Abstract 10797: Applying Cardiotoxicity Risk Prediction Scores in Women with Early-Stage Breast

Background: Cardiotoxicity is a common complication of anthracyclines and HER2-targeted therapies in breast cancer patients. Despite recognized risk factors, accurate prediction of cardiotoxicity risk remains a challenge. We applied 3 published risk prediction scores (Table) to a cohort of women with HER2+ breast cancer treated with anthracycline and trastuzumab and described their predictive accuracy. Methods: Women with stage I-III HER2+ breast cancer treated at the Princess Margaret Cancer Centre (Toronto) between 2006-2019 were identified retrospectively. Those with a pre-therapy MUGA scan or 2D echocardiography and ≥ 3 follow up scans using the same modality were included. CREC criteria were applied to define cardiotoxicity. Three risk prediction scores (Ezaz et al., Romond et al., and HFA-ICOS proformas) were used to classify patients into low/medium/high cardiotoxicity risk according to pretreatment characteristics. The proportion of patients who developed cardiotoxicity across risk categories were compared within each score using chi-square test. Results: Of 629 women (mean age 52.4 ± 10.9 years), 151 (24%) met CREC cardiotoxicity criteria. With the Ezaz et al. score , proportion of patients that developed cardiotoxicity was 24.4% (n = 141/577), 20% (n = 9/45) and 14.3% (n =1/7), in the low, medium, and high predicted risk groups, respectively (p = 0.665). With Romond et al. score , this was 17.9% (n = 40/223), 24.5% (n = 50/204) and 30.5% (n = 61/200), respectively (p = 0.01). HFA-ICOS was 15.5% (n = 30/193), 26.9% (n = 109/404), and 37.5% (n = 12/32), respectively (p = 0.002). Conclusion: In women with early stage HER2+ breast cancer receiving trastuzumab therapy, risk scores by Romond et al., and the HFA-ICOS proformas predict cardiotoxicity risk. However, for both these scores the incidence of cardiotoxicity in the “low” predicted risk group remains high (15.5%-17.9%). Therefore, there remains a need to develop more accurate risk prediction scores.