Comparison of Three-Dimensional Multi-Echo in Steady-State Acquisition and Short-Tau Inversion Recovery Sequences in Brachial Plexus Magnetic Resonance Neurography

BACKGROUND AND PURPOSE: The short-tau inversion recovery fast spin echo (STIR-FSE) pulse sequence is commonly used for 3D magnetic resonance neurography (MRN) of the brachial plexus (BP) due to its low B₀ susceptibility and robust fat suppression. An alternative 3D acquisition is a multi-echo in steady-state acquisition (MENSA-NERVE) sequence that typically provides more robust vascular suppression, but poorer fat suppression due to its use of a water excitation pulse. MATERIALS AND METHODS: A total of 50 subjects (24 female, median age [range]: 34 years [14-73]) who underwent unilateral BP MRN at 3-Telsa where non-contrast STIR-FSE and MENSA-NERVE were sequentially acquired and retrospectively analyzed. Sequences were assessed by 2 radiologists for vascular suppression, fat suppression, and visualization of the BP nerve roots, trunks, and cords as well as of the suprascapular nerve and axillary nerve. Pooled measurements were analyzed using the Wilcoxon signed-rank test, and agreement proportion was used to assess inter-rater/inter-sequence diagnostic agreement. RESULTS: MENSA-NERVE outperformed STIR-FSE for vascular suppression (p<0.001). Fat suppression assessments favored STIR-FSE over MENSA-NERVE (p=0.05 to <0.001), with incomplete fat suppression obscuring nerve visualization in MENSA-NERVE primarily observed only at the lung apex. MENSA-NERVE provided superior visualization of the suprascapular and axillary nerves (p=0.004 to <0.001). Interrater agreement of visualized pathology was good to excellent (agreement=0.74-0.92) for both sequences. Inter-sequence agreement was also excellent (agreement=0.80-0.88). CONCLUSION: MENSA-NERVE provided superior vascular suppression and nerve conspicuity as compared to non-contrast STIR-FSE and with less robust fat suppression at the neck-shoulder junction and lung apex.