Myocardial infarct size was estimated by three methods in a canine model, using Tc-99m pyrophosphate at 24 and 48 hr after coronary ligation. A gamma camera provided anterior, LAO, and lateral views, and was then fitted with multipinhole (MPH) and rotating-slanthole (RSH) collimators for tomographic studies, processed by computer to display frontal sections of the chest. Infarct weight was measured postmortem for comparison. All transmural infarcts were detected by all three imaging techniques. RSH tomography was superior to both MPH tomography and planar imaging for the detection of nontransmural infarction. Infarcts as small as 1.0 g were detected. Estimates of infarct volume measured from RSH slices showed an excellent correlation with infarct weight (r = 0.89) and were reproducible within acceptable limits. Estimates on infarct volume measured from MPH slices demonstrated a significantly poorer correlation with infarct weight (r = 0.48, p less than 0.01). Both tomographic techniques may improve infarct visualization by suppressing overlying activity and increasing contrast between infarct and background, but both produce significant blur artifacts that hamper their utilization by inexperienced observers.