Abstract
OBJECTIVE: Trastuzumab-based therapy is a cornerstone for HER2-positive breast cancer but carries a risk of significant cardiotoxicity. This study aims to investigate the long-term incidence of cardiovascular adverse events (CVDs), identify a comprehensive set of risk factors, and develop a robust predictive model for trastuzumab-induced cardiotoxicity (TIC) in a well-characterized, single-center patient cohort. METHODS: We retrospectively analyzed 600 HER2-positive breast cancer patients on trastuzumab-based regimens from 2018-2023. Patients were divided into CVD (n=100) and non-CVD (n=500) groups based on cardiotoxicity occurrence during a median 3.6-year follow-up. We analyzed baseline characteristics, treatment protocols, and serial monitoring data including ECG, NT-proBNP, LVEF, left ventricular global longitudinal strain (LVGLS), and cardiac biomarkers (creatine kinase (CK), CK-MB, and hs-cTnI). RESULTS: The cumulative incidence of CVDs was 16.7%. Cardiotoxicity events included symptomatic heart failure (n=11), asymptomatic LVEF decline (n=51), significant LVGLS reduction (n=29), and significant arrhythmias (n=9). Significant baseline predictors of cardiotoxicity included age >60 years, pre-existing hyperlipidemia, and elevated NT-proBNP levels (p<0.05). Treatment with anthracycline-based chemotherapy and chest radiotherapy were also strongly associated with increased CVD risk. During follow-up, the CVD group exhibited a significantly greater decline in LVEF (baseline vs. follow-up: 64.1% vs. 48.8%) and LVGLS (-20.9% vs. -15.3%) compared to the non-CVD group (p<0.001). In multivariate logistic regression analysis, the strongest independent predictors for CVDs were a post-treatment LVEF decline >10% (OR 5.75, 95% CI 3.95-8.41), a post-treatment relative LVGLS decline >15% from baseline (OR 4.42, 95% CI 3.10-6.22), and elevated hs-cTnI (OR 4.10, 95% CI 2.91-5.74). A predictive model incorporating both baseline and on-treatment factors showed excellent discrimination (AUC = 0.88). CONCLUSION: Cardiotoxicity remains a major concern in long-term trastuzumab therapy. This single-center study highlights the critical importance of integrating baseline risk stratification with serial monitoring of advanced echocardiographic parameters like LVGLS and sensitive biomarkers like hs-cTnI. Our comprehensive predictive model offers a powerful tool for early identification of at-risk patients, guiding personalized surveillance and facilitating timely implementation of cardioprotective strategies to mitigate the risk of irreversible cardiac damage in this patient population.