The blood biomarker panel identified 13% more breast cancer cases missed by Gail model and improved lung cancer risk classification with 93% specificity and 42% sensitivity.
Does a blood biomarker panel improve risk assessment for breast and lung cancers in women undergoing screening mammography?
A blood biomarker panel improves risk assessment for breast and lung cancers, including in never-smokers, beyond standard clinical models.
Absolute Event Rate: 0% vs 0%
Abstract Background: The MERIT cohort (Mammography, Early Detection, Risk Assessment, and Imaging Technologies) has enrolled 8,000 participants undergoing annual screening mammography at MD Anderson Cancer Center since 2017. The goal is to integrate clinical and imaging data with blood-based biomarker profiles to estimate risk of developing breast and other cancers. Annual blood collection enables evaluation of longitudinal algorithms for changes in biomarker levels. Here we report the performance of prespecified blood biomarker algorithms for assessing the risk of harboring or developing breast and lung cancers within one year of blood draw. Methods: Biomarker combination rules and decision thresholds were prespecified and locked. For longitudinal modeling, a parametric empirical Bayes (PEB) approach was applied to serial biomarker data to refine risk estimates. Results: Biomarker levels were assayed in 1,762 non-cancer controls, 221 breast cancer cases, and 21 lung cancer cases. Limiting the analyses to samples collected within one year of diagnosis included 104 breast cancer cases and 12 lung cancer cases. For breast cancer, using a 1.67% 5-year risk threshold (the Gail model “elevated-risk” cut-point), the biomarker panel classified an additional 13% of women with breast cancer at elevated risk who were missed by the Gail model. Incorporating 120 serial samples and applying the longitudinal algorithm further improved risk classification of breast cancer by an additional 9% without increasing the false-positive rate. For lung cancer, the biomarker panel achieved 93.0% specificity (95% CI: 89.0 to 96.0) and 42.0% sensitivity (95% CI: 33.0 to 50.0) at a prespecified 1% 6-year risk threshold, comparable to the U.S. Preventive Services Task Force eligibility criteria for low-dose CT (LDCT) screening. Among never-smokers, the panel classified 40% of women who developed lung cancer as elevated risk at 90% specificity. Incorporating 59 serial samples and applying the longitudinal algorithm resulted in a further 7% improvement in classification at fixed specificity. Conclusions: A blood biomarker panel for breast cancer risk assessment identified women at increased risk who were missed by the Gail model. A blood biomarker panel for lung cancer identified women at increased risk for lung cancer, including never-smokers who may benefit from LDCT screening. Longitudinal blood collection and application of PEB further improved risk assessment without increasing false positives, supporting the utility of blood testing for breast and lung cancer risks in the context of mammography screening. Citation Format: J. Dennison, E. Irajizad, O. Weaver, E. Ostrin, J. Fahrmann, J. Vykoukal, J. Leung, H. Khoshfekr Rudsari, S. Khoramisarvestani, N. Kettner, S. Hanash. A blood test for personalized breast and lung cancer risks: findings from the MD Anderson MERIT study abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS3-05-25.
Dennison et al. (Tue,) reported a other. The blood biomarker panel identified 13% more breast cancer cases missed by Gail model and improved lung cancer risk classification with 93% specificity and 42% sensitivity.