The morphologic description and measurements of endarterectomy specimens are usually believed to be accurate and are used as the gold standard against which the findings of diagnostic procedures are judged. Pathology data on 289 endarterectomy specimens from five participating centers and the corresponding angiography and B-mode ultrasonography data provided a basis for scrutinizing the validity of using the morphologic measurements as a standard. Discrepancies of > 1 mm between pathology and angiography measurements of minimum residual lumen occurred in 35% of the cases and between pathology and B-mode ultrasonography measurements in 64% of the cases. Discrepancies of > 1 mm between pathology-and angiography-measured lesion width occurred in 81% of the cases and between pathology and B-mode ultrasonography measurements in 64% of the cases. The cases representing mismatches of > 1 mm at one participating center were subjected to a rigorous review, with remeasurement of all morphologic features, in an attempt to explain the discrepancies. Various types of artifactual distortion of the specimens, the presence of slit-like and occluded lumens that were likely related to loss of perfusion pressure, and an inability to match planes of interrogation used in angiography and B-mode ultrasonography with pathology planes contributed significantly to the existence of mismatches. On the other hand, fixation and decalcifkation produced minimal and insignificant distortional changes. We conclude that the acquisition of quantitative data from endarterectomy specimens and the acceptance of morphologic data as a standard are limited by a number of problems that can be defined but have been difficult to resolve. (Stroke 1988; 19:289-296) T he morphologic description and measurements of endarterectomy specimens are usually believed to be accurate and are used as the gold standard against which the findings of diagnostic procedures are compared to assess their validity. Such was the approach used in a multicenter National Institutes of Health-spon- Supported by National Heart, Lung, and Blood Institute contracts N01- HV-129I6, 12904, 12912, 12913, 12914, 12915, and 12917. Address for correspondence: Eric A. Schenk, MD, Department of Pathology, University of Rochester Medical Center, Rochester, NY 14642.Received March 13, 1987; accepted October 6, 1987. sored study to evaluate B-mode ultrasonographic imaging (B-scan) of the carotid arteries. Five clinical centers and one data coordinating center collaborated in this study.The main purpose of the study was to assess the accuracy with which B-scan describes lumen diameter, plaque thickness, and percent stenosis within 1.5 cm of the carotid artery bifurcation. The B-scan findings were compared with those obtained by both conventional and digital angiography and by pathologic examination of the endarterectomy specimen. Details about the overall design, the methods for ultrasonography and angiography, and the findings have been published.'-2 This communication focuses on the patholo...