Artificial Intelligence in Tuberculosis Imaging: A Global Bibliometric Analysis of Research Trends and Collaborations.
Authors
Affiliations (3)
Affiliations (3)
- Centre of Diagnostic Imaging, Therapeutic and Investigative Studies (CODTIS), Faculty of Health Sciences, The National University of Malaysia (UKM), 50300 Kuala Lumpur, Malaysia; Department of Imaging, Zunyi Bozhou District People's Hospital, Zunyi 563100, China.
- Centre of Diagnostic Imaging, Therapeutic and Investigative Studies (CODTIS), Faculty of Health Sciences, The National University of Malaysia (UKM), 50300 Kuala Lumpur, Malaysia. Electronic address: [email protected].
- Department of Radiology, Hospital Canselor Tuanku Muhriz, Jalan Yaacob Latif, Bandar Tun Razak, Cheras, 56000 Kuala Lumpur, Malaysia.
Abstract
Tuberculosis (TB), a leading infectious cause of death, remains a global health challenge. Imaging is central to diagnosis and screening, while artificial intelligence (AI) is increasingly applied to chest X-rays (CXR) and computed tomography (CT). However, no bibliometric study has comprehensively mapped publication trends, collaborations, modalities, technological evolution, and emerging research hotspots in AI-driven TB imaging. Publications on AI-based TB imaging were retrieved from Web of Science (WoS) and Scopus (2000 to July 2025), and a bibliometric analysis was performed on English-language articles and reviews. CiteSpace and VOSviewer were used to analyze publication trends, international and institutional collaborations, imaging modalities, keyword co-occurrence networks, clustering, and influential papers. A total of 556 publications were identified. Annual publication output increased rapidly, with a compound annual growth rate of 21.2 % from 2000 to 2024. India and China led in output. CXR was the dominant imaging modality. Technological progress followed three phases: segmentation-based computer-aided diagnosis (CAD) systems, the deep learning revolution, and clinical deployment of AI systems. Keyword co-occurrence clustering identified major research domains, including "machine learning," "computer-aided detection," and "tuberculosis detection." Influential studies demonstrate a clear shift from algorithm development to large-scale screening validation and clinical use. This bibliometric study reveals that AI in TB imaging has expanded rapidly, progressing from basic early CAD to clinically relevant deep learning-driven applications. CXR remains essential for large-scale screening. Future efforts emphasize interpretable and generalizable AI solutions to maximize global impact on TB diagnosis and control. The findings highlight the value of integrating AI-based imaging into clinical and public health practice. With international collaboration and technical support, CXR-based AI can facilitate TB detection in high-burden regions.