Budget impact of implementing AI-enabled chest X-ray based incidental pulmonary nodule detection for early lung cancer diagnosis: models from Colombia, Costa Rica, Mexico, Thailand and Vietnam

OBJECTIVE: To assess the budget impact of incidental pulmonary nodule (IPN) detection using an artificial intelligence-software for chest X-ray (CXR) interpretation - qXR - for early lung cancer (LC) detection in Vietnam, Colombia, Thailand, Costa Rica, and Mexico. METHODS: Country-specific hybrid decision-tree and budget impact models were developed from the public payer perspective over a 5-year time horizon. Each compared current symptomatic LC detection with a scenario adding AI-enabled Chest X-Ray for IPN detection and referral to low-dose computed tomography. Costs included diagnostics, set-up, treatment, healthcare resource use, and end-of-life care, expressed in 2024 USD. One-way sensitivity and scenario analyses tested parameter and structural uncertainty. RESULTS: Over the 5-year time horizon, the implementation of AI-enabled CXR for IPN detection is estimated to avert 24,763 premature deaths across the five countries, improving survival and lowering long-term expenditures. Following the introduction of qXR, more costs were incurred due to additional patients being diagnosed and treated. This resulted in fewer late-stage patients and cost savings. The breakeven point was reached in year 3 for Vietnam, Thailand, Colombia, and Mexico and in year 4 for Costa Rica. Cost savings were primarily driven by the stage shift towards earlier detection, where treatment costs are lower and survival outcomes are better. CONCLUSIONS: Integrating AI-enabled CXR for IPN detection into national hospital workflows may enable earlier LC detection and achieve cost neutrality within 5 years in low- and middle-income countries. These findings support the economic feasibility and scalability of AI-assisted imaging as part of national lung cancer control strategies.