AI-enhanced cardiac digital twins extend drug proarrhythmic risk assessment through experimental data uncertainty propagation and overdose exploration: A loperamide case study

Drug-induced QT interval prolongation is a key biomarker of proarrhythmic risk and central to drug cardiac safety evaluation alongside in vitro assays and animal studies, yet current preclinical frameworks provide limited insight into how experimental uncertainty and extreme exposures translate into real-world arrhythmic risk despite both factors critically modulating outcomes. To address this, we used sex-specific machine learning surrogate models trained on 3D cardiac digital twins-mechanistic electrophysiology models of anatomically detailed ventricles that integrate multichannel ion channel block data-combining the realism of 3D simulations with high-throughput capability to enable rapid, ethically unconstrained assessment of proarrhythmic risk. We illustrate the approach using loperamide, which is safe at therapeutic doses but linked to fatal arrhythmias at extreme exposures, through two analyses: propagating experimental IC50 and Hill coefficient variability to quantify its effect on predicted QT prolongation and arrhythmic probability, and simulating extreme exposures to identify sex-specific arrhythmogenic thresholds. Experimental variability substantially broadened uncertainty in predicted QT prolongation across exposure levels, while extreme exposure simulations identified arrhythmogenic thresholds of approximately 107-109×C