</italic> adjustments, thereby expanding the primer’s dynamic regulation range. (<bold>D</bold>) Equivalent amplification efficiency of primers with varying initial Δ<italic toggle="yes">G</italic>° and tag Δ<italic toggle="yes">G</italic>° values. More negative tag Δ<italic toggle="yes">G</italic>° values correspond to greater dynamic range expansion. (<bold>E</bold>) Comparison between tag primers (tag Δ<italic toggle="yes">G</italic>° = −12 kcal mol<sup>−1</sup>) and conventional primers across different initial Δ<italic toggle="yes">G</italic>° values, showing substantially broader dynamic range for tag primers (defined as the <italic toggle="yes">x</italic>-axis range corresponding to 20 to 80% efficiency values: blue, 1.75; green, 4.0). (<bold>F</bold>) Free energy–efficiency response curves and (<bold>G</bold>) dynamic range comparison for tag primers versus conventional primers at different temperatures. Data were fitted from 1000 sequences accounting for sequence-specific Δ<italic toggle="yes">G</italic>° temperature dependencies. Tag primers consistently exhibit larger dynamic ranges across temperatures. (<bold>H</bold>) Free energy–efficiency response curves and (<bold>I</bold>) dynamic range comparison at different extension times. Tag primers maintain superior dynamic ranges regardless of extension time.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="sciadv.aec9175-f2.jpg"/></fig><p>We first established a theoretical model integrating primer hybridization and polymerase extension to validate the feasibility of tag primer (Supplementary Text and figs. S1 to S3). For a primer with an initial standard binding free energy Δ<italic toggle="yes">G</italic>° = −6 kcal mol<sup>−1</sup> (calculated at 60°C, 0.18 M Na<sup>+</sup>) and a tag region with initial standard binding energy of Δ<italic toggle="yes">G</italic>° = −12 kcal mol<sup>−1</sup>, the proportion of LT increased from 0% at cycle 0 to 25% by cycle 7, exceeding 90% after cycle 13 (<xref rid="F2" ref-type="fig">Fig. 2B</xref>). Correspondingly, the amplification efficiency rose from near 0 to 100%, matching the cumulative output of a conventional primer with 67% efficiency over 15 cycles (<xref rid="F2" ref-type="fig">Fig. 2C</xref> and fig. S4). Systematic simulations revealed that rational design could achieve any desired efficiency within this continuum (<xref rid="F2" ref-type="fig">Fig. 2D</xref>). Notably, increasing the binding strength of the tag Δ<italic toggle="yes">G</italic> substantially expanded the dynamic range compared to conventional primers (tag Δ<italic toggle="yes">G</italic>° = 0 kcal mol<sup>−1</sup>; <xref rid="F2" ref-type="fig">Fig. 2, D and E</xref>).</p><p>Additional parameters like annealing temperature and extension time also modulate amplification efficiency (fig. S5). We therefore systematically evaluated their effects on dynamic range. Temperature primarily affects sequence-specific Δ<italic toggle="yes">G</italic> (fig. S6). Using Δ<italic toggle="yes">G</italic>° referenced to 60°C, we simulated efficiency variations across temperatures. While temperature dramatically shifted equivalent amplification efficiency curves (e.g., 57.8% at 60°C versus 95.6% at 55°C versus 15.2% at 65°C), it minimally affected conventional primers’ dynamic range. Tag primers showed slight range expansion in dynamic range at elevated temperatures (<xref rid="F2" ref-type="fig">Fig. 2F</xref>) and consistently outperformed conventional primers across all conditions (<xref rid="F2" ref-type="fig">Fig. 2G</xref> and fig. S7). Extension time exerted a weaker influence on both efficiency distribution (<xref rid="F2" ref-type="fig">Fig. 2H</xref>) and dynamic range (<xref rid="F2" ref-type="fig">Fig. 2I</xref>), although tag primers maintained superior tunability across all conditions.</p></sec><sec disp-level="2"><title>Experimental validation and performance characterization of engineered PA</title><p>Following theoretical validation of primer-tag feasibility, we compiled experimental data from 2483 independent reaction records (TA: <italic toggle="yes">n</italic> = 1255; PA: <italic toggle="yes">n</italic> = 1228) to compare both primer systems. Data were generated by targeting multiple plasmid sites using various primer with varied configurations (with/without tags, sequence truncations) under different reaction protocols. To streamline subsequent analysis and prediction, we typically modified either the forward (FP) or reverse primer (RP) while maintaining the other as a conventional primer (Δ<italic toggle="yes">G</italic>° ≈ −12 kcal mol<sup>−1</sup>, efficiency ≈ 100%), using cycle threshold (Ct) values as a proxy for amplification yield (<xref rid="F3" ref-type="fig">Fig. 3A</xref>).</p><fig position="float" id="F3" fig-type="image" specific-use="distribute" orientation="portrait"><label>Fig. 3.</label><caption><title>Experimental validation and performance analysis of tag-primer designs.</title><p>(<bold>A</bold>) Amplification data collection using quantified plasmids (6000 copies/μl) with various primer designs and reaction protocols. Reactions showing no amplification within 50 cycles were recorded as Ct = 50. (<bold>B</bold>) Comparative efficiency modulation between tag primers and conventional primers through sequence truncation. Representative primer sets (including both types) were progressively shortened and tested at different temperatures (Ct values shown). The right panel displays the shared template-binding region. Conventional primers exhibit abrupt Ct transitions (green to purple), while tag primers demonstrate gradual, tunable efficiency shifts through combined sequence truncation and temperature adjustment. (<bold>C</bold>) Efficiency distributions versus free energy (calculated at 55° to 63°C). Tag primers show substantially broader dynamic ranges. (<bold>D</bold>) Ct distributions of conventional versus tag-primer reactions. Conventional primers yield bimodal extremes (low or high efficiency), whereas tag primers produce a near-uniform continuum across intermediate Ct values, indicating programmable, gradient-level amplification control. (<bold>E</bold>) Temperature-dependent Ct distributions. Higher temperatures induce right-skewed distributions due to reduced amplification efficiency. (<bold>F</bold>) Extension time effects. Longer extensions marginally improve efficiency (lower Ct values). For clarity, only data with Ct values greater than 25 at 30 s are shown for comparison. (<bold>G</bold>) Parameter influence ranking via linear regression. Tag presence dominates efficiency modulation, followed by free energy and temperature.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="sciadv.aec9175-f3.jpg"/></fig><p>Conventional primers exhibited drastic efficiency reduction upon single-base truncations (theoretical median efficiency drop: 81%), with experimental Ct values showing discontinuous jumps [e.g., 22 (green) to 50 (purple) at 60°C]. In contrast, tag primers enabled gradual efficiency modulation—Each truncation step induced modest Ct increments (median ΔCt ≈ 2 cycles), with a theoretical median efficiency drop of 33% (<xref rid="F3" ref-type="fig">Fig. 3B</xref> and figs. S8 and S9). Curve fitting of Ct values against Δ<italic toggle="yes">G</italic> revealed that tag primers achieved a 4.0 kcal mol<sup>−1</sup> dynamic range with 2.7-fold wider than conventional primers (1.5 kcal mol<sup>−1</sup>; <xref rid="F3" ref-type="fig">Fig. 3C</xref> and figs. S10 and S11). This expanded tunability was further reflected in Ct value distributions: Conventional primers clustered at extreme values (Ct = 22 or 50), whereas tag primers showed a uniform distribution across intermediate Ct values (<xref rid="F3" ref-type="fig">Fig. 3D</xref>), further confirming precise programmable control. To confirm the generalizability of tag strategy, we evaluated four distinct tag sequences with different GC content across multiple temperatures and primer lengths (fig. S12). All four tag sequences demonstrated consistent continuous tunability, maintaining gradual Ct modulation across different temperatures despite their sequence diversity. This robustness confirms that the PA strategy is sequence-independent and relies primarily on thermodynamic principles rather than specific sequence features, establishing broad applicability for primer design across diverse sequence contexts.</p><p>Beyond programmable tunability, tag primers demonstrated superior specificity in discriminating single-nucleotide variations. Systematic evaluation across three amplicons with primers containing single-base mismatches at varying positions revealed that tag primers exhibited substantially larger ΔCt differences between matched and mismatched scenarios compared to conventional primers (mean ΔCt: 12.0 versus 3.2 cycles; fig. S13). This enhanced discrimination arises from the truncated binding region in tag primers, which amplifies the energetic penalty of single-base mismatches that compounds exponentially over iterative PCR cycles, enabling robust discrimination between highly similar templates in multiplexed applications.</p><p>Having established tag primers’ tunability and specificity, we further characterized how reaction conditions affect amplification performance beyond sequence design. Temperature-dependent analysis across the full dataset indicated that higher temperatures reduced amplification efficiency, manifesting as right-skewed Ct distributions with elevated means, primarily attributable to the Δ<italic toggle="yes">G</italic> = Δ<italic toggle="yes">H</italic> – <italic toggle="yes">T</italic>Δ<italic toggle="yes">S</italic> relationship (<xref rid="F3" ref-type="fig">Fig. 3E</xref> and figs. S14 and S15). In contrast, longer extension times marginally reduced Ct values (<xref rid="F3" ref-type="fig">Fig. 3F</xref> and fig. S16).</p><p>To quantitatively evaluate the relative contributions of different parameters to Ct values, we performed linear regression analysis to determine weighting coefficients. Tag incorporation showed the strongest effect, reducing Ct by 6.1 cycles compared to conventional primers. Thermodynamic stability (Δ<italic toggle="yes">G</italic>) exerted the next largest influence, increasing Ct by 3.7 ± 0.1 per kcal mol<sup>−1</sup>. Temperature demonstrated moderate impact (1.3 ± 0.1 Ct elevation per °C), while extension time had the weakest effect (0.02 ± 0.01 Ct decrease per second; equivalent to 1.8 cycles from 30 to 120 s). Separate regression analyses revealed attenuated parametric influences in tag primers compared to conventional designs, further confirming tag primers’ enhanced tunability (<xref rid="F3" ref-type="fig">Fig. 3G</xref> and figs. S17 and S18).</p></sec><sec disp-level="2"><title>Precise prediction of amplification efficiency using ensemble learning models</title><p>Accurate prediction of amplification efficiency is essential to guide primer design and minimize experimental optimization (<xref rid="R19" ref-type="bibr"><italic toggle="yes">19</italic></xref>, <xref rid="R22" ref-type="bibr"><italic toggle="yes">22</italic></xref>, <xref rid="R23" ref-type="bibr"><italic toggle="yes">23</italic></xref>). However, conventional theoretical models that primarily rely on global thermodynamic parameters such as Δ<italic toggle="yes">G</italic>°, temperature, and extension time often exhibit poor predictive performance. In our preliminary analyses, primers with identical Δ<italic toggle="yes">G</italic>° under the same reaction conditions exhibited substantial variation in amplification efficiency, highlighting the importance of sequence-specific features beyond simplified thermodynamic frameworks (<xref rid="F3" ref-type="fig">Fig. 3B</xref>). To address this, we developed a machine learning model integrating diverse predictive factors, which include thermodynamic parameters, primer sequence biophysical characteristics, and global reaction conditions. By combining these heterogeneous features with tag-primer designs, our approach substantially improved the prediction precision (<xref rid="F4" ref-type="fig">Fig. 4A</xref> and figs. S19 and S20).</p><fig position="float" id="F4" fig-type="image" specific-use="distribute" orientation="portrait"><label>Fig. 4.</label><caption><title>Machine learning models for predicting amplification efficiency.</title><p>(<bold>A</bold>) Architecture of the ensemble learning–based prediction model. Sequence-specific biophysical features and global reaction parameters (temperature and time) were input into multiple base models, with meta-model stacking generating the final Ct value predictions. (<bold>B</bold>) Pareto plot comparing root mean square error (RMSE) and prediction time across models. The selected stacked model (black circle) achieved the lowest RMSE. (<bold>C</bold>) Performance comparison of different models across datasets. The “With tag” dataset showed superior prediction accuracy for all models, with the stacked model performing best within each dataset. (<bold>D</bold>) Scatter plot of predicted versus experimental values for the stacked model (<italic toggle="yes">r</italic><sup>2</sup> = 0.862). (<bold>E</bold>) Residual distribution with yellow dashed lines indicating the 80% confidence interval. Most predictions fell within ±4.5-cycle error range. (<bold>F</bold>) Feature importance ranking and (<bold>G</bold>) SHAP analysis. Thermodynamic parameters (<italic toggle="yes">T</italic><sub>m</sub>, temperature, and binding probability) dominated predictive power. (<bold>H</bold>) Feature correlation analysis showing strong <italic toggle="yes">T</italic><sub>m</sub>-Δ<italic toggle="yes">G</italic> interdependence, explaining Δ<italic toggle="yes">G</italic>’s lower independent ranking due to information redundancy.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="sciadv.aec9175-f4.jpg"/></fig><p>In our framework, we used an ensemble learning architecture that leverages multiple base learners, including generalized linear models (GLM), gradient boosting machines (GBM), random forest (RF), and XGBoost. In the first tier, multiple submodels were trained for each algorithm type using 70% of the dataset (TA: <italic toggle="yes">n</italic> = 1255; PA: <italic toggle="yes">n</italic> = 1228). Optimal performers or randomly selected subsets were then used to generate meta-features for the second-tier stacked model (<xref rid="F4" ref-type="fig">Fig. 4A</xref>). This ensemble approach effectively compensated for the limitations of individual models, achieving superior accuracy with the lowest root mean square error among all tested methods (<xref rid="F4" ref-type="fig">Fig. 4B</xref>) while maintaining practical inference speeds (~0.5 ms per sequence). Most notably, when combined with tag-primer designs, the model achieved a prediction precision of <italic toggle="yes">R</italic><sup>2</sup> = 0.862 compared to 0.624 for conventional theoretical models (<xref rid="F4" ref-type="fig">Fig. 4, C and D</xref>, and figs. S21 to S23). Residual analysis further confirmed that 80% of predictions fell within ±4.5 cycles of experimental values (<xref rid="F4" ref-type="fig">Fig. 4E</xref> and fig. S24), establishing a robust foundation for precision amplification control.</p><p>Feature importance analysis identified key thermodynamic parameters as the dominant predictors across both base learners and the stacked model. These included melting temperature (<italic toggle="yes">T</italic><sub>m</sub>), reaction temperature, binding probability, and Δ<italic toggle="yes">G</italic>°, all consistently contributing the strongest predictive power (<xref rid="F4" ref-type="fig">Fig. 4F</xref> and figs. S25 and S26). This aligns with the kinetic saturation principle: Given sufficient reaction time (30 to 120 s), thermodynamic stability ultimately governs amplification efficiency. SHAP (SHapley Additive exPlanations) value analysis further validated thermodynamic dominance of thermodynamics in model decisions (<xref rid="F4" ref-type="fig">Fig. 4G</xref>). The reduced independent contribution of ∆<italic toggle="yes">G</italic> may be due to information overlap with correlated features like <italic toggle="yes">T</italic><sub>m</sub>. This interpretation was corroborated by strong <italic toggle="yes">T</italic><sub>m</sub>-Δ<italic toggle="yes">G</italic>° correlation in SHAP-based feature dependence plots (<xref rid="F4" ref-type="fig">Fig. 4H</xref>).</p></sec><sec disp-level="2"><title>Development of an intelligent DNA storage operating system</title><p>DNA storage has emerged as a transformative solution for escalating data storage demands, leveraging DNA’s exceptional attributes including ultrahigh storage density, extended half-life, and minimal energy requirements (<xref rid="R24" ref-type="bibr"><italic toggle="yes">24</italic></xref>–<xref rid="R26" ref-type="bibr"><italic toggle="yes">26</italic></xref>). Nucleic acid amplification serves as a pivotal technology for file random access within this framework. The inherent complexity of DNA-encoded binary sequences presents unique challenges for amplification fidelity, making it an ideal test bed for our methodology. We have adapted our approach not only for conventional file retrieval but also engineered an intelligent access system capable of user-customized output (<xref rid="F5" ref-type="fig">Fig. 5A</xref>).</p><fig position="float" id="F5" fig-type="image" specific-use="distribute" orientation="portrait"><label>Fig. 5.</label><caption><title>Machine learning (ML)–guided intelligent DNA storage system with personalized amplification control.</title><p>(<bold>A</bold>) Schematic of the intelligent DNA storage operating system (map image credit: ICAO GIS). Under ML model guidance, the system modulates file-specific amplification efficiencies through optimized primer combinations and reaction condition adjustments, enabling differential molecular copy number control and customized readout outcomes within a single access operation. (<bold>B</bold> to <bold>D</bold>) File-specific (B) sequencing coverage, (C) decoding results, and (D) decoding accuracy under different primer combinations (green circles: high-efficiency primers; gray circles: low-efficiency primers). Results demonstrate programmable readout customization through primer efficiency regulation. (<bold>E</bold>) Continuous image clarity modulation via primer efficiency adjustment. (<bold>F</bold>) Cost-throughput optimization through preview-level control. Simulation assumes 25 M reads per run (1 M reads required for full file decoding), showing how PA increases throughput while reducing per-file costs at fixed sequencing capacity. (<bold>G</bold>) Condition-dependent file encryption/retrieval mechanism. Identical primers exhibit divergent amplification efficiencies under different reaction conditions, enabling cryptographic functionality. (<bold>H</bold>) Experimental validation showing file readout (left: 63°C, 30 s; right: 50°C, 120 s) with corresponding decoding accuracies.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="sciadv.aec9175-f5.jpg"/></fig><p>The system operates through strategic primer design and reaction condition optimization guided by our predictive models. By differentially regulating amplification efficiencies across files, we generate distinct molecular copy number distributions that translate into variable sequencing coverage. This coverage modulation enables (i) simultaneous multifile access with customized readout quality and (ii) intelligent resource allocation based on user-defined priorities (<xref rid="F5" ref-type="fig">Fig. 5, B to D</xref>). Experimental validation demonstrated that high-efficiency primers produced near-perfect decoding accuracy (≈100%), while low-efficiency primers generated “preview-mode” outputs (≈50% accuracy) at identical sequencing depths—effectively enabling tiered data access within a single run, with preview mode enhancing storage density by approximately an order of magnitude.</p><p>Further investigation of a fixed dual-file system revealed continuous clarity modulation (83 to 23% accuracy) through gradual primer efficiency reduction (<xref rid="F5" ref-type="fig">Fig. 5E</xref>). This precision control enables dramatic resource optimization: preview-quality images required two orders-of-magnitude less sequencing coverage than full-resolution decoding under fixed total sequencing capacity (<xref rid="F5" ref-type="fig">Fig. 5F</xref>). While extending this strategy to multiplex scenarios, increasing file numbers minimally affects PA’s tuning capability. However, the corresponding growth in primer pool size elevates cross-reactivity risks, particularly dimer formation, which may compromise modulation precision (fig. S27). Nevertheless, these challenges can be mitigated through stringent dimer screening and primer orthogonality validation during design.</p><p>Beyond primer-based control, we implemented condition-dependent file encryption through thermodynamic tuning. Designed “hidden” files exhibited near-zero amplification (7% accuracy) under standard conditions (63°C, 30s) but achieved normal detection (60% accuracy) when switched to optimized parameters (50°C, 120 s) (<xref rid="F5" ref-type="fig">Fig. 5, G to H</xref>). Although decoded images showed partial quality reduction, the stark contrast between conditions conclusively established condition-triggered file access as a viable paradigm.</p></sec><sec disp-level="2"><title>PA for low-abundance RNA variant analysis</title><p>High-throughput sequencing has become a cornerstone technology in molecular diagnostics, with its exceptional sensitivity, throughput, and multiplexing capacity enabling widespread clinical applications in disease detection and precision medicine (<xref rid="R27" ref-type="bibr"><italic toggle="yes">27</italic></xref>, <xref rid="R28" ref-type="bibr"><italic toggle="yes">28</italic></xref>). Among these techniques, targeted sequencing demonstrates particular value for RNA fusion detection and mutation screening due to its superior sequencing depth, specificity, and cost-effectiveness (<xref rid="R29" ref-type="bibr"><italic toggle="yes">29</italic></xref>, <xref rid="R30" ref-type="bibr"><italic toggle="yes">30</italic></xref>).</p><p>A critical challenge emerges from the substantial variation in RNA expression levels across genes, often spanning several orders of magnitude in clinical samples. Under uniform sequencing depth, this expression imbalance leads to inadequate coverage of low-abundance transcripts, increasing false-negative rates and compromising detection accuracy (<xref rid="F6" ref-type="fig">Fig. 6A</xref>, left). Our tunable amplification strategy addresses this limitation by selectively modulating primer efficiencies to enhance representation of low-expression genes. This approach achieves balanced detection limits across targets without increasing total sequencing depth, substantially improving both sensitivity and diagnostic reliability (<xref rid="F6" ref-type="fig">Fig. 6A</xref>, right).</p><fig position="float" id="F6" fig-type="image" specific-use="distribute" orientation="portrait"><label>Fig. 6.</label><caption><title>Application of PA in low-abundance RNA variant analysis.</title><p>(<bold>A</bold>) Comparison of sequencing outcomes between different amplification methods. Conventional uniform amplification (left) results in overrepresentation of highly expressed reference genes, limiting detection of low-abundance targets due to read resource competition. In contrast, our tunable amplification strategy (right) enhances target gene representation through primer efficiency modulation, enabling effective detection at lower sequencing depths. (<bold>B</bold>) Circos plot of RNA sequencing data from clinical samples, demonstrating expression level variations spanning multiple orders of magnitude. (<bold>C</bold>) Read distribution patterns comparing conventional and tunable amplification methods across sequencing depths, using <italic toggle="yes">FGFR3</italic> (target) and <italic toggle="yes">ACTB</italic> (reference) for <italic toggle="yes">FGFR3-TACC3</italic> fusion detection. The tunable strategy substantially improves <italic toggle="yes">FGFR3</italic>’s relative abundance in sequencing data. (<bold>D</bold>) Detection sensitivity analysis across sequencing depths and fusion frequencies. The tunable approach achieves ≈100-fold lower detection limits at 10<sup>4</sup> sequencing depths compared to conventional methods.</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" position="float" orientation="portrait" xlink:href="sciadv.aec9175-f6.jpg"/></fig><p>We validated this system using clinical patient samples exhibiting natural expression variation (<xref rid="F6" ref-type="fig">Fig. 6B</xref>). Focusing on FGFR3—a transmembrane tyrosine kinase receptor whose variants are implicated in cervical cancer, bladder cancer, and other malignancies (<xref rid="R31" ref-type="bibr"><italic toggle="yes">31</italic></xref>, <xref rid="R32" ref-type="bibr"><italic toggle="yes">32</italic></xref>)—we evaluated detection of <italic toggle="yes">FGFR3-TACC3</italic> fusion events using <italic toggle="yes">ACTB</italic> (a classical housekeeping gene) as reference. The 26-fold expression disparity between <italic toggle="yes">FGFR3</italic> and <italic toggle="yes">ACTB</italic> resulted in <italic toggle="yes">FGFR3</italic> capturing only 3.7% of sequencing reads under conventional amplification, with <italic toggle="yes">ACTB</italic> dominating the profile (<xref rid="F6" ref-type="fig">Fig. 6C</xref>). At a simulated sequencing depth of 10,000×, the detection limit for TA was 10%, whereas our PA strategy achieved a detection limit of 0.1%, representing a 100-fold sensitivity improvement. Even at 100,000× sequencing depth, the minimum detectable fusion frequency of TA remained 1%, far exceeding the clinically required 0.03% threshold (<xref rid="F6" ref-type="fig">Fig. 6D</xref>, left). Implementation of our PA strategy equalized gene representation, enabling reliable detection of 0.03% fusion events at equivalent sequencing depth and without additional costs (<xref rid="F6" ref-type="fig">Fig. 6D</xref>, right).</p></sec></sec><sec sec-type="discussion" disp-level="1"><title>DISCUSSION</title><p>Our programmable nucleic acid amplification system enables precise control over amplification efficiency for different nucleic acid templates, representing a substantial advance over conventional uniform amplification methods. By combining this tunability with our machine learning prediction model, we have developed an amplification technology with broad application prospects across multiple fields.</p><p>In DNA data storage, our intelligent access system provides practical value by enabling differential readout of multiple files within a single access operation. Guided by machine learning predictions, the system regulates amplification efficiency to achieve customized file retrieval while optimizing resource utilization and enhancing information security. Notably, our approach generates variable-resolution outputs from a single high-definition file, eliminating the need to store separate preview versions. This design simplification improves system compatibility and scalability. Furthermore, continuous fine-tuning of amplification efficiency allows smooth gradation of readout clarity from full resolution to complete concealment, with flexible control through either primer design optimization or reaction condition adjustment.</p><p>The PA approach also demonstrates critical utility in clinical diagnostics, particularly for low-abundance RNA variants analysis. As evidenced in our <italic toggle="yes">FGFR3-TACC3</italic> fusion case study, this technology enables sensitive detection of rare RNA variants (down to 0.03% frequency) without increasing sequencing depth. This capability addresses a fundamental limitation in current RNA sequencing applications including bulk sequencing, single-cell analysis, and spatial transcriptomics, where expression level variations often span several orders of magnitude. By enabling on-demand efficiency adjustment for different targets, our method provides practical solutions for biomarker discovery and mechanistic studies while reducing false-negative rates in clinical testing.</p><p>Despite its promising performance, PA has certain limitations that warrant consideration. First, the machine learning model, trained on dataset from specific PCR systems, may exhibit limited generalizability across different experimental conditions or sample types. To ensure robust model evaluation, k-fold cross-validation can be used to verify stability and reliability across diverse data subsets. The model can serve as a pretrained foundation, enabling researchers to fine-tune it with limited data from their own experimental systems for rapid adaptation to different application scenarios. In addition, PA faces technical challenges common to conventional PCR. For instance, templates with extreme GC content, long amplicons, dimer formation, or extremely low starting copy numbers may compromise amplification efficiency and modulation precision. These limitations can be addressed through two complementary strategies: optimizing primer design to avoid extreme sequence features during the design phase, and improving reaction conditions through high-fidelity polymerases, high-GC buffers, or adjusted cycling parameters. Moving forward, collecting diverse external datasets and establishing model repositories tailored to different PCR systems will further validate and expand the applicability of PA approach.</p><p>While these limitations require continued attention, the fundamental advantages of PA—precise efficiency control and machine learning–guided optimization—position this technology for broad impact across molecular biology. Beyond these demonstrated applications, we anticipate that this PA technology will find broad utility in diverse molecular analysis scenarios including single-cell sequencing and spatial transcriptomics. The ability to prioritize specific targets through customized amplification efficiency control offers a more efficient and reliable solution for multiplex detection systems. Future work will focus on systematically addressing current limitations while further enhancing PA’s programmability and precision, ultimately broadening its applications in biomedical research and clinical diagnostics.</p></sec><sec sec-type="materials|methods" disp-level="1"><title>MATERIALS AND METHODS</title><sec disp-level="2"><title>Theoretical simulation of nucleic acid amplification</title><p>A kinetic model was developed to mechanistically simulate nucleic acid amplification, decomposing the PCR process into multiple repetitive cycles. Each cycle comprised primer-template hybridization and polymerase-mediated primer extension. The hybridization kinetics were parameterized using the law of mass action, with forward rate constants (<italic toggle="yes">k</italic><sub>f</sub>) assumed to be ~10<sup>6</sup> M<sup>−1</sup> s<sup>−1</sup> for oligonucleotides of several tens of bases. Reverse rate constants (<italic toggle="yes">k</italic><sub>r</sub>) were estimated from the Δ<italic toggle="yes">G</italic>° via a thermodynamic relationship. The resulting system of ordinary differential equations was numerically solved using MATLAB’s stiff solver ode23s, with the output of one cycle serving as the initial state for the next. The model assumed constant enzyme activity, no secondary structures, and no side reactions. Detailed equations, parameter derivations, and simulation workflow are provided in Supplementary Text.</p></sec><sec disp-level="2"><title>Oligonucleotide synthesis and purification</title><p>Oligonucleotide pools were synthesized by Twist Bioscience (San Francisco, USA) and supplied as lyophilized DNA. All other individual DNA oligonucleotides were ordered from Sangon Biotech Co. (Shanghai, China). Unmodified oligonucleotides shorter than 59 nucleotides were purified using high-affinity purification, whereas those ≥59 nucleotides in length were purified by high-performance liquid chromatography.</p></sec><sec disp-level="2"><title>Nucleic acid amplification data acquisition and processing</title><p>Nucleic acid amplification experiments were conducted on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad) using 96-well plates, with Blue SYBR Green Master Mix (YEASEN, catalog no. 11184ES03) as the reaction reagent. Plasmids containing multiple target sites were used as templates. For each target site, several primers were designed by base insertion/deletion and by including or excluding tag sequences. Primer pairs were assembled by combining one programmable primer (either FP or RP) with a conventional primer (Δ<italic toggle="yes">G</italic>° ≈ −12 kcal mol<sup>−1</sup>), thereby varying only a single primer within each experimental group to reduce complexity.</p><p>All reactions were performed at a fixed plasmid template concentration (6000 copies/μl). Each 10 μl reaction consisted of 5 μl of Blue 2× Master Mix, 1 μl of DNA template, 1 μl of FP (4 μM), 1 μl of RP (4 μM), and nuclease-free water to the final volume. A two-step cycling protocol was applied: initial denaturation at 95°C for 3 min, followed by 50 cycles of denaturation at 95°C for 10 s, and annealing/extension at a designated temperature for 30 s. Each primer set was tested under multiple annealing/extension temperatures, with one temperature applied per reaction plate.</p><p>For each condition, three parallel replicates were included. Upon completion of amplification, a standard melting curve analysis was performed. Melting curve profiles were inspected to exclude reactions with abnormal amplification peaks. To minimize the influence of outliers, the median Ct value from the three replicates was used as the final result for subsequent modeling.</p></sec><sec disp-level="2"><title>Model training and interpretability analysis</title><p>A regression model was developed to predict PCR amplification performance, formulated as the direct estimation of Ct values from primer sequence features and experimental conditions. The dataset comprised multiple primers tested under varying temperatures, extension times, and tag configurations. To prevent data leakage, data were split by primer identity (70% training, 30% validation), ensuring that all results for a given primer were assigned exclusively to one set. Three datasets were modeled separately: without tag, with tag, and the combined total dataset (including a binary “tag” feature).</p><p>Features included reaction parameters (temperature and time), primer properties (length, GC content, and last nucleotide), and thermodynamic stability (Δ<italic toggle="yes">G</italic>, <italic toggle="yes">T</italic><sub>m</sub>, and average binding probability). Model training was performed using the H<sub>2</sub>O AutoML framework, generating multiple base learners (GLM, DRF, GBM, XGBoost, and XRT) and combining top models into a stacked ensemble for final prediction.</p><p>Interpretability analysis was conducted on the stacked ensemble to assess feature influence. Global importance was determined using model-agnostic permutation importance, local effects were quantified via SHAP values, marginal effects were visualized with partial dependence plots, and feature interaction patterns were evaluated using Pearson correlation of SHAP contributions. More details are provided in Supplementary Text.</p></sec><sec disp-level="2"><title>Random access of target files and NGS library preparation</title><p>To retrieve specific files, primer pairs were selected on the basis of the desired access pattern, allowing the use of programmable primers with tailored amplification efficiencies to achieve different retrieval outcomes. FPs and RPs for the chosen targets were mixed to a final concentration of 4 μM. Target sequences were amplified using Phusion DNA polymerase (Thermo Fisher Scientific, catalog no. F-530 L) in a 50-μl reaction containing 2 μl of the oligo pool, 2.5 μl of FP mix, 2.5 μl of RP mix, 0.5 μl of polymerase, 1 μl of 10 mM dNTPs, 10 μl of 5× high-fidelity (HF) buffer, and 31.5 μl of nuclease-free water. PCR was performed with an initial denaturation at 98°C for 2 min, followed by 12 cycles of 98°C for 20 s and 63°C for 10 s. The resulting amplicons were purified using a magnetic bead–based cleanup kit (Vazyme, catalog no. N411-02).</p><p>Adapter sequences were then appended in a second PCR step. For this, 15 μl of purified amplicon was combined with 2.5 μl of adapter FP mix, 2.5 μl of adapter RP mix, 0.5 μl of polymerase, 1 μl of 10 mM dNTPs, 10 μl of 5× HF buffer, and 18.5 μl of water, giving a total volume of 50 μl. Thermocycling was carried out with an initial 98°C for 2 min, followed by three cycles of 98°C for 20 s and 63°C for 10 s. PCR products were again bead-purified to remove residual primers and nucleotides.</p><p>The purified products were diluted 100-fold and used as templates in a quantitative PCR (qPCR) quantification assay. Each 10 μl of reaction contained 5 μl of Blue 2× Master Mix (YEASEN), 3 μl of diluted amplicon, 1 μl of N5 primer (5× diluted), and 1 μl of N7 primer (5× diluted). The qPCR program began with 95°C for 3 min, followed by 40 cycles of 95°C for 10 s and 60°C for 30 s. The obtained Ct values were used to determine the optimal cycle number for indexing.</p><p>For indexing, 15 μl of diluted amplicon was mixed with 1 μl of N5 index primer, 1 μl of N7 index primer, 0.5 μl of polymerase, 1 μl of 10 mM dNTPs, 10 μl of 5× HF buffer, and 21.5 μl of water, to a final volume of 50 μl. PCR was performed with an initial denaturation at 98°C for 2 min, followed by Ct + four cycles of 98°C for 20 s and 63°C for 10 s. Indexed products were purified, quantified using a commercial kit (Sangon, catalog no. N608301-0500), and pooled in equimolar amounts before next-generation sequencing.</p></sec><sec disp-level="2"><title>Data encoding and decoding</title><p>Digital image files were first divided into smaller information blocks, each assigned a unique address. These blocks were then converted into DNA sequences according to a predefined coding scheme, after which primer sequences were appended to both ends. The resulting oligonucleotide set was synthesized as a pooled library by a commercial DNA synthesis provider.</p><p>For data retrieval, sequencing results were first normalized to the same sequencing depth (typically 30× coverage) by down-sampling. The processed reads were then examined for correct primer sequences to confirm validity and identify the corresponding file. From each valid read, the address and payload regions were extracted, and payloads were grouped by their address labels. For each address, the most frequently occurring payload sequence—provided that it surpassed a predefined occurrence threshold—was retained for reconstruction. Any pixel positions where a valid payload could not be recovered were replaced with white pixels to preserve the structural layout of the decoded image.</p></sec></sec></body><back><ack><title>Acknowledgments</title><sec sec-type="funding"><title>Funding:</title><p>This work was supported the National Natural Science Foundation of China (nos. 22574102 and 22174094), the Fundamental Research Funds for the Central Universities (YG2023QNA33 and YG2025ZD28), Young Leading Scientists Cultivation Plan supported by Shanghai Municipal Education Commission (ZXWH1082101), the Shanghai Science and Technology Committee (24Y22800300), Sjtu Trans-med Awards Research (20240202), and the New Cornerstone Science Foundation supported by Shanghai Pilot Program for Basic Research-Shanghai Jiao Tong University (21TQ1400222)</p></sec><sec sec-type="author-contributions"><title>Author contributions:</title><p>P.S., C.F., and Z.W. conceived the research. P.S. and Z.W. designed the sequences. Z.W. and W.H. performed the wet experiments. Z.W., Y.W., X.X., and P.S. wrote the data analysis scripts and performed the simulations. H.L., F.W., and X.Z. participated in data analysis and discussions. P.S., C.F., and Z.W. analyzed data and wrote the manuscript. P.S., C.F., Z.W., H.L., F.W., and X.Z. reviewed and edited the manuscript.</p></sec><sec sec-type="COI-statement"><title>Competing interests:</title><p>The authors declare that they have no competing interests.</p></sec><sec sec-type="data-availability"><title>Data, code, and materials availability</title><p>All data and code needed to evaluate and reproduce the results in the paper are present in the paper and/or the Supplementary Materials. This study did not generate new materials. The code used in this study is available on GitHub (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://github.com/NABMElab/programmable-amplification" ext-link-type="uri">https://github.com/NABMElab/programmable-amplification</ext-link>) and archived in Zenodo (<ext-link xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="https://doi.org/10.5281/zenodo.17900013" ext-link-type="uri">https://doi.org/10.5281/zenodo.17900013</ext-link>) as auxiliary resources.</p></sec></ack><sec sec-type="supplementary-material"><title>Supplementary Materials</title><sec><title>This PDF file includes:</title><supplementary-material position="float" content-type="local-data" orientation="portrait"><caption><p>Supplementary Text</p><p>Figs. S1 to S27</p><p>Tables S1 to S3</p></caption><media xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="sciadv.aec9175_sm.pdf" position="float" orientation="portrait"/></supplementary-material></sec></sec><ref-list><title>REFERENCES</title><ref id="R1"><label>1</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>H.</given-names>
<surname>Lin</surname></string-name>, <string-name name-style="western"><given-names>W.</given-names>
<surname>Yu</surname></string-name>, <string-name name-style="western"><given-names>K. A.</given-names>
<surname>Sabet</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Bogumil</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Zhao</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Hambalek</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Lin</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Chandrasekaran</surname></string-name>, <string-name name-style="western"><given-names>O.</given-names>
<surname>Garner</surname></string-name>, <string-name name-style="western"><given-names>D.</given-names>
<surname>Di Carlo</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Emaminejad</surname></string-name></person-group>,
<article-title>Ferrobotic swarms enable accessible and adaptable automated viral testing</article-title>. <source>Nature</source>
<volume>611</volume>,
<fpage>570</fpage>–<lpage>577</lpage> (<year>2022</year>).<pub-id pub-id-type="pmid">36352231</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41586-022-05408-3</pub-id><pub-id pub-id-type="pmcid">PMC9645323</pub-id></mixed-citation></ref><ref id="R2"><label>2</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>L. R.</given-names>
<surname>Wu</surname></string-name>, <string-name name-style="western"><given-names>S. X.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Wu</surname></string-name>, <string-name name-style="western"><given-names>A. A.</given-names>
<surname>Patel</surname></string-name>, <string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name></person-group>,
<article-title>Multiplexed enrichment of rare DNA variants via sequence-selective and temperature-robust amplification</article-title>. <source>Nat. Biomed. Eng.</source>
<volume>1</volume>,
<fpage>714</fpage>–<lpage>723</lpage> (<year>2017</year>).<pub-id pub-id-type="pmid">29805844</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41551-017-0126-5</pub-id><pub-id pub-id-type="pmcid">PMC5969535</pub-id></mixed-citation></ref><ref id="R3"><label>3</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>N. R.</given-names>
<surname>Blumenfeld</surname></string-name>, <string-name name-style="western"><given-names>M. A. E.</given-names>
<surname>Bolene</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Jaspan</surname></string-name>, <string-name name-style="western"><given-names>A. G.</given-names>
<surname>Ayers</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Zarrandikoetxea</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Freudman</surname></string-name>, <string-name name-style="western"><given-names>N.</given-names>
<surname>Shah</surname></string-name>, <string-name name-style="western"><given-names>A. M.</given-names>
<surname>Tolwani</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Hu</surname></string-name>, <string-name name-style="western"><given-names>T. L.</given-names>
<surname>Chern</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Rogot</surname></string-name>, <string-name name-style="western"><given-names>V.</given-names>
<surname>Behnam</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Sekhar</surname></string-name>, <string-name name-style="western"><given-names>X.</given-names>
<surname>Liu</surname></string-name>, <string-name name-style="western"><given-names>B.</given-names>
<surname>Onalir</surname></string-name>, <string-name name-style="western"><given-names>R.</given-names>
<surname>Kasumi</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Sanogo</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Human</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Murakami</surname></string-name>, <string-name name-style="western"><given-names>G. S.</given-names>
<surname>Totapally</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Fasciano</surname></string-name>, <string-name name-style="western"><given-names>S. K.</given-names>
<surname>Sia</surname></string-name></person-group>,
<article-title>Multiplexed reverse-transcriptase quantitative polymerase chain reaction using plasmonic nanoparticles for point-of-care COVID-19 diagnosis</article-title>. <source>Nat. Nanotechnol.</source>
<volume>17</volume>,
<fpage>984</fpage>–<lpage>992</lpage> (<year>2022</year>).<pub-id pub-id-type="pmid">35879456</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41565-022-01175-4</pub-id></mixed-citation></ref><ref id="R4"><label>4</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>J. Y.</given-names>
<surname>Kishi</surname></string-name>, <string-name name-style="western"><given-names>N.</given-names>
<surname>Liu</surname></string-name>, <string-name name-style="western"><given-names>E. R.</given-names>
<surname>West</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Sheng</surname></string-name>, <string-name name-style="western"><given-names>J. J.</given-names>
<surname>Jordanides</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Serrata</surname></string-name>, <string-name name-style="western"><given-names>C. L.</given-names>
<surname>Cepko</surname></string-name>, <string-name name-style="western"><given-names>S. K.</given-names>
<surname>Saka</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>Yin</surname></string-name></person-group>,
<article-title>Light-Seq: Light-directed in situ barcoding of biomolecules in fixed cells and tissues for spatially indexed sequencing</article-title>. <source>Nat. Methods</source>
<volume>19</volume>,
<fpage>1393</fpage>–<lpage>1402</lpage> (<year>2022</year>).<pub-id pub-id-type="pmid">36216958</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41592-022-01604-1</pub-id><pub-id pub-id-type="pmcid">PMC9636025</pub-id></mixed-citation></ref><ref id="R5"><label>5</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>S. G.</given-names>
<surname>Rodriques</surname></string-name>, <string-name name-style="western"><given-names>R. R.</given-names>
<surname>Stickels</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Goeva</surname></string-name>, <string-name name-style="western"><given-names>C. A.</given-names>
<surname>Martin</surname></string-name>, <string-name name-style="western"><given-names>E.</given-names>
<surname>Murray</surname></string-name>, <string-name name-style="western"><given-names>C. R.</given-names>
<surname>Vanderburg</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Welch</surname></string-name>, <string-name name-style="western"><given-names>L. M.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>F.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>E. Z.</given-names>
<surname>Macosko</surname></string-name></person-group>,
<article-title>Slide-seq: A scalable technology for measuring genome-wide expression at high spatial resolution</article-title>. <source>Science</source>
<volume>363</volume>,
<fpage>1463</fpage>–<lpage>1467</lpage> (<year>2019</year>).<pub-id pub-id-type="pmid">30923225</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.aaw1219</pub-id><pub-id pub-id-type="pmcid">PMC6927209</pub-id></mixed-citation></ref><ref id="R6"><label>6</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>A. C.</given-names>
<surname>Payne</surname></string-name>, <string-name name-style="western"><given-names>Z. D.</given-names>
<surname>Chiang</surname></string-name>, <string-name name-style="western"><given-names>P. L.</given-names>
<surname>Reginato</surname></string-name>, <string-name name-style="western"><given-names>S. M.</given-names>
<surname>Mangiameli</surname></string-name>, <string-name name-style="western"><given-names>E. M.</given-names>
<surname>Murray</surname></string-name>, <string-name name-style="western"><given-names>C.</given-names>
<surname>Yao</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Markoulaki</surname></string-name>, <string-name name-style="western"><given-names>A. S.</given-names>
<surname>Earl</surname></string-name>, <string-name name-style="western"><given-names>A. S.</given-names>
<surname>Labade</surname></string-name>, <string-name name-style="western"><given-names>R.</given-names>
<surname>Jaenisch</surname></string-name>, <string-name name-style="western"><given-names>G. M.</given-names>
<surname>Church</surname></string-name>, <string-name name-style="western"><given-names>E. S.</given-names>
<surname>Boyden</surname></string-name>, <string-name name-style="western"><given-names>J. D.</given-names>
<surname>Buenrostro</surname></string-name>, <string-name name-style="western"><given-names>F.</given-names>
<surname>Chen</surname></string-name></person-group>,
<article-title>In situ genome sequencing resolves DNA sequence and structure in intact biological samples</article-title>. <source>Science</source>
<volume>371</volume>,
<fpage>908</fpage> (<year>2021</year>).<pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.aay3446</pub-id><pub-id pub-id-type="pmcid">PMC7962746</pub-id><pub-id pub-id-type="pmid">33384301</pub-id></mixed-citation></ref><ref id="R7"><label>7</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>P.</given-names>
<surname>Song</surname></string-name>, <string-name name-style="western"><given-names>S. X.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>Y. H.</given-names>
<surname>Yan</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Pinto</surname></string-name>, <string-name name-style="western"><given-names>L. Y.</given-names>
<surname>Cheng</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>Dai</surname></string-name>, <string-name name-style="western"><given-names>A. A.</given-names>
<surname>Patel</surname></string-name>, <string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name></person-group>,
<article-title>Selective multiplexed enrichment for the detection and quantitation of low-fraction DNA variants via low-depth sequencing</article-title>. <source>Nat. Biomed. Eng.</source>
<volume>5</volume>,
<fpage>690</fpage>–<lpage>701</lpage> (<year>2021</year>).<pub-id pub-id-type="pmid">33941896</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41551-021-00713-0</pub-id><pub-id pub-id-type="pmcid">PMC9631981</pub-id></mixed-citation></ref><ref id="R8"><label>8</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>S. D.</given-names>
<surname>Lemaire</surname></string-name>, <string-name name-style="western"><given-names>D.</given-names>
<surname>Turek</surname></string-name>, <string-name name-style="western"><given-names>D.</given-names>
<surname>Landsman</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Colotte</surname></string-name>, <string-name name-style="western"><given-names>T. F. A.</given-names>
<surname>de Greef</surname></string-name></person-group>,
<article-title>Challenges and opportunities in DNA computing and data storage</article-title>. <source>Nat. Nanotechnol.</source>
<volume>20</volume>,
<fpage>710</fpage>–<lpage>714</lpage> (<year>2025</year>).<pub-id pub-id-type="pmid">40494931</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41565-025-01937-w</pub-id></mixed-citation></ref><ref id="R9"><label>9</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>L.</given-names>
<surname>Ceze</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Nivala</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Strauss</surname></string-name></person-group>,
<article-title>Molecular digital data storage using DNA</article-title>. <source>Nat. Rev. Genet.</source>
<volume>20</volume>,
<fpage>456</fpage>–<lpage>466</lpage> (<year>2019</year>).<pub-id pub-id-type="pmid">31068682</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41576-019-0125-3</pub-id></mixed-citation></ref><ref id="R10"><label>10</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>N.</given-names>
<surname>Goldman</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>Bertone</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>C.</given-names>
<surname>Dessimoz</surname></string-name>, <string-name name-style="western"><given-names>E. M.</given-names>
<surname>Leproust</surname></string-name>, <string-name name-style="western"><given-names>B.</given-names>
<surname>Sipos</surname></string-name>, <string-name name-style="western"><given-names>E.</given-names>
<surname>Birney</surname></string-name></person-group>,
<article-title>Towards practical, high-capacity, low-maintenance information storage in synthesized DNA</article-title>. <source>Nature</source>
<volume>494</volume>,
<fpage>77</fpage>–<lpage>80</lpage> (<year>2013</year>).<pub-id pub-id-type="pmid">23354052</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nature11875</pub-id><pub-id pub-id-type="pmcid">PMC3672958</pub-id></mixed-citation></ref><ref id="R11"><label>11</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>G. M.</given-names>
<surname>Church</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Gao</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Kosuri</surname></string-name></person-group>,
<article-title>Next-generation digital information storage in DNA</article-title>. <source>Science</source>
<volume>337</volume>,
<fpage>1628</fpage> (<year>2012</year>).<pub-id pub-id-type="pmid">22903519</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1126/science.1226355</pub-id></mixed-citation></ref><ref id="R12"><label>12</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>K. J.</given-names>
<surname>Tomek</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Volkel</surname></string-name>, <string-name name-style="western"><given-names>E. W.</given-names>
<surname>Indermaur</surname></string-name>, <string-name name-style="western"><given-names>J. M.</given-names>
<surname>Tuck</surname></string-name>, <string-name name-style="western"><given-names>A. J.</given-names>
<surname>Keung</surname></string-name></person-group>,
<article-title>Promiscuous molecules for smarter file operations in DNA-based data storage</article-title>. <source>Nat. Commun.</source>
<volume>12</volume>,
<fpage>3518</fpage> (<year>2021</year>).<pub-id pub-id-type="pmid">34112775</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-021-23669-w</pub-id><pub-id pub-id-type="pmcid">PMC8192770</pub-id></mixed-citation></ref><ref id="R13"><label>13</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>D.</given-names>
<surname>Sims</surname></string-name>, <string-name name-style="western"><given-names>I.</given-names>
<surname>Sudbery</surname></string-name>, <string-name name-style="western"><given-names>N. E.</given-names>
<surname>Ilott</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Heger</surname></string-name>, <string-name name-style="western"><given-names>C. P.</given-names>
<surname>Ponting</surname></string-name></person-group>,
<article-title>Sequencing depth and coverage: Key considerations in genomic analyses</article-title>. <source>Nat. Rev. Genet.</source>
<volume>15</volume>,
<fpage>121</fpage>–<lpage>132</lpage> (<year>2014</year>).<pub-id pub-id-type="pmid">24434847</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nrg3642</pub-id></mixed-citation></ref><ref id="R14"><label>14</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>S.</given-names>
<surname>Kosuri</surname></string-name>, <string-name name-style="western"><given-names>G. M.</given-names>
<surname>Church</surname></string-name></person-group>,
<article-title>Large-scale de novo DNA synthesis: Technologies and applications</article-title>. <source>Nat. Methods</source>
<volume>11</volume>,
<fpage>499</fpage>–<lpage>507</lpage> (<year>2014</year>).<pub-id pub-id-type="pmid">24781323</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nmeth.2918</pub-id><pub-id pub-id-type="pmcid">PMC7098426</pub-id></mixed-citation></ref><ref id="R15"><label>15</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>A.</given-names>
<surname>Hoose</surname></string-name>, <string-name name-style="western"><given-names>R.</given-names>
<surname>Vellacott</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Storch</surname></string-name>, <string-name name-style="western"><given-names>P. S.</given-names>
<surname>Freemont</surname></string-name>, <string-name name-style="western"><given-names>M. G.</given-names>
<surname>Ryadnov</surname></string-name></person-group>,
<article-title>DNA synthesis technologies to close the gene writing gap</article-title>. <source>Nat. Rev. Chem.</source>
<volume>7</volume>,
<fpage>144</fpage>–<lpage>161</lpage> (<year>2023</year>).<pub-id pub-id-type="pmid">36714378</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41570-022-00456-9</pub-id><pub-id pub-id-type="pmcid">PMC9869848</pub-id></mixed-citation></ref><ref id="R16"><label>16</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>Z.</given-names>
<surname>Weng</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Li</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Wu</surname></string-name>, <string-name name-style="western"><given-names>X.</given-names>
<surname>Xiu</surname></string-name>, <string-name name-style="western"><given-names>F.</given-names>
<surname>Wang</surname></string-name>, <string-name name-style="western"><given-names>X.</given-names>
<surname>Zuo</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>Song</surname></string-name>, <string-name name-style="western"><given-names>C.</given-names>
<surname>Fan</surname></string-name></person-group>,
<article-title>Massively parallel homogeneous amplification of chip-scale DNA for DNA information storage (MPHAC-DIS)</article-title>. <source>Nat. Commun.</source>
<volume>16</volume>,
<fpage>667</fpage> (<year>2025</year>).<pub-id pub-id-type="pmid">39809776</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-025-55986-9</pub-id><pub-id pub-id-type="pmcid">PMC11733265</pub-id></mixed-citation></ref><ref id="R17"><label>17</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>L. R.</given-names>
<surname>Wu</surname></string-name>, <string-name name-style="western"><given-names>J. S.</given-names>
<surname>Wang</surname></string-name>, <string-name name-style="western"><given-names>J. Z.</given-names>
<surname>Fang</surname></string-name>, <string-name name-style="western"><given-names>E. R.</given-names>
<surname>Evans</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Pinto</surname></string-name>, <string-name name-style="western"><given-names>I.</given-names>
<surname>Pekker</surname></string-name>, <string-name name-style="western"><given-names>R.</given-names>
<surname>Boykin</surname></string-name>, <string-name name-style="western"><given-names>C.</given-names>
<surname>Ngouenet</surname></string-name>, <string-name name-style="western"><given-names>P. J.</given-names>
<surname>Webster</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Beechem</surname></string-name>, <string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name></person-group>,
<article-title>Continuously tunable nucleic acid hybridization probes</article-title>. <source>Nat. Methods</source>
<volume>12</volume>,
<fpage>1191</fpage>–<lpage>1196</lpage> (<year>2015</year>).<pub-id pub-id-type="pmid">26480474</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nmeth.3626</pub-id><pub-id pub-id-type="pmcid">PMC4666732</pub-id></mixed-citation></ref><ref id="R18"><label>18</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name>, <string-name name-style="western"><given-names>S. X.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>Yin</surname></string-name></person-group>,
<article-title>Optimizing the specificity of nucleic acid hybridization</article-title>. <source>Nat. Chem.</source>
<volume>4</volume>,
<fpage>208</fpage>–<lpage>214</lpage> (<year>2012</year>).<pub-id pub-id-type="pmid">22354435</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nchem.1246</pub-id><pub-id pub-id-type="pmcid">PMC4238961</pub-id></mixed-citation></ref><ref id="R19"><label>19</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>J. X.</given-names>
<surname>Zhang</surname></string-name>, <string-name name-style="western"><given-names>J. Z.</given-names>
<surname>Fang</surname></string-name>, <string-name name-style="western"><given-names>W.</given-names>
<surname>Duan</surname></string-name>, <string-name name-style="western"><given-names>L. R.</given-names>
<surname>Wu</surname></string-name>, <string-name name-style="western"><given-names>A. W.</given-names>
<surname>Zhang</surname></string-name>, <string-name name-style="western"><given-names>N.</given-names>
<surname>Dalchau</surname></string-name>, <string-name name-style="western"><given-names>B.</given-names>
<surname>Yordanov</surname></string-name>, <string-name name-style="western"><given-names>R.</given-names>
<surname>Petersen</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Phillips</surname></string-name>, <string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name></person-group>,
<article-title>Predicting DNA hybridization kinetics from sequence</article-title>. <source>Nat. Chem.</source>
<volume>10</volume>,
<fpage>91</fpage>–<lpage>98</lpage> (<year>2018</year>).<pub-id pub-id-type="pmid">29256499</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nchem.2877</pub-id><pub-id pub-id-type="pmcid">PMC5739081</pub-id></mixed-citation></ref><ref id="R20"><label>20</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>C.</given-names>
<surname>Bee</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Queen</surname></string-name>, <string-name name-style="western"><given-names>D.</given-names>
<surname>Ward</surname></string-name>, <string-name name-style="western"><given-names>X.</given-names>
<surname>Liu</surname></string-name>, <string-name name-style="western"><given-names>L.</given-names>
<surname>Organick</surname></string-name>, <string-name name-style="western"><given-names>G.</given-names>
<surname>Seelig</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Strauss</surname></string-name>, <string-name name-style="western"><given-names>L.</given-names>
<surname>Ceze</surname></string-name></person-group>,
<article-title>Molecular-level similarity search brings computing to DNA data storage</article-title>. <source>Nat. Commun.</source>
<volume>12</volume>,
<fpage>4764</fpage> (<year>2021</year>).<pub-id pub-id-type="pmid">34362913</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-021-24991-z</pub-id><pub-id pub-id-type="pmcid">PMC8346626</pub-id></mixed-citation></ref><ref id="R21"><label>21</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>J. S.</given-names>
<surname>Wang</surname></string-name>, <string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name></person-group>,
<article-title>Simulation-guided DNA probe design for consistently ultraspecific hybridization</article-title>. <source>Nat. Chem.</source>
<volume>7</volume>,
<fpage>545</fpage>–<lpage>553</lpage> (<year>2015</year>).<pub-id pub-id-type="pmid">26100802</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nchem.2266</pub-id><pub-id pub-id-type="pmcid">PMC4479422</pub-id></mixed-citation></ref><ref id="R22"><label>22</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>N. G.</given-names>
<surname>Xie</surname></string-name>, <string-name name-style="western"><given-names>M. X.</given-names>
<surname>Wang</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>Song</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Mao</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Wang</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Yang</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Luo</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Ren</surname></string-name>, <string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name></person-group>,
<article-title>Designing highly multiplex PCR primer sets with simulated annealing design using dimer likelihood estimation (SADDLE)</article-title>. <source>Nat. Commun.</source>
<volume>13</volume>,
<fpage>1881</fpage> (<year>2022</year>).<pub-id pub-id-type="pmid">35410464</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-022-29500-4</pub-id><pub-id pub-id-type="pmcid">PMC9001684</pub-id></mixed-citation></ref><ref id="R23"><label>23</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>J. X.</given-names>
<surname>Zhang</surname></string-name>, <string-name name-style="western"><given-names>B.</given-names>
<surname>Yordanov</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Gaunt</surname></string-name>, <string-name name-style="western"><given-names>M. X.</given-names>
<surname>Wang</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>Dai</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Chen</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Zhang</surname></string-name>, <string-name name-style="western"><given-names>J. Z.</given-names>
<surname>Fang</surname></string-name>, <string-name name-style="western"><given-names>N.</given-names>
<surname>Dalchau</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Li</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Phillips</surname></string-name>, <string-name name-style="western"><given-names>D. Y.</given-names>
<surname>Zhang</surname></string-name></person-group>,
<article-title>A deep learning model for predicting next-generation sequencing depth from DNA sequence</article-title>. <source>Nat. Commun.</source>
<volume>12</volume>,
<fpage>4387</fpage> (<year>2021</year>).<pub-id pub-id-type="pmid">34282137</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41467-021-24497-8</pub-id><pub-id pub-id-type="pmcid">PMC8290051</pub-id></mixed-citation></ref><ref id="R24"><label>24</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>D.</given-names>
<surname>Bar-Lev</surname></string-name>, <string-name name-style="western"><given-names>O.</given-names>
<surname>Sabary</surname></string-name>, <string-name name-style="western"><given-names>E.</given-names>
<surname>Yaakobi</surname></string-name></person-group>,
<article-title>The zettabyte era is in our DNA</article-title>. <source>Nat. Comput. Science</source>
<volume>4</volume>,
<fpage>813</fpage>–<lpage>817</lpage> (<year>2024</year>).<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s43588-024-00717-1</pub-id><pub-id pub-id-type="pmid">39516373</pub-id></mixed-citation></ref><ref id="R25"><label>25</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>J.</given-names>
<surname>Koch</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Gantenbein</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Masania</surname></string-name>, <string-name name-style="western"><given-names>W. J.</given-names>
<surname>Stark</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Erlich</surname></string-name>, <string-name name-style="western"><given-names>R. N.</given-names>
<surname>Grass</surname></string-name></person-group>,
<article-title>A DNA-of-things storage architecture to create materials with embedded memory</article-title>. <source>Nat. Biotechnol.</source>
<volume>38</volume>,
<fpage>39</fpage>–<lpage>43</lpage> (<year>2020</year>).<pub-id pub-id-type="pmid">31819259</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41587-019-0356-z</pub-id></mixed-citation></ref><ref id="R26"><label>26</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>C.</given-names>
<surname>Zhang</surname></string-name>, <string-name name-style="western"><given-names>R.</given-names>
<surname>Wu</surname></string-name>, <string-name name-style="western"><given-names>F.</given-names>
<surname>Sun</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Lin</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Liang</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Teng</surname></string-name>, <string-name name-style="western"><given-names>N.</given-names>
<surname>Liu</surname></string-name>, <string-name name-style="western"><given-names>Q.</given-names>
<surname>Ouyang</surname></string-name>, <string-name name-style="western"><given-names>L.</given-names>
<surname>Qian</surname></string-name>, <string-name name-style="western"><given-names>H.</given-names>
<surname>Yan</surname></string-name></person-group>,
<article-title>Parallel molecular data storage by printing epigenetic bits on DNA</article-title>. <source>Nature</source>
<volume>634</volume>,
<fpage>824</fpage>–<lpage>832</lpage> (<year>2024</year>).<pub-id pub-id-type="pmid">39443776</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/s41586-024-08040-5</pub-id><pub-id pub-id-type="pmcid">PMC11499255</pub-id></mixed-citation></ref><ref id="R27"><label>27</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>S.</given-names>
<surname>Goodwin</surname></string-name>, <string-name name-style="western"><given-names>J. D.</given-names>
<surname>Mcpherson</surname></string-name>, <string-name name-style="western"><given-names>W. R.</given-names>
<surname>Mccombie</surname></string-name></person-group>,
<article-title>Coming of age: Ten years of next-generation sequencing technologies</article-title>. <source>Nat. Rev. Genet.</source>
<volume>17</volume>,
<fpage>333</fpage>–<lpage>351</lpage> (<year>2016</year>).<pub-id pub-id-type="pmid">27184599</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nrg.2016.49</pub-id><pub-id pub-id-type="pmcid">PMC10373632</pub-id></mixed-citation></ref><ref id="R28"><label>28</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>J.</given-names>
<surname>Shendure</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Balasubramanian</surname></string-name>, <string-name name-style="western"><given-names>G. M.</given-names>
<surname>Church</surname></string-name>, <string-name name-style="western"><given-names>W.</given-names>
<surname>Gilbert</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Rogers</surname></string-name>, <string-name name-style="western"><given-names>J. A.</given-names>
<surname>Schloss</surname></string-name>, <string-name name-style="western"><given-names>R. H.</given-names>
<surname>Waterston</surname></string-name></person-group>,
<article-title>DNA sequencing at 40: Past, present and future</article-title>. <source>Nature (London)</source>
<volume>550</volume>,
<fpage>345</fpage>–<lpage>353</lpage> (<year>2017</year>).<pub-id pub-id-type="pmid">29019985</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nature24286</pub-id></mixed-citation></ref><ref id="R29"><label>29</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>L.</given-names>
<surname>Mamanova</surname></string-name>, <string-name name-style="western"><given-names>A. J.</given-names>
<surname>Coffey</surname></string-name>, <string-name name-style="western"><given-names>C. E.</given-names>
<surname>Scott</surname></string-name>, <string-name name-style="western"><given-names>I.</given-names>
<surname>Kozarewa</surname></string-name>, <string-name name-style="western"><given-names>E. H.</given-names>
<surname>Turner</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Kumar</surname></string-name>, <string-name name-style="western"><given-names>E.</given-names>
<surname>Howard</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Shendure</surname></string-name>, <string-name name-style="western"><given-names>D. J.</given-names>
<surname>Turner</surname></string-name></person-group>,
<article-title>Target-enrichment strategies for next-generation sequencing</article-title>. <source>Nat. Methods</source>
<volume>7</volume>,
<fpage>111</fpage>–<lpage>118</lpage> (<year>2010</year>).<pub-id pub-id-type="pmid">20111037</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nmeth.1419</pub-id></mixed-citation></ref><ref id="R30"><label>30</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>J. J.</given-names>
<surname>Salk</surname></string-name>, <string-name name-style="western"><given-names>M. W.</given-names>
<surname>Schmitt</surname></string-name>, <string-name name-style="western"><given-names>L. A.</given-names>
<surname>Loeb</surname></string-name></person-group>,
<article-title>Enhancing the accuracy of next-generation sequencing for detecting rare and subclonal mutations</article-title>. <source>Nat. Rev. Genet.</source>
<volume>19</volume>,
<fpage>269</fpage>–<lpage>285</lpage> (<year>2018</year>).<pub-id pub-id-type="pmid">29576615</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nrg.2017.117</pub-id><pub-id pub-id-type="pmcid">PMC6485430</pub-id></mixed-citation></ref><ref id="R31"><label>31</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><string-name name-style="western"><given-names>Y.</given-names>
<surname>Loriot</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Necchi</surname></string-name>, <string-name name-style="western"><given-names>S. H.</given-names>
<surname>Park</surname></string-name>, <string-name name-style="western"><given-names>J.</given-names>
<surname>Garcia-Donas</surname></string-name>, <string-name name-style="western"><given-names>R.</given-names>
<surname>Huddart</surname></string-name>, <string-name name-style="western"><given-names>E.</given-names>
<surname>Burgess</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Fleming</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Rezazadeh</surname></string-name>, <string-name name-style="western"><given-names>B.</given-names>
<surname>Mellado</surname></string-name>, <string-name name-style="western"><given-names>S.</given-names>
<surname>Varlamov</surname></string-name>, <string-name name-style="western"><given-names>M.</given-names>
<surname>Joshi</surname></string-name>, <string-name name-style="western"><given-names>I.</given-names>
<surname>Duran</surname></string-name>, <string-name name-style="western"><given-names>S. T.</given-names>
<surname>Tagawa</surname></string-name>, <string-name name-style="western"><given-names>Y.</given-names>
<surname>Zakharia</surname></string-name>, <string-name name-style="western"><given-names>B.</given-names>
<surname>Zhong</surname></string-name>, <string-name name-style="western"><given-names>K.</given-names>
<surname>Stuyckens</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Santiago-Walker</surname></string-name>, <string-name name-style="western"><given-names>P.</given-names>
<surname>De Porre</surname></string-name>, <string-name name-style="western"><given-names>A. O.</given-names>
<surname>Hagan</surname></string-name>, <string-name name-style="western"><given-names>A.</given-names>
<surname>Avadhani</surname></string-name>, <string-name name-style="western"><given-names>A. O.</given-names>
<surname>Siefker-Radtke</surname></string-name></person-group>,
<article-title>Erdafitinib in locally advanced or metastatic urothelial carcinoma</article-title>. <source>N. Engl. J. Med.</source>
<volume>381</volume>,
<fpage>338</fpage>–<lpage>348</lpage> (<year>2019</year>).<pub-id pub-id-type="pmid">31340094</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1056/NEJMoa1817323</pub-id></mixed-citation></ref><ref id="R32"><label>32</label><mixed-citation publication-type="journal"><person-group person-group-type="author"><collab id="coll01">The Cancer Genome Atlas Research Network</collab></person-group>,
<article-title>Integrated genomic and molecular characterization of cervical cancer</article-title>. <source>Nature</source>
<volume>543</volume>,
<fpage>378</fpage>–<lpage>384</lpage> (<year>2017</year>).<pub-id pub-id-type="pmid">28112728</pub-id>
<pub-id pub-id-type="doi" assigning-authority="pmc">10.1038/nature21386</pub-id><pub-id pub-id-type="pmcid">PMC5354998</pub-id></mixed-citation></ref></ref-list></back></article></pmc-articleset>