{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,9]],"date-time":"2026-09-09T12:41:37Z","timestamp":1788957697579,"version":"build-2803163510"},"reference-count":63,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T00:00:00Z","timestamp":1785369600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Digit. Health"],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Early detection of drug-resistant tuberculosis (DR-TB) is critical for timely treatment and limiting transmission. Differentiating DR-TB from drug-sensitive TB (DS-TB) remains challenging due to overlapping manifestations and subtle differences in chest X-rays (CXRs).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We developed an explainable multimodal machine learning framework integrating CXR embeddings from a Data-Efficient Image Transformer (DEiT) with patient-level data, including clinical, interpreted radiological, demographic, socioeconomic, and behavioral variables. Using 6,279 subjects from the TB Portals program, model performance was evaluated across three settings comprising CXR embeddings only, patient-level features only, and multimodal. Interpretability was analyzed using Shapley Additive Explanations (SHAP).<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>A multimodal XGBoost model achieved a precision-recall area under the curve (PR-AUC) of 0.9258 and receiver operating characteristic area under the curve (ROC-AUC) of 0.8718. CXR-embeddings-only classifiers performed lower (PR-AUC: 0.8334; ROC-AUC: 0.7369), while patient-level XGBoost yielded highest performance. The multimodal model improved interpretability through complementary structural information. SHAP identified CXR embeddings as primary global drivers. Among patient-level variables, daily contacts and employment status were the strongest contributors, reflecting exposure intensity and socioeconomic conditions. Comorbidity showed moderate effects, while education and case definition had smaller contributions. Key radiological factors included non-tuberculosis abnormalities and mediastinal lymph node involvement, whereas medium-density stabalized fibrotic nodules, medium nodules, and small cavities demonstrated minimal association. Interaction analysis revealed higher risk with increasing comorbidity in the presence of morphological alterations; BMI and age exhibited modest influence. Independent review by two clinicians confirmed plausibility and consistency of SHAP explanations.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>This interpretable multimodal framework supports early DR-TB detection and targeted interventions in high-burden settings.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fdgth.2026.1828522","type":"journal-article","created":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T04:21:31Z","timestamp":1785385291000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["Interpretable multimodal machine learning for diagnosis of drug-resistant tuberculosis"],"prefix":"10.3389","volume":"8","author":[{"given":"Joan Jonathan","family":"Mnyambo","sequence":"first","affiliation":[{"name":"School of Engineering, Computing and Mathematics, University of Plymouth","place":["Plymouth, United Kingdom"]},{"name":"Department of Informatics and Information Technology, Sokoine University of Agriculture","place":["Morogoro, Tanzania"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Amir","family":"Aly","sequence":"additional","affiliation":[{"name":"School of Engineering, Computing and Mathematics, University of Plymouth","place":["Plymouth, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Emmanuel","family":"Ifeachor","sequence":"additional","affiliation":[{"name":"School of Engineering, Computing and Mathematics, University of Plymouth","place":["Plymouth, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yinghui","family":"Wei","sequence":"additional","affiliation":[{"name":"University of Exeter Medical School, Faculty of Health and Life Sciences, University of Exeter","place":["Exeter, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shang-Ming","family":"Zhou","sequence":"additional","affiliation":[{"name":"School of Nursing and Midwifery, University of Plymouth","place":["Plymouth, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Stephen","family":"Mullin","sequence":"additional","affiliation":[{"name":"Peninsula Medical School, University of Plymouth","place":["Plymouth, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,30]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"e60280","DOI":"10.7759\/cureus.60280","article-title":"Advancements in artificial intelligence for the diagnosis of multidrug resistance and extensively drug-resistant tuberculosis: a comprehensive review","volume":"16","author":"Geethalakshmi","year":"2024","journal-title":"Cureus"},{"key":"B2","volume-title":"Global Tuberculosis Report 2025","year":"2025"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1374703","DOI":"10.3389\/fpubh.2024.1374703","article-title":"Opinion review of drug-resistant tuberculosis in West Africa: tackling the challenges for effective control","volume":"12","author":"Otchere","year":"2024","journal-title":"Front Public Health"},{"key":"B4","doi-asserted-by":"publisher","first-page":"1510026","DOI":"10.3389\/fmicb.2024.1510026","article-title":"Revolutionizing diagnosis of pulmonary Mycobacterium tuberculosis based on CT: a systematic review of imaging analysis through deep learning","volume":"15","author":"Zhang","year":"2025","journal-title":"Front Microbiol"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1773","DOI":"10.1038\/s41591-022-01981-2","article-title":"Multimodal biomedical AI","volume":"28","author":"Acosta","year":"2022","journal-title":"Nat Med"},{"key":"B6","doi-asserted-by":"publisher","first-page":"62198","DOI":"10.1109\/ACCESS.2025.3558878","article-title":"A systematic review on application of multimodal learning and explainable AI in tuberculosis detection","volume":"13","author":"Nansamba","year":"2025","journal-title":"IEEE Access"},{"key":"B7","doi-asserted-by":"publisher","first-page":"17","DOI":"10.1007\/s12553-023-00727-5","article-title":"ML technologies for diagnosing and treatment of tuberculosis: a survey","volume":"13","author":"Jonathan","year":"2023","journal-title":"Health Technol (Berl)"},{"key":"B8","doi-asserted-by":"publisher","first-page":"1075","DOI":"10.1016\/j.cmi.2024.05.006","article-title":"Chest X-ray alone is insufficient for predicting drug-resistant pulmonary tuberculosis","volume":"30","author":"Pedersen","year":"2024","journal-title":"Clin Microbiol Infect"},{"key":"B9","author":"Mnyambo","year":""},{"key":"B10","doi-asserted-by":"publisher","first-page":"e70246","DOI":"10.1002\/mp.70246","article-title":"An exploratory study on integrating radiomics with vision transformers for enhancing medical imaging classification accuracy","volume":"53","author":"Yang","year":"2026","journal-title":"Med Phys"},{"key":"B11","author":"Yang","year":""},{"key":"B12","author":"Touvron","year":""},{"key":"B13","doi-asserted-by":"publisher","first-page":"3267","DOI":"10.1128\/JCM.01013-17","article-title":"The TB portals: an open-access, web-based platform for global drug-resistant-tuberculosis data sharing and analysis","volume":"55","author":"Rosenthal","year":"2017","journal-title":"J Clin Microbiol"},{"key":"B14","author":"Mosca","year":""},{"key":"B15","doi-asserted-by":"publisher","first-page":"307","DOI":"10.1186\/s12880-024-01481-4","article-title":"A CT-based radiomics analysis for differentiating drug-resistant and drug-sensitive pulmonary tuberculosis","volume":"24","author":"Jiang","year":"2024","journal-title":"BMC Med Imaging"},{"key":"B16","doi-asserted-by":"publisher","first-page":"134","DOI":"10.1016\/j.jgar.2025.04.013","article-title":"Multimodal radiomics integrating deep learning and clinical features for diagnosing multidrug-resistant tuberculosis in HIV\/AIDS patients","volume":"43","author":"Song","year":"2025","journal-title":"J Glob Antimicrob Resist"},{"key":"B17","doi-asserted-by":"publisher","first-page":"675","DOI":"10.21037\/qims-21-290","article-title":"Differentiating between drug-sensitive and drug-resistant tuberculosis with machine learning for clinical and radiological features","volume":"12","author":"Yang","year":"2022","journal-title":"Quant Imaging Med Surg"},{"key":"B18","author":"Yang","year":""},{"key":"B19","doi-asserted-by":"publisher","first-page":"1004","DOI":"10.1186\/s12879-025-11439-8","article-title":"Prevalence and contributing factors of drug-resistant tuberculosis (DR-TB) in Iran: a systematic review","volume":"25","author":"Daneshi","year":"2025","journal-title":"BMC Infect Dis"},{"key":"B20","doi-asserted-by":"publisher","first-page":"84228","DOI":"10.1109\/ACCESS.2023.3298750","article-title":"Combining radiological and genomic TB portals data for drug resistance analysis","volume":"11","author":"Bui","year":"2023","journal-title":"IEEE Access"},{"key":"B21","doi-asserted-by":"publisher","first-page":"1508736","DOI":"10.3389\/fmed.2025.1508736","article-title":"Prediction of active drug-resistant pulmonary tuberculosis based on CT radiomics: construction and validation of independent models and combined models for residual pulmonary parenchyma","volume":"12","author":"Liu","year":"2025","journal-title":"Front Med (Lausanne)"},{"key":"B22","doi-asserted-by":"publisher","first-page":"343","DOI":"10.3390\/computers13120343","article-title":"Hybrid vision transformer and convolutional neural network for multi-class and multi-label classification of tuberculosis anomalies on chest X-ray","volume":"13","author":"Yulvina","year":"2024","journal-title":"Computers"},{"key":"B23","doi-asserted-by":"publisher","first-page":"96","DOI":"10.1186\/s12880-025-01630-3","article-title":"Enhanced tuberculosis detection using vision transformers and explainable AI with Grad-CAM on chest X-rays","volume":"25","author":"Vanitha","year":"2025","journal-title":"BMC Med Imaging"},{"key":"B24","doi-asserted-by":"publisher","first-page":"7","DOI":"10.1186\/s44398-025-00008-3","article-title":"PaliGemma-CXR: a multi-task multimodal model for tuberculosis chest X-ray interpretation","volume":"1","author":"Musinguzi","year":"2025","journal-title":"BMC Artif Intell"},{"key":"B25","doi-asserted-by":"publisher","first-page":"1834","DOI":"10.3390\/sym15101834","article-title":"Radiomics and its feature selection: a review","volume":"15","author":"Zhang","year":"2023","journal-title":"Symmetry"},{"key":"B26","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1016\/j.imed.2022.07.002","article-title":"Transformers in medical image analysis","volume":"3","author":"He","year":"2023","journal-title":"Intell Med"},{"key":"B27","doi-asserted-by":"publisher","first-page":"1393","DOI":"10.1007\/s12559-024-10259-3","article-title":"TB-CXRNET: Tuberculosis and drug-resistant tuberculosis detection technique using chest X-ray images","volume":"16","author":"Rahman","year":"2024","journal-title":"Cognit Comput"},{"key":"B28","doi-asserted-by":"publisher","first-page":"e71350","DOI":"10.1002\/hsr2.71350","article-title":"Accuracy of machine learning in identifying drug resistance in tuberculosis: A systematic review and meta-analysis","volume":"8","author":"Wei","year":"2025","journal-title":"Health Sci Rep"},{"key":"B29","doi-asserted-by":"publisher","first-page":"vbaf050","DOI":"10.1093\/bioadv\/vbaf050","article-title":"Machine learning-based approach for identification of new resistance associated mutations from whole genome sequences of mycobacterium tuberculosis","volume":"5","author":"Pal","year":"2025","journal-title":"Bioinform Adv"},{"key":"B30","doi-asserted-by":"publisher","first-page":"48","DOI":"10.1186\/s12863-025-01338-x","article-title":"Machine learning-based prediction of antimicrobial resistance and identification of AMR-related SNPs in Mycobacterium tuberculosis","volume":"26","author":"Xu","year":"2025","journal-title":"BMC Genomic Data"},{"key":"B31","doi-asserted-by":"publisher","first-page":"43","DOI":"10.1080\/10696679.2024.2376052","article-title":"How to treat missing data in survey research","volume":"33","author":"Popovich","year":"2025","journal-title":"J Mark Theory Pract"},{"key":"B32","doi-asserted-by":"publisher","first-page":"012001","DOI":"10.1088\/1742-6596\/2071\/1\/012001","article-title":"Identifying drug-resistant tuberculosis from chest X-ray images using a simple convolutional neural network","volume":"2071","author":"Ureta","year":"2021","journal-title":"J Phys Conf Ser"},{"key":"B33","doi-asserted-by":"publisher","first-page":"109025","DOI":"10.1016\/j.isci.2024.109025","article-title":"Integrative analysis of multimodal patient data identifies personalized predictors of tuberculosis treatment prognosis","volume":"27","author":"Sambarey","year":"2024","journal-title":"iScience"},{"key":"B34","doi-asserted-by":"publisher","first-page":"570","DOI":"10.1109\/TPDS.2025.3526238","article-title":"A comparative study of sampling methods with cross-validation in the fedhome framework","volume":"36","author":"Ahmadi","year":"2025","journal-title":"IEEE Trans Parallel Distrib Syst"},{"key":"B35","doi-asserted-by":"publisher","first-page":"665","DOI":"10.1186\/s12885-025-14040-z","article-title":"Evaluating key predictors of breast cancer through survival: a comparison of aft frailty models with lasso, ridge, and elastic net regularization","volume":"25","author":"Bosson-Amedenu","year":"2025","journal-title":"BMC Cancer"},{"key":"B36","doi-asserted-by":"publisher","first-page":"993","DOI":"10.1016\/j.aej.2022.10.069","article-title":"A novel multi-level feature selection method for radiomics","volume":"66","author":"Wang","year":"2023","journal-title":"Alex Eng J"},{"key":"B37","doi-asserted-by":"publisher","first-page":"433","DOI":"10.1186\/s12859-019-2994-z","article-title":"An elastic-net logistic regression approach to generate classifiers and gene signatures for types of immune cells and T helper cell subsets","volume":"20","author":"Torang","year":"2019","journal-title":"BMC Bioinf"},{"key":"B38","doi-asserted-by":"publisher","first-page":"e0246159","DOI":"10.1371\/journal.pone.0246159","article-title":"HDSI: high dimensional selection with interactions algorithm on feature selection and testing","volume":"16","author":"Jain","year":"2021","journal-title":"PLoS ONE"},{"key":"B39","doi-asserted-by":"publisher","first-page":"2012","DOI":"10.1093\/bioinformatics\/btaa535","article-title":"Predictive and interpretable models via the stacked elastic net","volume":"37","author":"Rauschenberger","year":"2021","journal-title":"Bioinformatics"},{"key":"B40","doi-asserted-by":"publisher","first-page":"301","DOI":"10.1111\/j.1467-9868.2005.00503.x","article-title":"Regularization and variable selection via the elastic net","volume":"67","author":"Zou","year":"2005","journal-title":"J R Stat Soc Ser B Stat Methodol"},{"key":"B41","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18637\/jss.v106.i01","article-title":"Elastic net regularization paths for all generalized linear models","volume":"106","author":"Tay","year":"2023","journal-title":"J Stat Softw"},{"key":"B42","doi-asserted-by":"publisher","first-page":"197","DOI":"10.2147\/IJNRD.S461028","article-title":"Comparative analysis of logistic regression, gradient boosted trees, SVM, and random forest algorithms for prediction of acute kidney injury requiring dialysis after cardiac surgery","volume":"17","author":"Omar","year":"2024","journal-title":"Int J Nephrol Renovasc Dis"},{"key":"B43","first-page":"26","article-title":"Predictive modeling of obesity and cardiovascular disease risk: a random forest approach","volume":"7","author":"Nasser","year":"2023","journal-title":"Int J Acad Inf Syst Res (IJAISR)"},{"key":"B44","doi-asserted-by":"publisher","first-page":"e0298102","DOI":"10.1371\/journal.pone.0298102","article-title":"Brain tumor classification in VIT-B\/16 based on relative position encoding and residual MLP","volume":"19","author":"Hong","year":"2024","journal-title":"PLoS ONE"},{"key":"B45","doi-asserted-by":"publisher","first-page":"35716","DOI":"10.1109\/ACCESS.2024.3371151","article-title":"Classification of paediatric pneumonia using modified DenseNet-121 deep-learning model","volume":"12","author":"Arulananth","year":"2024","journal-title":"IEEE Access"},{"key":"B46","author":"Sankari","year":""},{"key":"B47","author":"Liang","year":""},{"key":"B48","author":"Netayawijit","year":""},{"key":"B49","doi-asserted-by":"publisher","first-page":"e70056","DOI":"10.1111\/cts.70056","article-title":"Practical guide to SHAP analysis: explaining supervised machine learning model predictions in drug development","volume":"17","author":"Ponce-Bobadilla","year":"2024","journal-title":"Clin Transl Sci"},{"key":"B50","doi-asserted-by":"publisher","first-page":"62","DOI":"10.1186\/s13017-025-00635-1","article-title":"A radiomics-clinical predictive model for difficult laparoscopic cholecystectomy based on preoperative CT imaging: a retrospective single center study","volume":"20","author":"Sun","year":"2025","journal-title":"World J Emerg Surg"},{"key":"B51","author":"Mitra","year":""},{"key":"B52","doi-asserted-by":"publisher","first-page":"109569","DOI":"10.1016\/j.compbiomed.2024.109569","article-title":"Local interpretable model-agnostic explanation approach for medical imaging analysis: a systematic literature review","volume":"185","author":"Hassan","year":"2025","journal-title":"Comput Biol Med"},{"key":"B53","first-page":"S291","article-title":"Utilizing radiological images for predicting drug resistance of lung tuberculosis","volume":"10","author":"Kovalev","year":"2015","journal-title":"Int J Comput Assist Radiol Surg"},{"key":"B54","doi-asserted-by":"publisher","first-page":"1915","DOI":"10.1007\/s11548-018-1857-9","article-title":"Detecting drug-resistant tuberculosis in chest radiographs","volume":"13","author":"Jaeger","year":"2018","journal-title":"Int J Comput Assist Radiol Surg"},{"key":"B55","author":"Tulo","year":""},{"key":"B56","doi-asserted-by":"publisher","first-page":"236","DOI":"10.1002\/msd2.70017","article-title":"Explainable artificial intelligence (XAI) for material design and engineering applications: a quantitative computational framework","volume":"5","author":"Liu","year":"2025","journal-title":"Int J Mech Syst Dyn"},{"key":"B57","doi-asserted-by":"publisher","first-page":"469","DOI":"10.1016\/j.jinf.2018.10.004","article-title":"Risk factors of multidrug-resistant tuberculosis: a global systematic review and meta-analysis","volume":"77","author":"Pradipta","year":"2018","journal-title":"J Infect"},{"key":"B58","doi-asserted-by":"publisher","first-page":"161","DOI":"10.21037\/qims.2018.03.06","article-title":"Radiological signs associated with pulmonary multi-drug resistant tuberculosis: an analysis of published evidence","volume":"8","author":"W\u00e1ng","year":"2018","journal-title":"Quant Imaging Med Surg"},{"key":"B59","doi-asserted-by":"publisher","first-page":"e044349","DOI":"10.1136\/bmjopen-2020-044349","article-title":"Risk factors for drug-resistant tuberculosis, the association between comorbidity status and drug-resistant patterns: a retrospective study of previously treated pulmonary tuberculosis in Shandong, China, during 2004\u20132019","volume":"11","author":"Tao","year":"2021","journal-title":"BMJ Open"},{"key":"B60","doi-asserted-by":"publisher","first-page":"e0220946","DOI":"10.1371\/journal.pone.0220946","article-title":"Clinical predictors of drug-resistant tuberculosis in Mexico","volume":"14","author":"Flores-Trevi\u00f1o","year":"2019","journal-title":"PLoS ONE"},{"key":"B61","doi-asserted-by":"publisher","first-page":"84","DOI":"10.5083\/ejcm\/25-02-10","article-title":"Evaluation of imaging features of drug-sensitive and drug-resistant pulmonary tuberculosis","volume":"15","author":"Shukla","year":"2025","journal-title":"Eur J Cardiovasc Med"},{"key":"B62","doi-asserted-by":"publisher","first-page":"317","DOI":"10.3947\/ic.2016.48.4.317","article-title":"Association between body mass index and sputum culture conversion among South Korean patients with multidrug resistant tuberculosis in a tuberculosis referral hospital","volume":"48","author":"Park","year":"2016","journal-title":"Infect Chemother"},{"key":"B63","doi-asserted-by":"publisher","first-page":"4228","DOI":"10.1002\/pc.70421","article-title":"An archimedes optimization algorithm based extreme gradient boosting model for predicting the bending strength of uv cured glass fiber reinforced polymer composites","volume":"47","author":"Zhang","year":"2026","journal-title":"Polym Compos"}],"container-title":["Frontiers in Digital Health"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1828522\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,31]],"date-time":"2026-07-31T13:43:38Z","timestamp":1785505418000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdgth.2026.1828522\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,30]]},"references-count":63,"alternative-id":["10.3389\/fdgth.2026.1828522"],"URL":"https:\/\/doi.org\/10.3389\/fdgth.2026.1828522","relation":{},"ISSN":["2673-253X"],"issn-type":[{"value":"2673-253X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,30]]},"article-number":"1828522"}}