FLAIR-Imaging in multiple sclerosis: A comparative MR-Study at 1.5 and 3.0 Tesla

Purpose: The FLAIR technique is one of the most sensitive techniques for the visualization of white matter lesions (WML) in the brain. However, operating at higher field strength imaging parameters have to be adapted and optimized. Aim of this prospective study was 1) to identify the optimal TE for FLAIR-imaging at 3T assessing three different echo-times qualitatively and quantitatively (TE 100, 120 and 140ms) and 2) to evaluate the diagnostic efficacy of high-field 3T MRI FLAIR-Imaging in comparison to conventional 1.5T MRI in patients with proven multiple sclerosis.

Method and materials: In 30 consecutive patients MRI was performed with a 3T system (Gyroscan Intera, Philips Medical Systems). Protocols were modified for 3T using the increased SNR to acquire more and thinner slices while maintaining comparable scan times with regard to the standard protocols at 1.5T (3T: 27 slices, 4mm; 1.5T: 20 slices, 6mm). Axial FLAIR, sagittal FLAIR, axial T2w-TSE and axial T1w-FFE images were obtained. Axial FLAIR-images were acquired with a TE of 100, 120 and 140 ms while TR (11000 ms), TI (2600 ms) and reconstructed voxel size (0.45×0.45×4.00mm) were kept constant. 13 patients with clinically proven multiple sclerosis (4 m, 9 f, mean age 38yrs) were further examined at a 1.5T system (Gyroscan Intera, PMS). FLAIR-images were ranked by two observers for each patient qualitatively in matched pairs (different TE or different field strength) with regard to artifacts, gray-white differentiation, image noise and overall diagnostic value. Quantitative SI-measurements were obtained in gray matter, white matter, CSF and representative white matter lesions (WML). In patients with multiple sclerosis the total number of lesions detected was noted for each field strength. SNR and CNR values were calculated; statistical significance was determined by Wilcoxon signed-rank test and Student t-test.

Results: Comparing different echo times at 3T a TE of 100 and 120 ms proved superior in all qualitative categories when compared to 140 ms (p<0.0001). In the quantitative assessment 100 and 120 ms had a significantly higher CNR with regard to gray-white differentiation and lesion conspicuity when compared to 140 ms. The SNR of CSF was lowest for 100 ms and highest for 140 ms. CNR of WML was highest for 120 ms (19.5) intermediate for 100 ms (17.5) and lowest for 140 ms (15.5) (p<0.003). In patients with multiple sclerosis 3T was judged superior to 1.5T for lesion conspicuity (p<0.0009) and overall diagnostic value (p<0.008), whereas no difference was found for gray-white differentiation and image noise. 3T was inferior to 1.5T with regard to artifacts (p<0.0002). SNR of CSF was lower at 3T (8.2 vs. 11.5, p<0.05), but CNR for WML was not different at both field strengths (19.8 vs. 19.4). However, significantly more WML were detected at 3T than at 1.5T (175 vs. 124, p<0.0007).

Conclusion:

The optimal Echo Time for FLAIR-imaging at 3T is 120 ms, due to the significantly higher CNR of WM lesions, as compared to 100 and 140 ms. A TE of 140 ms is not suited for FLAIR-Imaging at 3T. In patients with multiple sclerosis high-field 3T MRI proved superior to 1.5T in terms of lesion detection, using a TE of 120 ms for FLAIR-Imaging. Trading of higher SNR at 3T for better spatial resolution the same CNR level could be maintained, increasing lesion detectability. Thus, high field MRI may further strengthen the role of MRI as the most sensitive paraclinical test available for early diagnosis of multiple sclerosis.