Feasibility of Shoulder MRI Protocol Using Deep Learning - assisted Iterative Algorithm Protocols: Comparison with Standard MRI Protocols

Motivation: Feasibility of Shoulder MRI Protocol Using Deep Learning - assisted Iterative Algorithm Protocols: Comparison with Standard MRI Protocols Goal(s): This study aims to explore the potential of DLIA for accelerating MRI of the shoulder Approach: 160 patients were divided into the standard group (SD) and the DLIA group. SD was processed using the SD sequence and SD reconstruction techniques, while DLIA utilized the DLIA sequence and DLIA reconstruction. Results: The SD group scan time 875 seconds, DLIA 540 seconds. Subjective showed no statistically significant changes. Compared SD group, DLIA group had significantly higher subjective image quality (P < 0.05). Impact: Shoulder MRI images using DLIA protocol could shorten the scanning time while maintaining comparable image quality and diagnostic capability.