Mapping The Intellectual Structure of AI in Tuberculosis Research: A Keyword Co-occurrence and Bibliometric Analysis
Abstract
Artificial intelligence (AI) has become increasingly essential in global tuberculosis (TB) control, offering advances in diagnostic accuracy, drug-resistance prediction, treatment monitoring, and public health decision-making. As research activity accelerates, a structured bibliometric assessment is necessary to map scientific progress and highlight emerging directions. This study conducted a bibliometric analysis using data retrieved from Scopus, from which 57 articles were included after rigorous cleaning and screening. Citation metrics were computed using Publish or Perish, while VOSviewer was employed to visualize keyword co-occurrence networks. Citation impact was evaluated using total local citation score (TLCS) and total global citation score (TGCS). Findings show that annual publication output remained low until 2023 but increased sharply from 2023 to 2025, reflecting rapid global expansion of AI-related TB research. The Indian Journal of Tuberculosis and the International Journal of Biomedical Imaging emerged as influential journals, with the latter demonstrating the highest TGCS, underscoring its broad international impact. Despite this global momentum, collaboration among authors remains limited, with several isolated research clusters rather than a unified network. Keyword analysis revealed major themes centered on AI-assisted TB prediction, AI for monitoring treatment responses and drug-resistant TB patterns, machine-learning models developed for TB screening, and digital platforms and remote health technologies into TB management and disease prevention. Overall, AI research in TB is expanding rapidly, driven by global recognition of AI as a critical instrument for accelerating detection, optimizing treatment strategies, and strengthening TB control efforts worldwide. Keywords: Artificial intelligence, tuberculosis management, bibliometric analysis, predictive machine learning