Abstract Background Effective cardio-oncology management depends on precise baseline cardiovascular risk evaluation. The available risk assessment score developed by the Heart Failure Association (HFA) and the International Cardio-Oncology Society (ICOS) needs to be externally validated in different populations and with an assessment event-free survival. Purpose To evaluate the effectiveness and reliability of the HFA-ICOS risk score in predicting cardiotoxicity in breast cancer patients over a three-year follow-up. Methods All patients from the PREDICATE study (ISRCTN12628444) were included in this analysis. In this prospective single-center study we included 114 breast cancer patients: 101 received anthracycline therapy, and 13 combined anthracycline and trastuzumab treatment. In all patients baseline was calculated cardiovascular toxicity risk by the HFA-ICOS score. The combined endpoint was a new decrease in global longitudinal strain (GLS) of 15%, a new reduction in left ventricular ejection fraction (LVEF) of 10%, and discontinuation of chemotherapy due to cardiovascular complications and cardiovascular death. Statistics included analysis of variance or Kruskal–Wallis rank sum test for continuous variables and Chi-Square or Fisher exact test for categorical data, Kaplan-Meier survival curves, multivariative logistic regression analysis and ROC analysis. Results Baseline 55.3% (n=63) were classified as low risk for cardiotoxicity, 36.8% (n=42) as moderate, and 7.9% (n=9) as high. Demographic characteristics, comorbidities, and cardiovascular risk factors are presented at Figure 1. Median follow-up was 991 (IQR 920; 1062) days. The event-free survival rate for the low-risk group was 75%: 1045.2 days (95% CI: 963.2; 1127.1), with 15 events. The survival rate for the moderate-risk group was 62%: 993.2 (95% CI: 882.3; 1104.1), with 14 events. For the high-risk group, the survival rate was 38%: 689.0 (95% CI: 406.1; 971.9), with five events (log-rank 0.038). The HFA-ICOS risk stratification demonstrated discrimination with an AUC of 0.64 (95% CI: 0.51, 0.76), with a sensitivity of 26.5% and a specificity of 95%; the positive predictive value was 84.1%, and the negative predictive value was 56.4%, yielding an overall accuracy of 74.6% (Figure 2). Conclusions Our prospective single-center study suggests that the HFA-ICOS risk score has an acceptable level of predictive validity for assessing the risk of a combined endpoint, comprising both echocardiographic markers of LV dysfunction and CV mortality in breast cancer patients, particularly those treated with anthracyclines.Figure 1 Figure 2
Amanova et al. (Fri,) studied this question.