Abstract P369: Coronary Heart Disease and Mortality Following a Breast Cancer Diagnosis

Abstract

Background: Coronary heart disease (CHD) is a leading cause of death for breast cancer survivors. While cardiotoxic breast cancer treatments are linked with cardiovascular injury and an increased risk of CHD, current risk prediction models for CHD do not consider the effect of cancer treatments and likely underestimate CHD and mortality risk in this population. Methods: We conducted support vector machine, decision tree, and logistic regression models on electronic health record data to evaluate the independent and joint effects of cardiovascular health (CVH) and cancer treatments on post-treatment CHD and all-cause mortality occurring within 10 years among 1934 women diagnosed with breast cancer in 2006 and 2007 at our medical center. CVH was classified as poor, intermediate, or ideal according to 5 factors: smoking, body mass index, blood pressure, glucose/hemoglobin A1c, and cholesterol. We considered anthracyclines and hormone therapies as cardiotoxic cancer treatments in these analyses. Results: Women with ideal CVH scores had a lower risk of CHD (23.7%) and death (14.3%) compared to those with poor CVH scores (61.9% for CHD, 32.5% for death). Women exposed to cancer treatments had a higher risk of CHD (58.9%) and death (37.5%) compared to women who were not exposed (29.1% for CHD, 15.4% for death). Meanwhile, the joint effects of poor CVH and receipt of cardiotoxic cancer treatments conferred a much higher risk of CHD (75.9%) and death (39.5%) compared to women with ideal CVH who did not receive treatments (20.8% for CHD, 11.2% for death). All approaches ( Figure 1 ) demonstrated the highest area under the curve (AUC) for models considering the joint effects of CVH and cancer treatments (range: 0.70-0.75), followed by the independent effects of CVH (range: 0.65-0.66) and the independent effects of cancer treatments (range: 0.59-0.63). Conclusions: Cancer treatment data should be considered along with traditional cardiovascular risk factor data in the prediction of CHD and mortality among breast cancer survivors.