ABSTRACT Recent technological advancements may now allow shorter reliable magnetic resonance neurography (MRN) scans of the brachial plexus (BP) even at 1.5 T. The objective of this study was to compare ultra‐fast bilateral coronal 3D T2w TSE STIR (3D‐IR) with unilateral sagittal 2D T2w TSE Dixon (2D‐DX) in terms of image quality, diagnostic reliability, and time efficiency for BP MRN at 1.5 T. Free‐breathing 3D‐IR and 2D‐DX images acquired at 1.5 T were retrospectively collected from an equal number of healthy and pathologic cases and blindly compared by three raters. Pediatric cases, different acquisition methods, and low‐quality scans were excluded. Image quality was assessed objectively and subjectively. Image interpretability was assessed for each BP component. Diagnostic accuracy was evaluated against final radiologic reports. Inter‐rater agreement was calculated. The MRN scans of 72 BP from 36 patients (13 females, 50.3 ± 17.6 y) were included. 2D‐DX performed better for signal intensity (all p < 0.001), nerve‐to‐muscle contrast ratio ( p = 0.016), signal‐ and contrast‐to‐noise ratios (all p < 0.001), sharpness of C6 ( p = 0.001) and MT ( p = 0.004), general depiction of nerves ( p = 0.007), and inter‐rater agreement, while 3D‐IR was superior for noise intensity ( p < 0.001), nerve‐to‐fat contrast ratio ( p = 0.024) and vascular suppression ( p < 0.001). Sharpness of the remaining nerves, overall neurographic quality, resistance to motion artifacts, nerve identifiability, and image interpretability were comparable between the two sequences. Diagnostic accuracy was similar for both techniques (3D‐IR: 79.6%; 2D‐DX: 78.7%), although bilateral coverage via 3D‐IR required 46% of the acquisition time compared to 2D‐DX (4′30″ vs. 8′20″). According to these results, 3D STIR seems a viable alternative to 2D Dixon for bilateral BP MRN at 1.5 T, offering a significant time advantage without compromising diagnostic reliability. A full‐3D abbreviated protocol supported by time‐efficient breathing control and advanced reconstruction tools could further optimize 1.5‐T BP MRN workflows.
Zecca et al. (Thu,) studied this question.