Risk Prediction Scores for Cardiotoxicity in Cancer Patients Treated With Cardiotoxic Anticancer Therapies in Europe

Cancer management has undergone major advances over the past 30 years. Several anticancer therapies are now highly effective, allowing many cancers to become chronic diseases through sequential lines of treatment. However, these therapies may be associated with severe cardiovascular adverse events. It is therefore essential to better understand the factors that determine, for a given patient and a given anticancer therapy, the occurrence of cardiovascular toxicity, and ideally to predict which patients are at highest risk. The 2022 ESC Cardio-Oncology Guidelines strongly recommend risk stratification prior to the initiation of cardiotoxic anticancer therapies. Current recommendations suggest the use of risk scores known as "HFA-ICOS" scores. However: Not all cardiotoxic anticancer therapies currently have an associated risk score; Existing scores are derived from small retrospective or prospective cohorts (typically fewer than 100 patients), and most have not undergone rigorous validation, leaving considerable room for improvement. The EU-CARTOX-SCORES protocol aims to develop new prediction scores for cardiotoxicity occurring within the first year after initiation of cardiotoxic anticancer therapies. This protocol is innovative through the integration of artificial intelligence and/or machine learning approaches alongside conventional statistical methods. These models will be interfaced with a dedicated cardio-oncology digital platform (CardioOncoPilot), currently being deployed across the European Union, ensuring standardized and high-quality data collection within routine clinical care. This is a prospective, multicenter, observational study designed as a European cardio-oncology registry, involving all European cardiologists using the platform in routine practice. The inclusion period will last 3 years, with a 12-month follow-up for each patient. At the time of data extraction, all available cases recorded in the CardioOncoPilot platform will be analyzed.