COST-EFFECTIVENESS OF AI-POWERED PRE-SCREENING FOR GERMLINE BRCA TESTING IN HR+/HER2-NEGATIVE EARLY BREAST CANCER IN FRANCE
Author(s)
Sukhvinder S. Johal, PhD1, Thomas Di Maio, MSc2, Arnaud Fouillet, MSc3, Esther Mouquet, MiM3, Hortense Deslandes, PharmD3.
1HTA & Modelling Science, AstraZeneca, Cambridge, United Kingdom, 2Diagnostics, AstraZeneca, Baar, Switzerland, 3Waiv, Paris, France.
1HTA & Modelling Science, AstraZeneca, Cambridge, United Kingdom, 2Diagnostics, AstraZeneca, Baar, Switzerland, 3Waiv, Paris, France.
OBJECTIVES: Germline BRCA mutation (gBRCAm) testing identifies HR+/HER2-negative early breast cancer (eBC) patients eligible for adjuvant olaparib, but testing remains suboptimal due to low mutation prevalence, next-generation sequencing (NGS) costs, and limited genetic counselling capacity. We evaluated the cost-effectiveness of AI-powered histopathology pre-screening to identify patients at elevated gBRCAm risk before confirmatory NGS testing in France.
METHODS: A decision-tree model compared current standard of care with AI pre-screening over a lifetime horizon. Under standard care, NGS testing rates were 70% for patients at high personal mutation risk and 40% for those at low risk. With pre-screening, 65% of low-risk patients underwent AI analysis of H&E whole-slide images before NGS referral; high-risk patients and remaining low-risk patients followed standard care. AI sensitivity and specificity were 93% and 41%, respectively. Costs, quality-adjusted life-years (QALYs), and life-years were derived from a French health technology assessment partitioned survival model for olaparib. The population included 38,600 HR+/HER2-negative eBC patients, 80% classified as low risk. AI pre-screening and NGS costs were €150 and €882.90, respectively.
RESULTS: AI pre-screening identified 303 additional gBRCAm patients versus standard care (1,214 vs. 911). Incremental costs were €13.7 million, with 344 additional QALYs and 421 additional life-years gained. The incremental cost-effectiveness ratio was €39,900/QALY, and incremental net monetary benefit was €3.5 million at a €50,000/QALY willingness-to-pay threshold. Tests needed to identify one gBRCAm-positive patient decreased from 19.49 to 17.67.
CONCLUSIONS: AI-powered histopathology pre-screening before confirmatory gBRCA NGS testing was cost-effective in France and improved identification of patients eligible for adjuvant olaparib. By increasing diagnostic yield and reducing tests per positive case, this strategy may support more efficient use of NGS and genetic counselling resources while expanding access to precision oncology.
METHODS: A decision-tree model compared current standard of care with AI pre-screening over a lifetime horizon. Under standard care, NGS testing rates were 70% for patients at high personal mutation risk and 40% for those at low risk. With pre-screening, 65% of low-risk patients underwent AI analysis of H&E whole-slide images before NGS referral; high-risk patients and remaining low-risk patients followed standard care. AI sensitivity and specificity were 93% and 41%, respectively. Costs, quality-adjusted life-years (QALYs), and life-years were derived from a French health technology assessment partitioned survival model for olaparib. The population included 38,600 HR+/HER2-negative eBC patients, 80% classified as low risk. AI pre-screening and NGS costs were €150 and €882.90, respectively.
RESULTS: AI pre-screening identified 303 additional gBRCAm patients versus standard care (1,214 vs. 911). Incremental costs were €13.7 million, with 344 additional QALYs and 421 additional life-years gained. The incremental cost-effectiveness ratio was €39,900/QALY, and incremental net monetary benefit was €3.5 million at a €50,000/QALY willingness-to-pay threshold. Tests needed to identify one gBRCAm-positive patient decreased from 19.49 to 17.67.
CONCLUSIONS: AI-powered histopathology pre-screening before confirmatory gBRCA NGS testing was cost-effective in France and improved identification of patients eligible for adjuvant olaparib. By increasing diagnostic yield and reducing tests per positive case, this strategy may support more efficient use of NGS and genetic counselling resources while expanding access to precision oncology.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
P22
Topic
Economic Evaluation, Medical Technologies, Study Approaches
Disease
Oncology, Personalized & Precision Medicine