An End-to-End Deep Learning Method for Reconstructing SMS-PI Accelerated Musculoskeletal MRI

Motivation: Deep learning (DL) image reconstruction techniques are often utilized in knee MRI with 6-fold acceleration. However, applying these techniques to all musculoskeletal (MSK) MRI and achieving higher acceleration remains a challenge. Goal(s): Develop a dedicated DL image reconstruction method that combines simultaneous multislice (SMS) and parallel imaging (PI) acceleration for MSK MRI. Approach: DL approach combining slice separation and the reconstruction of image data from k-space for SMS-PI-accelerated MSK MRI. It features adjustable denoising strength and enhanced super-resolution image quality. Results: Proposed methods demonstrate improved reconstruction of MSK MRI scans at 8-fold and 12-fold accelerations compared to previous approaches, showing greater overall generalizability. Impact: Proposed methods facilitate the reconstruction of 8-fold accelerated 2D-TSE-MR images across various planes, contrasts, and MSK regions. Preliminary results indicate the feasibility of DL reconstruction at 12-fold acceleration, potentially allowing for significant speed-ups compared to the slower standard of care.