BackgroundAreas of abnormal or heterogeneous conduction velocity (CV) are important ablation targets for ventricular tachycardias, yet precise assessment of CV in clinical contact mapping remains challenging. Numerous different CV estimation methods have been proposed. ObjectiveThis study aimed to compare the automated local activation time (LAT)–independent omnipolar-based CV estimation method termed wave speed (WS) with 4 established LAT-based methods to formally establish the quantitative differences between them. MethodsHigh-density contact maps in patients with structurally normal hearts during sinus rhythm (SR) and ventricular ectopy (VE) were retrospectively analyzed. CV was assessed and compared by 5 methods: omnipolar WS, gradient method, planar wavefront fitting, circular wavefront fitting, and radial basis function. CV variations based on electrogram (EGM) type (unipolar, bipolar, and omnipolar), catheter movement, and surrogate markers for catheter contact were analyzed. ResultsThe study included 23 patients (47.8% male; 45.7 ± 17.3 years) with 22 SR maps (11 left ventricle, 11 right ventricle) and 16 VE maps (9 left ventricle, 7 right ventricle). The WS algorithm yielded statistically significant higher CV estimates in SR (mean, 1.41 ± 0.18 m/s) and VE (mean, 1.23 ± 0.18 m/s) maps compared with all LAT-based estimation methods, with absolute differences ranging from 0.1 m/s to 0.81 m/s. Median pointwise differences in SR and VE between WS and LAT-based methods were high, ranging from 0.55 ± 0.15 m/s (WS vs planar wavefront fitting) to 0.67 ± 0.16 m/s (WS vs radial basis function). For LAT-based methods, use of unipolar EGMs yielded significantly higher CV estimates than bipolar or omnipolar EGMs in SR. ConclusionThe CV estimation method has an important, statistically significant impact on ventricular CV measurements. Future work will focus on how these differences affect identification of pathologic conduction slowing in scar-related substrate.