COST-UTILITY ANALYSIS OF AI-ASSISTED MAMMOGRAPHY FOR BREAST CANCER SCREENING IN THAILAND
Author(s)
Nichakorn Worakajit, PhD1, Nanthawan Pomkai, M.A.1, Khanit Pisawong, PharmD1, varalak srinonprasert, MD1, Pornpim Korpraphong, MD2, Hathaichanok Sumalee, M.S.W1, Pattara Leelahavarong, PhD1.
1Siriraj Health Policy Unit, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand, 2Department of Radiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand.
1Siriraj Health Policy Unit, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand, 2Department of Radiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand.
OBJECTIVES: Breast cancer screening access in Thailand remains limited due to radiologist workforce shortages. AI-assisted mammography (AI MMG) may help expand screening capacity. This study evaluated the cost-utility of Thai-developed AI MMG compared with standard guideline screening (mammography with targeted ultrasound for abnormal or dense breast findings) and current practice (mammography plus ultrasound for all women).
METHODS: A cost-utility analysis from government and societal perspectives was conducted using a decision tree linked to a Markov model. The decision tree captured screening outcomes, while the Markov model simulated lifetime progression between early-stage disease, late-stage disease, and death. Model inputs, including screening performance, transition probabilities, costs, and utility values, were derived from Thai medical audit data, Thanyarak Center records, and proprietary performance data provided by Thai AI MMG manufacturers (Inspectra Mammography Model). Costs and outcomes were discounted at 3% annually according to Thai HTA guidelines.
RESULTS: Among average-risk asymptomatic Thai women aged 40-70 years, AI MMG generated 16.54 discounted QALYs/patient, compared with 16.48 for current practice and 16.30 for standard guideline screening. Compared with the guideline approach, AI MMG improved health outcomes (+0.23 QALYs) and was cost-effective, with an incremental cost-effectiveness ratios (ICERs) of 12,155.54 THB (USD 370)/QALY gained, below Thailand’s willingness-to-pay threshold of 160,000 THB (USD 4,790)/QALY. AI MMG also reduced total costs from both government and societal perspectives compared with current practice and standard guideline screening by reducing unnecessary ultrasound referrals in AI-classified normal cases, thereby lowering imaging utilization and radiologist workload.
CONCLUSIONS: Thai-developed AI MMG may provide a cost-effective approach to expand breast cancer screening in Thailand. With high accuracy (sensitivity 96.6%; specificity 76.3%), AI MMG shows high potential for screening women with dense breasts and supporting scalable screening implementation. Integration into mobile screening services and radiologist-limited settings could improve access and reduce the financial burden on public health systems.
METHODS: A cost-utility analysis from government and societal perspectives was conducted using a decision tree linked to a Markov model. The decision tree captured screening outcomes, while the Markov model simulated lifetime progression between early-stage disease, late-stage disease, and death. Model inputs, including screening performance, transition probabilities, costs, and utility values, were derived from Thai medical audit data, Thanyarak Center records, and proprietary performance data provided by Thai AI MMG manufacturers (Inspectra Mammography Model). Costs and outcomes were discounted at 3% annually according to Thai HTA guidelines.
RESULTS: Among average-risk asymptomatic Thai women aged 40-70 years, AI MMG generated 16.54 discounted QALYs/patient, compared with 16.48 for current practice and 16.30 for standard guideline screening. Compared with the guideline approach, AI MMG improved health outcomes (+0.23 QALYs) and was cost-effective, with an incremental cost-effectiveness ratios (ICERs) of 12,155.54 THB (USD 370)/QALY gained, below Thailand’s willingness-to-pay threshold of 160,000 THB (USD 4,790)/QALY. AI MMG also reduced total costs from both government and societal perspectives compared with current practice and standard guideline screening by reducing unnecessary ultrasound referrals in AI-classified normal cases, thereby lowering imaging utilization and radiologist workload.
CONCLUSIONS: Thai-developed AI MMG may provide a cost-effective approach to expand breast cancer screening in Thailand. With high accuracy (sensitivity 96.6%; specificity 76.3%), AI MMG shows high potential for screening women with dense breasts and supporting scalable screening implementation. Integration into mobile screening services and radiologist-limited settings could improve access and reduce the financial burden on public health systems.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EE90
Topic
Economic Evaluation, Health Policy & Regulatory, Health Technology Assessment
Disease
Oncology