Cervical carotid MR angiography with multiple overlapping thin-slab acquisition: comparison with conventional angiography.

Abstract

Multiple overlapping thin-slab acquisition (MOTSA) is a hybrid time-of-flight MR angiography technique that combines the advantages of two- and three-dimensional methods. A venetian blind-like artifact, resulting from variations in signal intensity at slab boundaries, is a potentially significant disadvantage. The objective of this study was to determine the accuracy of MOTSA MR angiography for the measurement of stenosis and to evaluate the effectiveness of modifications to reduce the venetian blind artifact. Fifty-one consecutive patients undergoing conventional and MR angiography of the carotid arteries for suspected atherosclerotic disease were studied. Methods established in the North American Symptomatic Carotid Endarterectomy Trial (NASCET) were used to measure stenosis. Measurements from the conventional angiograms were compared with those from the MR angiograms. Modifications of the MR angiographic technique that were instituted after the initial 26 patients to reduce the venetian blind artifact included new acquisition parameters and the use of a postprocessing algorithm. After modification, highly significant improvements in Spearman ranked correlation coefficients were observed for MR angiography interobserver comparison (.573 to .974, p < .001) and comparison of conventional and MR angiography (.774 to .949, p < .001). A corresponding reduction was observed in the standard deviation of the difference measurement for percent stenosis for both interobserver (34.5% to 11.7%, p < .001) and intertechnique (22.6% to 12.6%, p < .001) comparisons. In the postmodification group, the interobserver correlation coefficients for conventional and MR angiography were not significantly different (.969 vs .974). The areas under the receiver-operating-characteristic curves used to predict surgically significant disease for the premodification and postmodification groups were 0.79 and 0.98, respectively, for observer A and 0.85 and 0.95, respectively, for observer B. Improved acquisition parameters and postprocessing significantly reduce the severity of the venetian blind artifact and increase the accuracy of MOTSA MR angiography in the measurement of percent stenosis. With these improvements, MOTSA MR angiography shows promise for noninvasive evaluation of stenosis of the carotid arteries.