Abstract / Summary
Abstract Rapid advances in oncology care have resulted in a growing population of cancer survivors, many of whom suffer from short- and long-term cancer therapy related cardiovascular toxicities (CTR-CVT). Cardio-oncology aims to diagnose and manage these cardiovascular conditions, but part of the challenge pertains to finding such individuals among the millions of cancer survivors. Medical informatics may help identify patients eligible to receive cardio-oncology care, assess the risk of cardiovascular complications, and predict toxicity by employing artificial intelligence (AI) algorithms. Uncoordinated development of the medical informatics solutions for cardio-oncology has produced disparate algorithms aiming to predict CTR-CVT by analysing features of patient records in the context of clinical guidelines developed by organizations such as the American Heart Association, the American College of Cardiology, and the European Society of Cardiology. These algorithm development efforts emphasize narrowly-defined clinical goals, rather than integrating an entire array of the applicable data elements to provide medical professionals with comprehensive analyses resembling care delivery in community settings and translatable into clinical workflows. The resulting mismatch between solutions developed by informaticists and the needs of practicing clinicians has led to difficulty with external validation and practical application of these informatics solutions. Additionally, the latest AI and medical device developments, such as digital twins and wearables, enable new solutions to support clinical decision making and patient education in cancer survivorship care. We aim to describe, summarize, and evaluate these developments in a cohesive vision and call to action for medical informaticists to develop relevant translatable solutions for cardio-oncology.