Integrative machine learning approaches with genomic data for predicting antitubercular drug resistance: A systematic review and meta-analysis

IF 3.7 3区医学 Q2 INFECTIOUS DISEASES

Journal of global antimicrobial resistance Pub Date : 2026-01-01 Epub Date: 2025-12-12 DOI:10.1016/j.jgar.2025.11.021

Rohan Shrivastava , Kashmi Sharma , Somesh Mishra , Poonam Parihar , Gobardhan Das , Khushhali Menaria Pandey , Anand Kumar Maurya , Vineet Kumar Sharma , Anamika Mishra , Sandeep Kushwaha , Rekha Khandia , Kuldeep Singh Yadav , Anvita Gupta Malhotra , Amit Agrawal , Jitendra Singh , Megha Katare Pandey

{"title":"Integrative machine learning approaches with genomic data for predicting antitubercular drug resistance: A systematic review and meta-analysis","authors":"Rohan Shrivastava , Kashmi Sharma , Somesh Mishra , Poonam Parihar , Gobardhan Das , Khushhali Menaria Pandey , Anand Kumar Maurya , Vineet Kumar Sharma , Anamika Mishra , Sandeep Kushwaha , Rekha Khandia , Kuldeep Singh Yadav , Anvita Gupta Malhotra , Amit Agrawal , Jitendra Singh , Megha Katare Pandey","doi":"10.1016/j.jgar.2025.11.021","DOIUrl":null,"url":null,"abstract":"<div><h3>Objectives</h3><div>Tuberculosis (TB) remains a leading infectious cause of death, with drug-resistant TB threatening control gains worldwide Rapid, accurate prediction of resistance to first-line agents is essential to guide therapy.</div><div>The objective of this meta-analysis is to evaluate the diagnostic performance of machine learning (ML) algorithms trained on genomic data for predicting phenotypic resistance to the antitubercular drugs.</div></div><div><h3>Methods</h3><div>We searched 4 databases for original studies applying ML to whole-genome or targeted sequencing for resistance prediction against rifampicin (RIF), isoniazid (INH), ethambutol (EMB), streptomycin (STM), and other routinely tested drugs. Random-effects meta-analyses (REML) pooled sensitivity and specificity; Area Under Curve (AUC) was meta-analysed after logit transformation. Publication bias was examined via funnel plots plus Egger’s and Begg’s tests with trim-and-fill adjustment.</div></div><div><h3>Results</h3><div>Seven eligible studies were included. Pooled performance was strongest for RIF and INH (RIF: sensitivity 0.90, specificity 0.95; INH: sensitivity 0.88, specificity 0.93). EMB and STM showed lower sensitivities despite reasonable AUCs. Forest plots for AUC (including logit scale), sensitivity, and specificity demonstrated drug-wise variation. Across studies, specificity (mean ≈ 0.92) exceeded sensitivity (mean ≈ 0.75). Bias diagnostics revealed marked funnel plot asymmetry; trim-and-fill imputation added 22, 29, and 23 studies for AUC, sensitivity, and specificity, respectively, yielding adjusted pooled estimates of ∼0.89 (AUC), 0.70 (sensitivity), and 0.93 (specificity).</div></div><div><h3>Conclusions</h3><div>ML models trained on genomic data demonstrate high diagnostic accuracy and robust discriminative ability for predicting first-line drug resistance—particularly for RIF and INH—although sensitivity remains variable across drugs and model types. Standardized external validation and calibration are needed before broad clinical deployment.</div></div>","PeriodicalId":15936,"journal":{"name":"Journal of global antimicrobial resistance","volume":"46 ","pages":"Pages 148-157"},"PeriodicalIF":3.7000,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of global antimicrobial resistance","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2213716525002681","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/12/12 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"INFECTIOUS DISEASES","Score":null,"Total":0}

引用次数: 0

Abstract

Objectives

Tuberculosis (TB) remains a leading infectious cause of death, with drug-resistant TB threatening control gains worldwide Rapid, accurate prediction of resistance to first-line agents is essential to guide therapy.

The objective of this meta-analysis is to evaluate the diagnostic performance of machine learning (ML) algorithms trained on genomic data for predicting phenotypic resistance to the antitubercular drugs.

Methods

We searched 4 databases for original studies applying ML to whole-genome or targeted sequencing for resistance prediction against rifampicin (RIF), isoniazid (INH), ethambutol (EMB), streptomycin (STM), and other routinely tested drugs. Random-effects meta-analyses (REML) pooled sensitivity and specificity; Area Under Curve (AUC) was meta-analysed after logit transformation. Publication bias was examined via funnel plots plus Egger’s and Begg’s tests with trim-and-fill adjustment.

Results

Seven eligible studies were included. Pooled performance was strongest for RIF and INH (RIF: sensitivity 0.90, specificity 0.95; INH: sensitivity 0.88, specificity 0.93). EMB and STM showed lower sensitivities despite reasonable AUCs. Forest plots for AUC (including logit scale), sensitivity, and specificity demonstrated drug-wise variation. Across studies, specificity (mean ≈ 0.92) exceeded sensitivity (mean ≈ 0.75). Bias diagnostics revealed marked funnel plot asymmetry; trim-and-fill imputation added 22, 29, and 23 studies for AUC, sensitivity, and specificity, respectively, yielding adjusted pooled estimates of ∼0.89 (AUC), 0.70 (sensitivity), and 0.93 (specificity).

Conclusions

ML models trained on genomic data demonstrate high diagnostic accuracy and robust discriminative ability for predicting first-line drug resistance—particularly for RIF and INH—although sensitivity remains variable across drugs and model types. Standardized external validation and calibration are needed before broad clinical deployment.

查看原文本刊更多论文

整合机器学习方法与基因组数据预测抗结核药物耐药性：系统回顾和荟萃分析。

目的：结核病（TB）仍然是主要的传染性死亡原因，耐药结核病威胁着世界范围内的控制成果，快速、准确地预测对一线药物的耐药性对于指导治疗至关重要。本荟萃分析的目的是评估基于基因组数据训练的机器学习（ML）算法的诊断性能，以预测抗结核药物的表型耐药性。方法：我们检索了4个数据库，检索了应用ML全基因组或靶向测序预测对利福平（RIF）、异烟肼（INH）、乙胺丁醇（EMB）、链霉素（STM）和其他常规检测药物耐药的原始研究。随机效应荟萃分析（REML）汇集了敏感性和特异性；曲线下面积（AUC）经logit变换后进行meta分析。发表偏倚通过漏斗图和Egger’s和Begg’s检验进行检验。结果：纳入了7项符合条件的研究。RIF和INH的综合表现最强（RIF：敏感性0.90，特异性0.95；INH：敏感性0.88，特异性0.93）。尽管auc合理，但EMB和STM的敏感性较低。AUC（包括logit量表）、敏感性和特异性的森林图显示了药物方面的差异。在所有研究中，特异性（平均值≈0.92）超过敏感性（平均值≈0.75）。偏倚诊断显示漏斗图不对称性明显；修整填充法分别增加了22、29和23项AUC、敏感性和特异性研究，调整后的合并估计值分别为0.89 （AUC）、0.70（敏感性）和0.93（特异性）。结论：基于基因组数据训练的ML模型在预测一线耐药性（尤其是RIF和inh）方面表现出较高的诊断准确性和强大的判别能力，尽管敏感性在药物和模型类型之间仍然存在差异。在广泛的临床部署之前，需要标准化的外部验证和校准。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of global antimicrobial resistance INFECTIOUS DISEASES-PHARMACOLOGY & PHARMACY

CiteScore

8.70

自引率

2.20%

发文量

285

审稿时长

34 weeks

期刊介绍： The Journal of Global Antimicrobial Resistance (JGAR) is a quarterly online journal run by an international Editorial Board that focuses on the global spread of antibiotic-resistant microbes. JGAR is a dedicated journal for all professionals working in research, health care, the environment and animal infection control, aiming to track the resistance threat worldwide and provides a single voice devoted to antimicrobial resistance (AMR). Featuring peer-reviewed and up to date research articles, reviews, short notes and hot topics JGAR covers the key topics related to antibacterial, antiviral, antifungal and antiparasitic resistance.