Magnetic resonance angiography minimizes need for arteriography after inadequate carotid duplex ultrasound scanning

PURPOSE: We prospectively evaluated whether magnetic resonance angiography (MRA) enabled definition of cerebrovascular anatomy after indeterminate or inadequate results at duplex ultrasound scanning to facilitate patient selection for carotid endarterectomy (CEA) and for technical planning. METHODS: After implementation of a protocol in October 1998 to minimize use of cerebral arteriography, MRA (arch/cervical two-dimensional and cranial three-dimensional time of flight technique) was performed in 138 consecutive patients with cerebrovascular occlusive disease and inconclusive duplex scans obtained by an ICAVL-approved laboratory. The ability of MRA to define anatomic features unresolved at duplex scanning was compared between categories of duplex scan inadequacies. Operative outcome was compared between patients requiring MRA before CEA (n = 66) and a concurrent cohort undergoing CEA on the basis of duplex scan results only (n = 69). RESULTS: Incomplete imaging of the carotid bifurcation, because of high bifurcation, long (>3 cm) internal carotid artery (ICA) plaque, or calcific shadows, was the most common reason for inadequate duplex scans (n = 74, 53%), followed by borderline severe ICA disease (23.17%), suspected extracervical disease (supra-aortic trunk, vertebral, or intracranial, 22, 16%), ICA near- occlusion (12.9%), and diffuse recurrent stenosis (7.5%). MRA enabled resolution of duplex scan inadequacies in 95% of patients with disease confined to the carotid bifurcation, and 90% of all patients, but was least accurate for delineation of extracervical lesions (77%) and near-occlusions (75%). In 5 of 8 patients (6%) arteriography was performed to determine operability of ICA near-occlusion or extracervical lesions. Combined stroke and death rates after CEA were not statistically different (P =.3) between patients requiring MRA (3 of 66, 4.6%) and the concurrent group in whom MRA was performed solely on the basis of duplex results (1 of 69, 1.5%). However, intraoperative technical adjustments (anatomy that precluded shunt use, extended endarterectomy length, ICA shortening due to tortuosity) were planned in 71% of patients (12 of 17) with MRA-defined anatomy, but only 36% of patients (4 of 11) with long CEA on the basis of duplex results only (P =.08). CONCLUSION: MRA replaces the need for cerebral arteriography in most patients after inadequate carotid duplex scanning. Delineation of cerebrovascular anatomy at MRA assists in determination of CEA candidacy and operative planning.