Optimizing deep learning-enhanced opportunistic osteoporosis screening from chest X-rays in Australia: Impact of follow-up and treatment adherence on clinical and economic outcomes

Abstract

Opportunistic osteoporosis screening using artificial intelligence (AI) applied to routine chest radiographs offers a novel, low-burden strategy to improve early detection and treatment initiation. This study evaluated the cost-effectiveness of AI-enhanced opportunistic osteoporosis screening compared with no screening among Australian women aged ≥50 years undergoing chest radiography. A microsimulation-based Markov model was developed to estimate quality-adjusted life years (QALYs) and lifetime costs in 2025 Australian dollars. The model incorporated the diagnostic performance of AI applied to chest radiographs, subsequent DXA confirmation, treatment initiation, and real-world medication persistence. It also explored the effect of enhanced follow-up and adherence on clinical and economic outcomes.

AI-enhanced opportunistic screening was associated with improved health outcomes (+0.0025 QALYs and − 0.0043 fractures per patient) and lower lifetime costs (–AUD59) compared with no screening, representing a dominant strategy. When scaled to 10,000 screened women aged ≥50 years, the model projected prevention of 43 fractures, 25 additional life-years, 25 QALYs gained, and AUD590,000 in healthcare savings. Optimizing follow-up and adherence could amplify clinical benefits by four- to five-fold.

AI-enhanced opportunistic osteoporosis screening from chest X-rays appears to be a cost-effective strategy within the Australian healthcare system, delivering better outcomes at lower cost. The results also suggest that linkage to DXA confirmation and subsequent treatment initiation may influence the magnitude of health and economic outcomes. These findings highlight that follow-up and persistence are not merely complementary but essential for realizing the full potential of cost-effective AI-driven osteoporosis screening.

More