Abstract P6-09-03: Development of a cardiac toxicity prediction tool for HER2 (+) breast cancer patients receiving trastuzumab

Background: In HER2 positive breast cancer, the addition of trastuzumab prolongs overall survival in both early stage and metastatic disease. However, moderate to severe cardiac toxicity is a potentially serious complication which can lead to dose reductions, delays, hospitalizations, and premature discontinuation of treatment. Patient care could be substantially improved if such cardiac events are accurately predicted through the use of validated and easy-to-use mathematical models. In this study, the development of a model to predict the risk of cardiac toxicity prior to initiation of trastuzumab therapy is described. Methods: Medical records of 498 HER2 positive breast cancer patients who received trastuzumab at the Ottawa Hospital Cancer Centre were identified. Charts were reviewed for potential cardiac toxicity risk factors and cardiac events. Potential cardiac toxicity variables included: cardiac risk factors, medications, previous chemotherapy/radiation, baseline left ventricular ejection fraction (LVEF), and exposure to anthracyclines. Cardiac toxicity was defined as: decreased LVEF greater or equal to 10% and to less than 50%, referral to the cardiac oncology service, or clinical symptoms of heart failure. General linear modeling for a discrete bivariate outcome was used to identify risk factors for cardiac toxicity using a backwards elimination process. Internal validation of the final regression coefficients was done using nonparametric bootstrapping. A risk scoring algorithm (range 0-100) was then derived from the final model coefficients. A receiver operating characteristic (ROC) curve analysis was then undertaken to measure the predictive accuracy of the final scoring algorithm. Results: Baseline LVEF, concomitant use of any cardiac medication or lipid lower drugs and doxorubicin based chemotherapy were identified as being important predictors for cardiac toxicity. The ROC curve analysis indicated good predictive accuracy with an area under the curve of 0.69 (95%CI: 0.64 to 0.74). Prior to the initiation of trastuzumab, patients with risk scores greater than 45 units would be considered at high risk for developing cardiac toxicity (likelihood ratio = 2.9). Conclusions: Risk of cardiac toxicity with trastuzumab is increased in patients with a low baseline LVEF, history of coronary artery disease and hyperlipidemia, as well as those receiving doxorubicin. The planned external validation and eventual clinical application of this prediction tool will be an important source of risk information for the practicing oncologist, and may enhance patient care by optimizing preventative therapies and selecting high risk patients for cardiac imaging and cardiology follow-up. Citation Format: Jeffrey A Sulpher, George Dranitsaris, Freya Crawley, Franco Dattilo, Maya Kovacs, Christopher Johnson, Susan Dent. Development of a cardiac toxicity prediction tool for HER2 (+) breast cancer patients receiving trastuzumab [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-09-03.