Real-time B-mode carotid imaging: A three-year multicenter experience

Abstract

This report attempts to establish the place of real-time B-mode carotid imaging in the diagnosis of carotid artery disease through an analysis of the results of 3 years of experience in three major noninvasive vascular laboratories. Over 7000 patients were studied noninvasively with real-time B-mode carotid imaging and oculoplethysmography (OPG). Angiographic comparisons of 1723 vessels form the basis of this study. All tests were interpreted by physicians who had no knowledge of the angiograms. Images and x-ray films were classified by diameter of stenosis: grade I (0% to 39%), grade II (40% to 69%), grade III (70% to 99%), and grade IV (total occlusion). The effect of the imaging experience was determined by dividing the study into two periods. The results of image quality vs. accuracy and of combining the anatomic study (scan) and physiologic study (OPG) were also evaluated. The overall data show a specificity of 87% (985 of 1139), a sensitivity for grade II disease of 72% (193 of 267), a sensitivity for grade III of 66% (133 of 201), and a sensitivity for grade IV of 64% (74 of 116). With experience each center showed improvement of the imaging technique in diagnosing grade III (p > 0.1: not significant) and grade IV disease (p < 0.0002: highly significant), although there was no improvement in the specificity and sensitivity of grade II disease. All scan errors were analyzed. Most errors were interpretation errors (27%: 90 of 338), scan/arteriogram mismatches (23%: 79 of 338), or poor-quality scans secondary to existing disease (22%: 75 of 338). There was a direct correlation of scan quality and accuracy, with a 97% specificity for grade I scans of good to excellent quality. When the scan and OPG agreed, there were uniform predictive values for all grades (88% to 93%). Sixteen of 79 scan/arteriogram mismatch vessels were operated on, and the scan proved more reliable in 86%. Real-time B-mode carotid imaging is a reliable technique for defining the normal carotid artery and is becoming increasingly sensitive in identifying existing disease. Despite its limitations, its strong points make it a valuable clinical tool.