Abstract P3-09-09: Assessing utility of breast cancer risk assessment tool in comparison to Tyrer-Cuzick model for determination of breast cancer risk and implications for chemoprevention

Abstract

Abstract Background Despite findings that the Tyrer-Cuzick (IBIS Breast Cancer Risk Evaluation Tool or TC) model is more predictive of breast cancer risk than the Gail model (NCI maintained Breast Cancer Risk Assessment Tool or BCRAT), BCRAT is commonly clinically used as per the United States Preventive Services Task Force (USPSTF), with a 5-year risk for breast cancer (BC) of greater than 3% on BCRAT, the benefits of preventive medication likely outweigh the risks. We aimed to compare the models, 1: to see if a 10 year risk estimate per the TC model reliably correlated with the 3% 5 year risk per BCRAT, and 2: to analyze the subset of patients with atypical hyperplasia (AH) and lobular carcinoma in situ (LCIS) who are known both to be at high risk for breast cancer and to benefit from chemoprevention. Our hypothesis is that BCRAT has limited utility in risk estimation, and the most comprehensive model for risk estimation and clinical decision making is TC. Methods 200 women ages 35-64 women followed in benign breast clinic were included. Risk estimations were run using BCRAT, TC version 7 (v7) and TC version 8 (v8). A Pearson's Correlation test was conducted to investigate the relationship between the TC models and the BCRAT model. A p-value < 0.05 was considered statistically significant. Results Analysis showed a positive moderate-strength relationship between the TC v7 10-year risk and the 5-year BCRAT risk for this population (R = 0.468, P<0.001) and a positive moderate-strength relationship between the TC v8 10-year risk and the 5-year BCRAT risk (R = 0.550, P<0.001). A TC v7 risk of 8.09% (95% confidence interval (CI): 7.42-8.75) and a TC v8 risk of 8.54% (95% CI: 7.85-9.24) corresponded to a BCRAT risk of 3%. However, much error was present when assessing consistency and correlation between the models. A total of 36 patients were diagnosed with AH, 2 patients were diagnosed with LCIS and 7 patients were diagnosed with both AH and LCIS. 11 patients who had AH had an estimated 5-year risk per the BCRAT model of <3%. Two of these patients were pre-menopausal and African American and one was pre-menopausal and Hispanic. Of the remaining 8 patients, all were under the age of 60. Of the 30 patients who had a BCRAT 5-year estimated risk of BC of >3% but no AH or LCIS, 12 had two first degree relatives with breast cancer and 16 had a first-degree relative with BC and at least two benign breast biopsies. Conclusion BCRAT is limited and caution is warranted with its use in assessing risk and for counseling around chemoprevention benefit. There is not reliable correlation between the 5 year BCRAT risk estimate and the 10 year TC risk estimate. Chemoprevention should be discussed for patients with AH, LCIS or 2+ first degree relatives with breast cancer. Further, BCRAT may underestimate risk in minority populations and others with AH. For a limited group of patients with moderate risk, dual modeling may be clinically useful in making chemopreventive recommendations. Citation Format: Pederson HJ, Yanda C, Kline M, Stephens M, Goraya ST, Grobmyer SR, Kattan MW. Assessing utility of breast cancer risk assessment tool in comparison to Tyrer-Cuzick model for determination of breast cancer risk and implications for chemoprevention [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P3-09-09.