Imaging and AI Techniques in Intrapulmonary Tuberculosis Diagnosis and Management

Pulmonary tuberculosis (PTB) remains a major health threat worldwide, resulting in millions of deaths yearly, despite ongoing global, regional, and national efforts to eradicate and control this highly infectious disease. The lungs are the primary organs affected, which can be permanently damaged or spread to other body parts if not treated early, medical imaging crucial in diagnosis, assessment, staging, monitoring, and guidance. Chest X-ray (CXR) is the standard initial screening tool for suspected PTB, providing an overall view of structures within the chest. Computed tomography (CT) provides more insights into lymphadenopathy and early bronchogenic spread. Positron emission tomography–computed tomography (PET/CT) is being explored for determining treatment response. Radiological appearances of PTB alongside details from diverse research across these imaging modalities were evaluated. A targeted screening approach involving taking X-ray services down to the doorstep of many in TB-endemic countries, including strategies such as campaigning, mass screening, active case finding, and contact tracing, certainly yielded better results. AI introduction in imaging has surprisingly been instrumental to the recent success and giant strides made in addressing PTB. Empirical studies have demonstrated the remarkable performance of AI techniques such as machine learning (ML), deep learning (DL), natural language processing (NLP), expert systems (ES), robotics, and fuzzy logic (FL) in PTB imaging tasks – worklist management, triaging/prioritization, dose optimization, diagnosis/reporting, and treatment outcome prediction. However, while it is necessary for models to undergo robust training and validation, it is imperative to address growing ethical and regulatory concerns regarding responsible AI use.