This retrospective study compared intravenous digital subtraction angiography (IV-DSA) with conventional
catheter angiography (CA) performed on 105 patients with suspected
cerebrovascular disease. Comparisons were made considering CA as the "gold standard." Factors reviewed for correlation included
carotid artery stenosis, ulceration, fibromuscular disease, vessel occlusion, and subclavian and intracranial lesions. Each IV-DSA was judged as to its quality. Eighty-seven patients (83%) had interpretable IV-DSA studies. IV-DSA differed from CA regarding
stenosis on 46 carotid arteries (26%). Of 47 excellent quality IV-DSA studies, 18 (38%) were in disagreement with CA regarding
stenosis. For the subcategory of patients having hemodynamically significant
stenosis (greater than 50%
luminal narrowing) according to CA, the sensitivity of IV-DSA was 85%. Specificity was 94% with an overall accuracy rate of 92%. The authors believe these high values are somewhat misleading. The IV-DSA interpretation of
stenosis differed from CA in 32 of 48 patients (67%) in this group with a mean difference of 30%. IV-DSA identified 3 of 11 intracranial lesions (27%). Surgical
therapy was modified in 13 of 87 patients (15%) on the basis of angiographic interpretation of
carotid artery stenosis compared with IV-DSA. We conclude that IV-DSA should be used as a screening tool and the decision for carotid artery surgery made solely on its findings should be reached with the knowledge of its limitations and that any patient with hemispheric cerebral symptoms warrants CA as the first diagnostic examination.