BACKGROUND: Tissue Doppler may be used to quantify regional left ventricular function but is limited by segmental variation of longitudinal velocity from base to apex and free to septal walls. We sought to overcome this by developing a composite of longitudinal and radial velocities. METHODS AND RESULTS: We examined 82 unselected patients undergoing a standard dobutamine echocardiogram. Longitudinal velocity was obtained in the basal and mid segments of each wall using tissue Doppler in the apical views. Radial velocities were derived in the same segments using an automated border detection system and centerline method with regional chords grouped according to segment location and temporally averaged. In 25 patients at low probability of coronary disease, the pattern of regional variation in longitudinal velocity (higher in the septum) was the opposite of radial velocity (higher in the free wall) and the combination was homogeneous. In 57 patients undergoing angiography, velocity in abnormal segments was less than normal segments using longitudinal (6.0 +/- 3.6 vs 9.0 +/- 2.2 cm/s, P =.01) and radial velocity (6.0 +/- 4.0 vs 8.0 +/- 3.9 cm/s, P =.02). However, the composite velocity permitted better separation of abnormal and normal segments (13.3 +/- 5.6 vs 17.5 +/- 4.2 cm/s, P =.001). There was no significant difference between the accuracy of this quantitative approach and expert visual wall motion analysis (81% vs 84%, P =.56). CONCLUSION: Regional variation of uni-dimensional myocardial velocities necessitates site-specific normal ranges, probably because of different fiber directions. Combined analysis of longitudinal and radial velocities allows the derivation of a composite velocity, which is homogeneous in all segments and may allow better separation of normal and abnormal myocardium.