Assessment Of The Gail Model In A Cohort Of Women With Atypical Hyperplasia.

5371 Background: Understanding an individual woman’s risk of developing breast cancer is of high importance if we are to tailor clinical management properly. We sought to evaluate the performance of the Gail model1 in a cohort of women with atypical hyperplasia, and to determine if other histopathological features might contribute to enhanced risk prediction in this cohort. Methods: The Mayo Clinic Benign Breast Disease (BBD) Cohort includes 9343 women who had an open breast biopsy between 1967 and 1991.2 Of these, 336 women had atypical hyperplasia, a group with significantly increased risk of a later breast cancer (RR∼4.0). Gail model risk factors, and others, were obtained via survey and medical record review. Lifetime risk (thirty-year probability) of breast cancer was computed for each woman. Logistic regression was used to assess the concordance between the predicted and observed lifetime risk. Proportional hazards regression, with bootstrap model selection, was used to identify a potential risk prediction model for this high-risk group of women. Results: In this atypia sub-cohort, 64 women experienced a breast cancer with an average follow-up of about 15 years. This number of events was slightly lower than the number predicted by the 30-year Gail model probabilities (rate ratio [95% CI] = 0.94 [0.74 - 1.20]). At the individual level, the concordance between observed breast cancer events and predicted lifetime probabilities of breast cancer was 0.59. This did not reach statistical significance (p=0.13), however, the number of events was low. The model selection process identified one covariate that was associated with breast cancer risk in this sub-cohort: the number of foci of atypia. Conclusions: Averaging risks across this atypia cohort, the Gail model prediction was on target, but the per-individual concordance between observed and predicted breast cancer was low. Knowledge of the number of foci of atypia provided additional information about breast cancer risk. The development of alternative risk models in this group, and in the entire BBD cohort, are in process. References: 1. Gail MH, Brinton LA, Byar DP, Corle DK, Green SB, Schairer C, Mulvihill JJ. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. JNCI 1989, 81(24):1879-1886 2. Hartmann LC, Sellers TA, Frost MH, Lingle WL, Degnim AC, Ghosh K, Vierkant RA, Maloney SD, Pankratz VS, Hillman DW, Suman VJ, Johnson J, Blake C, Tlsty T, Vachon CM, Melton LJ, Visscher DW. Benign breast disease and the risk of breast cancer. N Engl J Med 2005, 353(3):229-237.