Background: Studies have shown that the increase in heterogeneity during repolarization has higher correlation with torsades de pointes than QT prolongation itself. A challenge issue we have to address is what electrocardiographic (ECG) features are most likely linked to the heterogeneity changes. While we are identifying the best features, we still have to rely on QT interval, although accurate QT measurement is also difficult because of the complex nature of cardiac repolarization, ambiguous definition of the end of repolarization in surface ECG, and lack of a gold standard for accuracy validation. Methods: In this study, we applied a modeling approach to address both tasks: identify new morphology features and validate QT measurement accuracy. Theoretically, QT interval from torso ECGs should closely correlate with the maximum action potential (AP) duration (APD). An ion-channel–based cardiac cell model and a cell-to-torso forward model were used to generate many pairs of cells and corresponding ECGs. We then compared the APD of the simulated cardiac cells with the QT interval calculated from GE's (Milwaukee, WI) 12SL QT algorithm. By changing the parameters of the slow-potassium and rapid-potassium ionic (Ikr) channels, a table of APs was generated, with APD range of 377 to 500 milliseconds. To calculate the ECG from the AP at cell level, a simplified representation of this relation is Y = A × X, where Y is the potential on the body surface, X is the cell AP, and A is the transfer matrix, which is determined by the geometry shapes (acquired from GE [Milwaukee, WI] cardiac volume computed tomographic scanner) and the conductivities of different tissues. Finite element and boundary element methods were applied to calculate the transfer matrix. A T-wave morphology score formed by 3-segment least-square–fitting lines are also examined with the changes in the Ikr. Results: For the test of QT interval vs APD, the correlation coefficient is 0.99, and the root-mean-square difference is 17 milliseconds. The QT interval measured in torso ECGs is about 10 milliseconds shorter than the maximum APD. The correlation of the T-wave morphology score with the change in Ikr is 0.88. Conclusion: The modeling approach provides a more objective way for repolarization analysis and for validation of T-wave measurement accuracy. The QT interval measurement algorithm currently used in GE's cardiographs is highly correlated with the model-generated APD.