Artificial Intelligence for Tuberculosis Screening and Detection: From Evidence to Policy and Implementation

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Public Health

Background:

• Artificial intelligence (AI) is increasingly utilized for tuberculosis (TB) screening and diagnosis.
• Computer-aided detection (CAD) applied to chest radiography (CXR) is a prominent AI application in TB diagnostics.

Purpose of the Study:

• To conduct a state-of-the-art review of AI applications across the TB diagnosis pathway.
• To synthesize evidence on the programmatic value, diagnostic accuracy, and implementation of AI in TB care.

Main Methods:

• A narrative review synthesizing evidence from WHO policy documents, validation initiatives, and peer-reviewed studies (2010-2026).
• Selection process aligned with PRISMA principles for structured evidence synthesis.

Main Results:

• Computer-aided detection for chest radiography (CAD-CXR) is WHO-recommended for TB screening in specific contexts, showing comparable sensitivity to human readers.
• Implementation studies indicate improved screening efficiency and potential cost-effectiveness in high-burden settings.
• Other AI modalities show promise but require further validation; impact hinges on health system integration and patient-level outcomes.

Conclusions:

• AI holds significant potential to strengthen TB screening and diagnostic pathways.
• A differentiated approach is crucial: responsible scale-up of endorsed tools and rigorous evaluation of emerging technologies.
• Effective and equitable TB care depends on AI integration, considering programmatic value beyond diagnostic accuracy alone.