Extracranial carotid MR imaging at 3T

Some controversy exists over the accuracy and optimal parameters for carotid CE MR angiography at 1.5T. Spatial resolution remains more important than does temporal resolution to address the key question of vessel stenosis, based upon a review of the available literature that compares CE MR angiography with DSA. Specifically, CE MR angiograms with 0.9- to 1.2-mm resolution in all three planes before interpolation have a high reported sensitivity and specificity compared with DSA. To achieve this type of spatial resolution, cover the entire course of the carotid arteries from the aortic arch through the skull base, and achieve an absence of venous signal usually requires an elliptic-centric phase encoding CE MR angiogram that lasts for 50 to 60 seconds without the use of parallel imaging techniques. This near-millimeter resolution requires an accurate timing of the gadolinium bolus arrival to maximize intra-arterial SNR and to minimize venous contamination. Parallel imaging techniques can decrease the imaging time, but at a cost of some SNR. Initial experience with eight-channel or higher neurovascular coils at 3T indicates an increase in SNR/CNR compared with 1.5T. This should allow more consistent submillimeter-resolution carotid CE MR angiography with adequate SNR to maintain good IQ in a wide variety of clinical patients. Although a definite, prospective comparison of various CE MR angiography techniques, including a 20- to 30-second scan with 1.2- to 1.4-mm³ voxel resolution and 50- to 60-second scan with 0.9- to 1.1-mm³ voxel resolution at 1.5T, as well as 0.5- to 0.6-mm³ voxel resolution with scan time of 50 to 60 seconds at 3T versus rotational DSA does not exist, the expectation is that the higher resolution and increased SNR that has resulted from 3T carotid CE MR angiography will have high sensitivity and specificity in detecting severe carotid stenosis. The most exciting application of 3T for carotid artery imaging may not be the higher resolution CE MR angiogram, however. Early work has demonstrated the potential of 3T, combined with sensitive surface coils, to depict carotid plaque with sufficient SNR to identify important plaque components consistently in most patients. This could help move MR imaging of the carotid arteries away from a strict evaluation of luminal narrowing to a focused evaluation of plaque morphology. Much work needs to be done. Although there is a growing body of literature to support the contention that plaque morphology is a predictor of subsequent thromboembolic disease, the natural history of these various plaque components in a large number of patients needs to be elucidated. If plaque characterization proves to be an independent risk factor that predicts stroke, more aggressive clinical treatment option strategies may be devised for patients who are at the highest risk. Currently, plaque characterization at 3T requires a different set of coils compared with the global assessment of the entire course of the carotid arteries. Future generations of 16- to 32-channel carotid coils should be able to combine the best features of current 4- to 8-channel surface carotid coils and neurovascular coils. These will enable a comprehensive evaluation of the entire course of the carotid artery and detailed carotid bifurcation plaque characterization at 3T within a clinically acceptable 1-hour time frame. This comprehensive carotid artery evaluation with 3T MR imaging would be far superior to that which is possible with US or CT.