Abstract P5-14-14: A validated clinical risk prediction model for cardiac toxicity in patients treated with antiHER2 agents

Abstract

Abstract Background: Cardiac toxicity (CT) in the form of left ventricular systolic dysfunction is a well-known complication of anti-HER2 therapy. Prior clinical trial data has suggested potential factors associated with increased risk for CT. We sought to develop a clinical risk prediction model from real world patients to better identify patients who may be at higher risk for CT. Methods: Patients treated with anti-HER2 agents at all Mayo Clinic sites from Jan 2008 to July 2015 were reviewed. Patient data including age, race, body mass index, comorbidities, smoking history, baseline ejection fraction (EF), presence of left ventricular hypertrophy (LVH), prior cancer treatments, number of anti-HER2 agents used, and cardiac medication use were recorded. Factors were compared in patients with and without development of CT. Cox regression models were used to identify univariate and multivariate predictors of CT. A risk score was then assigned based on the final multivariable model. Results: 604 patients were included in final analysis. Median follow-up 2.9 yrs; median age: 57 yr; baseline EF: 64%. One yr CT rate was 7%. Independent predictors of CT were age, baseline EF, LVH, and anthracycline use (C index=0.74). Based on model coefficient, scores of 1, 1, 2 and 3 were assigned to age>55, EF <60%, anthracycline exposure and LVH, respectively. Two random samples of 300 patients were created out of the original cohort for internal validation of the risk score model. In each, the risk score was found to be a highly significant predictor for CT with C index of 0.72 {HR 2.13 (95% CI 1.57-2.90), p <0.001} and 0.77 {HR 2.45 (95% CI 1.74-3.44), p <0.001}. Conclusions: Patients with risk factors of older age, lower baseline EF, anthracycline use and LVH are at increased risk for CT when treated with anti-HER2 agents. Using a risk prediction model at therapy initiation may enhance identification of patients who benefit from cardiology referral, closer cardiac monitoring, and/or prophylactic treatment while on anti-HER2 targeted therapy. Table 1. Multivariable Cox regression model predicting cardiac dysfunctionParameterParameter EstimateHR (95%CI)p valueScoreAge >550.909552.48(1.27, 4.84)0.00751EF<600.634361.89(1.01, 3.53)0.04751Anthracycline Exposure1.046132.85(1.48, 5.46)0.00162Baseline LVH1.633565.12(2.04,12.85)<0.0013 Citation Format: Pooja P Advani, Lauren Cornell, Zhuo Li, Jordan Ray, Catalina Sanchez, Armando Villanueva, Taryn Smith, Alvaro Moreno-Aspitia. A validated clinical risk prediction model for cardiac toxicity in patients treated with antiHER2 agents [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-14-14.