Clinicopathological correlation of premortem coronary angiograms and postmortem data was performed to determine the degree of interobserver variability in the clinical interpretation of coronary angiograms using the pathological findings as the standard of accuracy. Comparison of the independent interpretations of the antemortem coronary angiograms by three cardiologists showed no substantial difference in overall diagnostic accuracy among the three observers. In the majority of instances where a substantial angiographic lesion was found on coronary angiography, it was verified pathologically. In false-positive or false-negative interpretations, a majority opinion was accurate when compared with a pathological lesion in 50% of the instances. Despite the fact that coronary angiography is an excellent diagnostic tool, there remain limitations of the accuracy of interpretation not solved by a majority opinion.