PURPOSE: : The purpose of the study was to identify patient characteristics that predict for increased cardiac exposure through dosimetric analysis of the anatomy of a cohort of women treated with left-sided tangential breast radiation. Statistical analyses estimations for the appropriate sample sizes required for detection of significant differences in cardiac mortality at 15 years were conducted, assuming a threshold V25 for radiation-induced coronary artery disease (CAD) beyond which women are at risk for radiation-induced coronary artery disease.

METHODS AND MATERIALS: : Detailed heart dosimetry was recorded. Clinical factors (age, history of CAD, diabetes, receipt of cardiotoxic agents, weight/body mass index) and anatomic factors (heart volume, breast volume, cardiac contact distance) were recorded for each patient.

RESULTS: : The average heart V25 was 3.57%. The median percentage of the heart included in the tangential beam was 4.02%. There were no clinical or anatomic factors that predict suboptimal heart anatomy (ie, V25 of >/=6%) on multivariate analysis. The sample size calculations using thresholds for induction of CAD of V25 >/=1%, 6%, and 10% yielded sample sizes of 1314, 9504, and 61,342, respectively; considering node-positive breast cancer mortality and 15% loss to follow-up, these change to 2237, 16,166, and 104,334, respectively.

CONCLUSIONS: : Current studies with modern radiotherapy techniques would be underpowered to detect a difference in cardiac mortality where only some women are at risk. The heart, chest wall, and breast have a complex relationship to tangential breast radiation, and their interplay prevented this anatomic metric's success.