Abstract

Tuberculosis (TB) is a major public health challenge in Nigeria, where late diagnosis is common despite ongoing national control efforts. Chest X-ray is a sensitive screening tool, but its usefulness is limited by a severe shortage of radiologists, leading to delays and missed early cases. Artificial intelligence (AI) offers a promising way to support clinicians by rapidly identifying abnormalities suggestive of TB, especially in resource-limited settings where experts are scarce.

This paper presents a practical framework for introducing AI-assisted Chest X-ray screening into Nigerian hospitals. It outlines key steps for local implementation, including developing Nigerian CXR datasets, validating AI models on local populations, integrating AI into routine clinical workflow, and training frontline health workers to use these tools effectively. The framework also considers infrastructure needs, including power supply reliability, offline functionality, and secure data management.

AI has the potential to shorten turnaround time, improve consistency of CXR interpretation, and expand screening capacity to underserved communities. However, challenges such as model bias, maintenance costs, and data governance concerns must be addressed from the outset. By proposing a phased roadmap and clear recommendations for policymakers, hospitals, researchers, and industry partners, this work aims to guide responsible and sustainable adoption of AI in TB screening.

With careful adaptation to local realities, AI could play an important role in strengthening Nigeria’s TB diagnostic pathway and improving equitable access to early care.

Key words: Tuberculosis, Artificial Intelligence, Radiologists, Workflow, Hospitals