Abstract: Torsade de Pointes (TdP) proarrhythmia is a major complication of therapeutic drugs that block the delayed rectifier current. QT interval prolongation, the principal marker used to screen drugs for proarrhythmia, is both insensitive and nonspecific. Consequently, better screening methods are needed. Drug-induced transmural dispersion of repolarization (TDR) is mechanistically linked to TdP. Therefore, we hypothesized that drug-induced enhancement of TDR is more predictive of proarrhythmia than QT interval. High-resolution transmural optical action potential mapping was performed in canine wedge preparations (n = 19) at baseline and after perfusion with 4 different QT prolonging drugs at clinically relevant concentrations. Two proarrhythmic drugs in patients (bepridil and E4031) were compared with 2 nonproarrhythmic drugs (risperidone and verapamil). Both groups prolonged the QT (all P < 0.02), least with the proarrhythmic drug bepridil, reaffirming that QT is a poor predictor of TdP. In contrast, TDR was enhanced only by proarrhythmic drugs (P < 0.03). Increased TDR was due to a preferential prolongation of midmyocardial cell, relative to epicardial cell, APD, whereas nonproarrhythmic drugs similarly prolonged both cell types. In contrast to QT prolongation, augmentation of TDR was induced by proarrhythmic but not nonproarrhythmic drugs, suggesting TDR is a superior preclinical marker of proarrhythmic risk during drug development.