The use of planar radionuclide ventriculography to evaluate global and segmental ventricular function is limited by the superimposition of structures in some projections and the gross segmental resolution of the planar technique. Preliminary reports have suggested the feasibility of tomographic gated radionuclide ventriculography with rotating detector systems. This study tested the hypotheses that 1) tomographic radionuclide ventriculography detects segmental dysfunction at rest not identified with multiview planar studies and single plane contrast ventriculography, and 2) ventricular volumes and ejection fraction calculated from these studies provide data similar to those obtained with angiography and planar radionuclide ventriculography. Gated blood pool tomograms were acquired over 180 degrees at 15 frames per cardiac cycle during the initial 90% of the cardiac cycle. Compared with the multiview planar technique tomographic ventriculography showed an increased sensitivity for detecting left ventricular segments with significant coronary artery stenosis (97 versus 74%, p less than 0.025) without any loss in specificity. Compared with both planar radionuclide and contrast ventriculography, tomographic radionuclide ventriculography also detected more noninfarcted left ventricular segments supplied by stenosed coronary arteries (81 versus 39 and 32%, respectively, p less than 0.01). Tomographic radionuclide ventriculographic measurements of left ventricular volumes and ejection fraction showed close correlations with angiographic and planar radionuclide determinations. Gated blood pool tomography is a sensitive method for the evaluation of segmental wall motion and an accurate method for the measurement of global left ventricular volumes and ejection fraction.