From data to decision: a clinical pipeline for wearable AI in early cardiotoxicity detection

Chemotherapy-induced cardiotoxicity (CIC) is a leading cause of morbidity in cancer survivors, as conventional surveillance often detects cardiac dysfunction only after significant injury. This review moves beyond summarizing emerging technologies to focus on the end-to-end clinical pipeline-from sensor data to actionable decision-making-for creating a proactive "early-warning" system. We examine how continuous monitoring with wearable devices and artificial intelligence (AI) can detect subclinical CIC by analyzing digital biomarkers like heart rate variability. Proof-of-concept studies show AI can predict LVEF decline with moderate to high accuracy, though evidence is limited by small cohorts and lacks external validation. Implementing this wearable-AI pipeline could fill a critical surveillance gap and enable timely cardioprotective interventions. However, large-scale validation and building clinician trust through interpretable models are crucial before routine clinical adoption.