Time-resolved contrast enhanced magnetic resonance angiography of the head and neck at 3.0 tesla: initial results.

OBJECTIVES

We sought to implement and evaluate a high-performance, extended field of view protocol for time-resolved contrast-enhanced magnetic resonance imaging (CEMRA) of the carotid circulation by using a dedicated neurovascular (NV) array coil.

MATERIALS AND METHODS

A total of 16 adult volunteers and 20 clinical patients with suspected cerebrovascular disease (15 male, 21 female, 25-82 years of age) were scanned with a fast 3D MRA sequence (TR/TE: 2.16/1 milliseconds, sampling BW: 1090 Hz/pixel), with echo-sharing and parallel acquisition. All studies were performed on a 3.0 T MR system using an 8-channel neurovascular array coil. After injection of 6 mL of gadodiamide at 3 mL/s, a coronal 3D data set with in-plane resolution of 1 x 1.3 was implemented for 10 consecutive measurements each 1.8 seconds apart. The subjects subsequently underwent high spatial-resolution (in-plane: 0.8 x 0.9) CEMRA for comparative analysis. The quality of segmental arterial anatomy and the presence and degree of the arterial stenosis were evaluated by 2 neuroradiologists. The interobserver variability was tested by kappa statistics and comparative analysis between the TR-CEMRA and high spatial-resolution CEMRA was evaluated by mean of the Spearman rank correlation coefficient.

RESULTS

Craniocervical arteries were visualized with good image quality and definition in the diagnostic range. Occlusive disease was detected in 42 (reader A) and 44 (reader B) arterial segments with excellent interobserver agreement (kappa =0.89; 95% confidence interval 0.82-0.96). There was a significant correlation between the TR-CEMRA and high spatial-resolution CEMRA (Rs = 0.91 and 0.93, for readers A and B, respectively) for the degree of stenosis. Three aneurysms, 3 AVMs, 1 AV-fistula, and 2 subclavian steals were detected by both observers and were confirmed by correlative imaging.

CONCLUSION

Time-resolved CEMRA at 3.0 T is reliable and versatile, providing 3-dimensional time-resolved data sets with high spatial (in plane: 1.3 x 1 mm2) and temporal (1.8 seconds) resolution over a large field of view. The higher signal-to-noise ratio gain at 3.0 T can be used effectively to improve performance of fast imaging and to support aggressive parallel acquisition protocols, as in the present study. Further clinical studies are required to establish the range of applications and the accuracy of the technique.