Predictors of surveillance mammography outcomes in women with a personal history of breast cancer

Kathryn P. Lowry, Lior Z. Braunstein, Konstantinos P. Economopoulos, Laura Salama, Constance D. Lehman, G. Scott Gazelle, Elkan F. Halpern, Catherine S. Giess, Alphonse G. Taghian, Janie M. Lee

To identify predictors of poor mammography surveillance outcomes based on clinico-pathologic features. This study was HIPAA compliant and IRB approved. We performed an electronic medical record review for a cohort of women with American Joint Committee on Cancer (AJCC) Stage I or II invasive breast cancer treated with breast conservation therapy who developed subsequent in-breast treatment recurrence (IBTR) or contralateral breast cancer (CBC). Poor surveillance outcome was defined as second breast cancer not detected by surveillance mammography, including interval cancers (diagnosed within 365 days of surveillance mammogram with negative results) and clinically detected cancers (diagnosed without a surveillance mammogram in the preceding 365 days). Univariate and multivariate logistic regression were performed to identify predictors of poor mammography surveillance outcome, including patient and primary tumor characteristics, breast density, mode of primary tumor detection, and time to second cancer diagnosis. 164 women met inclusion criteria (65 with IBTR, 99 with CBC); 124 had screen-detected second cancers. On univariate analysis, poor surveillance outcome (n = 40) was associated with age at primary cancer diagnosis < 50 years (p < 0.0001), AJCC stage II primary cancers (p = 0.007), and heterogeneously or extremely dense breasts (p = 0.04). On multivariate analysis, age < 50 years at primary breast cancer diagnosis remained a significant predictor of poor surveillance outcome (p = 0.001). Women younger than age 50 at primary breast cancer diagnosis are at risk of poor surveillance mammography outcomes, and may be appropriate candidates for more intensive clinical and imaging surveillance.