Cost-effectiveness of opportunistic osteoporosis screening using artificial-intelligence assisted chest radiographs in Japan

Abstract

Objectives: Osteoporosis is often undiagnosed due to the shortcomings of conventional screening, resulting in missed chances for early treatment. Advances in artificial intelligence (AI), particularly deep learning applied to chest X-rays, offer a new opportunity for opportunistic screening. This study assesses the cost-effectiveness of this approach in Japanese women aged ≥ 50 years. Methods: An economic model estimated the cost per quality-adjusted life year (QALY) gained (in 2024 Japanese Yen, ¥) for a strategy involving AI-assisted chest X-ray screening followed by treatment, compared to no screening. Patient trajectories were modeled using the AI system's diagnostic performance and aligned with the Japanese osteoporosis guidelines. Analyses were conducted for Japan overall, in Kure City (a high-fracture-incidence area), and in a lower-incidence scenario. Real-world medication persistence, the probabilities of dual-energy X-ray absorptiometry examination after screening detection, and treatment initiation rates were incorporated. Results: Nationwide in Japan, the cost per QALY gained from opportunistic osteoporosis screening was estimated at ¥189,713 for women aged ≥ 50, substantially lower than the accepted cost-effectiveness threshold of ¥5 million. In Kure City, opportunistic screening was dominant (lower total costs for more QALYs). In the lower-incidence scenario, 25% below the national average, the cost per QALY was ¥1,055,095, remaining below the threshold. Results were robust across all age-specific populations and in sensitivity analyses. Conclusions: Leveraging AI-assisted chest X-rays for incidental osteoporosis detection demonstrates strong economic viability for older Japanese women. This approach also proves to be a dominant strategy in areas with elevated fracture rates.

Ähnliche Arbeiten