Validation of the HFA-ICOS risk assessment tool with real-world data from a prospective cohort of breast cancer patients in treatment with anthracyclines and trastuzumab

Abstract Background The increasing prevalence of cancer therapy-related cardiovascular toxicity (CTRCD) poses a significant challenge to patient management, necessitating accurate risk assessments. Previous methodologies have struggled with precise risk stratification, partly due to varying definitions of CTRCD and the specificity of cancer treatments. Purpose This study aimed to evaluate the assessment tool proposed by the Heart Failure Association and the International Cardio-Oncology Society (HFA-ICOS) in predicting the risk of CTRCD in patients undergoing anthracycline and trastuzumab treatment for breast cancer at a national reference center. Methods A prospective cohort study was conducted, encompassing 172 patients treated from January 2022 to July 2023. Patients were assessed at baseline, three months, and six months post-treatment initiation, with exclusion criteria including baseline left ventricular ejection fraction (LVEF) <50%, history of heart failure, and stage IV breast cancer. The study utilized the HFA-ICOS tool for risk stratification and CTRCD was defined with the current criteria of the European Society of Cardiology guideline. Descriptive statistics were calculated, and incidence rates and Cox proportional-hazards models were developed according to the risk category. Results In our study, 138 (80.2%) patients were classified as low risk, 23 (13.4%) as moderate risk and 11 (6.4%) as high risk. We identified 81 (47.1%) patients that developed CTRCD in follow-up. Descriptive characteristics can be found in Table 1. The global incidence rate for CTRCD in our cohort was of 2.4 new cases per 1000 person-days, however no differences in incidence rates were found when stratifying by risk category (Figure 1A). Moreover, patients with moderate to high risk according to the HFA-ICOS tool did not show a significantly increased hazard ratio for cardiotoxicity (HR 0.41, 95%CI 0.15-1.12, p=0.082 and HR 1.39, 95%CI 0.50-3.86, p=0.5, respectively) as shown in Figure 1B. Conclusion While the HFA-ICOS tool provides a structured framework for assessing cardiotoxicity risk, its predictive value in this cohort was limited. The findings underscore the complexity of cancer therapy-related cardiotoxicity and the need for enhanced, individualized risk assessment tools. Future research should focus on refining risk prediction models, incorporating broader clinical and genetic factors to improve patient outcomes in cardio-oncology.Table 1.Clinical CharacteristicsFigure 1.Incidence rates and Cox models