Economic analysis of different throughput scenarios and implementation strategies of computer-aided detection software as a screening and triage test for pulmonary TB

This study evaluates the cost-effectiveness of Computer-Aided Detection (CAD) software in screening and triage for tuberculosis (TB), comparing its cost with radiologist-based readings in Pakistan’s national scale-up plan. The study analyzed four CAD software (CAD4TB, Lunit INSIGHT CXR, qXR, InferRead) using different pricing models and estimated costs based on active case finding (ACF) and facility-based CXR testing scenarios. Findings showed that CAD4TB and InferRead had the lowest per-screen cost (0.19–2.78 USD), whereas Lunit and qXR were comparable to radiologist costs (0.70–0.93 USD) only at high throughput levels. The projected cost for nationwide CAD implementation over four years ranged from 2.65–19.23 million USD, significantly lower than the 23.97 million USD required for radiologist-based screening. The study suggests that CAD software is a cost-effective solution for large-scale TB screening in high-burden countries, but selection should consider software pricing structures, screening volume, and operational efficiency to optimize cost savings and diagnostic coverage.