25 Years of Cardiac Ion Channel QSAR: From hERG Dominance to Multi-Channel Modelling

In 2013 the Comprehensive in vitro Proarrhythmia Assay (CiPA) initiative proposed multi-channel assessment for proarrhythmic risk evaluation, yet ligand-based predictive modelling has remained largely single-channel. To quantify and explain this mismatch, we survey 146 QSAR and machine learning studies on cardiac ion channels published between 2001 and early 2026. Across this period, 90% of studies target hERG as their sole ion channel endpoint—a proportion reflecting regulatory incentives, data availability, and research inertia rather than scientific necessity. Although public data for Nav1.5 and Cav1.2 were sufficient for modelling years earlier, both channels received only isolated studies until 2022. Since then, multi-channel modelling has expanded substantially, with multiple groups now addressing hERG, Nav1.5, and Cav1.2 jointly. In contrast, two CiPA channels, Kv4.3 and Kir2.1, remain entirely unmodelled due to genuine data scarcity. Closing the remaining gaps will require coordinated data release (or federated learning) for neglected channels, methods suited to data-scarce targets, and standardised benchmarks enabling rigorous progress tracking—none of which individual research groups can establish alone.