Cost-effectiveness of AI-enhanced breast cancer screening in Iran: a microsimulation study.

Abstract

Background: Recognized worldwide as a major health threat and Sustainable Development Goal priority, breast cancer places an especially heavy clinical and economic burden on Iran, making early detection imperative. To guide policy, we assessed the cost-effectiveness of ten mammography-based screening schedules.

Methods: We developed an individual-level microsimulation model simulating 10,000 Iranian women over a 40-year horizon. Screening scenarios varied by interval (annual, biennial, triennial), incorporating mammography alone and two AI-enhanced strategies. Analyses adopted a payer perspective, discounting costs and quality-adjusted life years (QALYs) at 5% annually. Incremental cost-effectiveness ratios (ICERs) were evaluated against a willingness-to-pay (WTP) threshold of USD 4,500 per QALY (Iran's 2023 GDP per capita). Probabilistic sensitivity analyses (PSA) assessment ensured result robustness.

Results: Compared with no screening, all screening schedules increased QALYs and total costs, while reducing breast-cancer incidence and mortality. Among the alternative tests, annual mammography interpreted by AI and a radiologist was the most cost-effective strategy, with an ICER of USD 4,064 per QALY, demonstrating a high level of cost-effectiveness relative to the next most effective strategy. Probabilistic sensitivity analysis confirmed robustness, with the selected strategy being cost-effective in 52% of simulations.

Conclusions: AI-enhanced mammography is more cost-effective than conventional screening. Concurrent AI-radiologist double reading delivers the greatest value at Iran's USD 4,500 /QALY threshold, while AI-triage with radiologist confirmation, though less efficient, offers a practical alternative in settings where imaging and biopsy capacity cannot be rapidly expanded.