AI-assisted Collaborative Reconstruction for Highly-accelerated DW-PROPELLER-EPI

Motivation: Diffusion-weighted PROPELLER-EPI (DW-PROPELLER-EPI) can achieve higher resolution for DWI compared to single-shot EPI, but it suffers from long scanning time and image blurring induced by geometric distortion unless accelerating the acquisition of each blade. Goal(s): To reduce noise amplification and improve the measurement of inter-shot phase variations related to highly-accelerated DW-PROPELLER-EPI. Approach: We proposed an AI-assisted Collaborative Reconstruction for PROPELLER-EPI using POCSMUSE (AI-CORPUSE) framework that can 1) reduce the noise amplification, and 2) improve inter-shot phase correction under high acceleration. Results: This method can accelerate the acquisition of DW-PROPELLER-EPI and achieve brain DWI with higher image quality. Impact: High geometric fidelity and high-resolution brain DWI with reasonable scan time may benefit clinical applications and neuroscience research.