Diffusion-Prepared Multi-Shot BSSFP Imaging with Gradient Stabilizer for High-Resolution and Distortion-Free Diffusion Imaging

Yu Gao, Fei Han, Z Zhou, X Zhong, Jack Neylon, Anand Santhanam, Yingli Yang, and Peng Hu

Medical Physics

AAPM: 60th Annual Meeting & Exhibition

Nashville, TN | 29 Jul 2018

Purpose

To propose a gradient stabilizer strategy to solve the k-space magnitude inconsistency problem associated with multi-shot diffusion-prepared approach.

Methods

A diffusion-prepared gradient-stabilized balanced steady-state free precession sequence (DP-GS-bSSFP, abbreviated as DP-GS. Sequenec without GS is abbreviated as DP) was proposed. The gradient stabilizer strategy converts malignant k-space magnitude inconsistency to phase inconsistency, so the images can be reconstructed using existing methods. Bioch simulations were performed to simulate the effect of gradient stabilizers with respect to phase error. A diffusion phantom was used to compare the ADC accuracy of DP-GS with diffusion-weighted single-shot Echo-planar imaging (DW-ssEPI). Five volunteers were recruited to evaluate the in-vivo performacnce for brain. DP without ECG gating, DP with ECG gating, and DP-GS without ECG gating were acquired ten times with b=500ms/mm2 in a single-shot manner to study the benefits of gradient stabilizer. Diffusion trace images and diffusion tensor images were acquired using four-shot DP-GS. ADC accuracy and geometric accuracy (quantified using landmark target registration error (TRE)) were compared with DW-ssEPI. MUSE was used for the multi-shot DP-GS reconstruction.

Results

Simulation showed that the proposed approach is insensitive to phase errors. It had satisfactory ADC accuracy on the phantom with <6% difference with DW-ssEPI for all diffusivity levels. In the repeated single-shot study, DP without ECG gating had severe signal void artificats and exhibitierd non-repeatable patterns which can be partially mitigated by ECG gating. Adding the gradient stabilizers provided stable signal magnitude accross all repetitions. Compared with DW-ssEPI, the four-shot DP-GS had substantially reduced distortion and susceptibility-related artifacts. High-resolution high-quality ADC maps and color-coded fractional anisotropy maps were generated using DP-GS. The mean/max TRE for DSW-ssEPI was around 2mm/10mm, while was about 0.6mm/1.5mm for DP-GS. Good grey matter (p=0.40) ADC correlations between the two methods were observed.

Conclusion

The proposed DP-GS approach provided high-resolution and distortion-free diffusion weighted and diffusion-tensor images.