Advances in Tuberculosis Diagnosis, Pathophysiology, Treatment, and Therapy Duration (2020–2025): An Umbrella Review of Systematic Reviews and Meta-Analyses

Elmukhtar Habas¹, Amnna Rayani², Ala Habas³, Mehdi Errayes⁴

Keywords: HIV, coinfection, systematic review, diagnosis, pathophysiology, treatment, therapy duration, immunocompromised patients, ML, AI

DOI: 10.63475/yjm.v5i1.0323

DOI URL: https://doi.org/10.63475/yjm.v5i1.0323

Publish Date: 22-04-2026

Download PDF

Citation: 0

Author Affiliation:

1 Professor/Senior Consultant, Department of Medicine, Hamad General Hospital, Qatar University, Doha, Qatar
2 Professor, University of Tripoli, Tripoli, Libya
3 Specialist, Tripoli Central Hospital, University of Tripoli, Tripoli, Libya
4. Senior Consultant, Department of Medicine, Hamad General Hospital, Doha, Qatar

Abstract

Tuberculosis (TB) remains a major global health threat despite advances in prevention and care, particularly among immunocompromised individuals and those with HIV coinfection. This umbrella review of systematic reviews and meta-analyses (January 2020–December 2025 synthesizes recent evidence on TB pathophysiology, diagnosis, treatment, and therapy duration, with emphasis on machine learning/artificial intelligence (ML/AI) applications and vulnerable populations. We searched PubMed, Scopus, Web of Science, and EMBASE for systematic reviews and meta-analyses using terms related to TB pathophysiology, diagnosis, treatment, therapy duration, and immunocompromised management, including ML/AI applications. Inclusion criteria: English-language publications, focus on human studies, and restriction from January 2020 to December 2025. As this is an umbrella review of existing systematic reviews and metaanalyses, findings were synthesized narratively. No new pooled analyses, meta-analyses, or meta-regressions were performed. Quantitative estimates are reproduced as reported in the source meta-analyses and attributed to those sources. Of 2000 identified records, 75 were included after screening and eligibility assessment (see PRISMA flow diagram). Significant findings encompass a deeper comprehension of immune dynamics in pathophysiology, increased diagnostic precision through AI and molecular methodologies, safer rifamycin-based therapies, and shorter treatment regimens demonstrating non-inferiority. AI and ML play a crucial role in enhancing predictive analytics and enabling personalization. In immunocompromised patients, HIV coinfection markedly deteriorates outcomes, while customized treatment regimens and AIbased predictions contribute to decreased mortality rates. Recent improvements in TB therapies enable the development of shorter and safer TB regimens, combined with integrated care for coinfections. Furthermore, including AI and ML has improved precision. However, challenges persist in resource-limited environments. Future research must emphasize implementation studies.

---