Improving echo-planar imaging-based diffusion weighted imaging of the head and neck using local shim coils

Purpose: To evaluate the performance of additional local shim coils within the head/neck surface coil for diffusion-weighted imaging (DWI) MRI of the head/neck region using a 3-Tesla scanner. Material and Methods: Ten healthy volunteers underwent single-shot echo-planar DWI and T1 gradient echo acquisition of the head/neck region, without (sEPI; sGRE) and with (cEPI; cGRE) the use of additional shim coils integrated into the head/neck coil. Additionally, B 0 field maps with and without additional shim coils were acquired using a dual-echo T1w GRE sequence. Acquisitions were performed on a clinical 3T scanner. In the DWI images, visual evaluation was performed for overall image quality (OAIQ), signal loss, distortions and ghosting. Failure of fat saturation was visually assessed for all sequences. Assessments were quantified using a 4-point scale by two radiologists in consensus. The effects on static field homogeneities were evaluated visually in the acquired field maps. Apparent diffusion coefficients (ADCs) were quantified in the spinal cord, spinal fluid and trapezius muscle for sEPI and cEPI. Results: Compared to sEPI, cEPI significantly improved OAIQ, distortion and ghosting and decreased signal loss (p≤0.008). Additionally, cEPI resulted in significantly improved fat saturation compared to sEPI, sGRE and cGRE (p<0.001). Regarding the static magnetic field, additional local shim coils resulted in marked homogenisation, especially in the lower/posterior neck. ADC values did not differ significantly (p≥0.08). Conclusions: Additional local shim coils integrated into the head/neck surface coil improve homogenisation of the static magnetic field and providing improved DWI quality compared to standard sEPI. For T1-weighted sGRE, additional coils have no significant effect on fat saturation.