SUMMARY To evaluate the accuracy of coronary angiography in identifying severe narrowing of the left main coronary artery (LMCA), the degree of narrowing observed by angiography was compared to that observed at necropsy in 28 patients with symptomatic coronary heart disease in whom angiography had been performed within 40 days of death. The angiograms were evaluated independently by three experienced angiographers. In 20 of the 28 patients (71%), the degree of narrowing of the LMCA was either underestimated (13 patients) or overestimated (10 patients) by two or three of three angiographers; of 84 angiographic judgments made by the three angiographers in the 28 patients, 54 (64%) were underestimates (33 judgments, 39%) or overestimates (21 judgments, 25%) of the degree of LMCA narrowing. Of 12 LMCAs narrowed 76-100% in cross-sectional area at necropsy, six were underestimated at preoperative angiography by two or three of three angiographers; of 12 LMCAs narrowed 51-75% in cross-sectional area at necropsy, all 12 were either under-or overestimated angiographically by two or three of three angiographers; of four LMCAs narrowed 26-50% in cross-section at necropsy, two were overestimated by two of three angiographers. Thus, angiographic determination of degrees of narrowing of the LMCA during life is subject to considerable error. The angiographic errors appear to have resulted primarily from the presence of atherosclerotic plaque in the LMCA and an insufficient number of angiographic projections.CORONARY BYPASS SURGERY appears to increase longevity in patients with significant stenosis of the left main coronary artery (LMCA), so identification of "significant" narrowing in this artery generally is considered to be an indication for the bypass operation regardless of the symptomatic status of the patient.1 Accurate identification of the degree of LMCA narrowing, therefore, is of paramount importance. Although coronary angiography is the most reliable means of identifying patients with significant stenosis of the LMCA, the accuracy of this technique has not been subjected to critical analysis. In the present report we examine the accuracy of coronary angiography in evaluating the LMCA by comparing the results of antemortem angiography to histologic observations at necropsy. At necropsy, the major extramural coronary arteries, including the LMCA, were excised intact from the heart, fixed in 10% buffered formalin, x-rayed, decalcified (if necessary), and cut transversely into 5-mm-long segments. (We did not use pressure perfusion of the coronary arteries before fixation.) The tissue was then dehydrated (alcohols), cleared (xylene), embedded in paraffin and cut. One section was stained by Movat's pentachrome method from each 5-mm segment. In addition, the original paraffin-embedded blocks of LMCA from six patients were recut at 6-,u intervals. Of the 1 160 sections cut, every twentieth section was stained by Movat's method and examined. Histologic evaluation of the degrees of cross-sectional area narrowing by atheroscle...