The Addition of a 3D Balanced Steady‐State Free Precession Pulse Sequence Improves Magnetic Resonance Imaging Identification of Certain Canine Cranial Nerves

In people, magnetic resonance imaging (MRI) with high-resolution, high T2-weighted (T2w) contrast-balanced steady-state free precession (b-SSFP) pulse sequences and 3D fast spoiled gradient-echo (FSPGR) pulse sequences can improve visualization of cranial nerves (CNs) and associated osseous foramina. This prospective, observational study aimed to determine whether the addition of b-SSFP and FSPGR sequences improves confidence in the identification of canine CNs compared to the standard imaging protocol at 1.5 T. The head of 10 canine cadavers was imaged, including transverse T1-weighted (T1w) and T2w fast spin echo (FSE); sagittal T2w FSE; 3D Fast Imaging Employing Steady-State Acquisition with Cycling (FIESTA-C); and 3D Liver Acquisition with Volume Acceleration (LAVA). Board-certified radiologists rated their confidence in identifying CNs II-XII using a four-point scale. Observers evaluated standard sequences alone, then with the addition of transverse FIESTA-C, and finally with the inclusion of FIESTA-C and transverse LAVA. Scores between protocols were compared with the Wilcoxon matched-pairs signed-rank test. Interobserver agreement was measured with weighted kappa. Median score increased when adding FIESTA-C for all observers for nerves VII, VIII, IX, and X, and for two observers for nerves III and XI. Matched-pairs comparison revealed significant differences for nerves VII and VIII for all observers, and for nerves III, IX, X, XI, and XII for two observers. Further addition of LAVA generally did not improve confidence. Interobserver agreement was overall moderate to substantial, but slight to fair for some nerves and pulse sequences. Addition of a 3D b-SSFP sequence may improve identification of some CNs, especially VII and VIII, which are somewhat commonly affected by middle ear pathology in dogs.