Application Research of AI-assisted Compressed Sensing Technology in MRI Scanning of the Knee Joint: 3D-MRI Perspective

Motivation: The broad clinical application of knee 3D-MRI has been constrained by scanning time. Goal(s): To investigate the potential of AI-assisted compressed sensing (ACS) in knee MRI to optimize the scanning process. Approach: 3D-ACS, 3D compressed sensing (CS), and 2D parallel acquisition technology (PAT) scans were performed. The 3D-ACS images underwent 3.5 mm/2.0 mm multiplanar reconstruction (MPR); radiologists evaluated the quality of images and diagnosed diseases. Results: 3D-ACS provided poorer bone structure visualization, improved cartilage visualization, and less satisfactory axial images with 3.5 mm/2.0 mm MPR than 2D-PAT. High levels of diagnostic agreement and accuracy were observed across all diagnoses. Impact: 3D-ACS provided poorer bone structure visualization, improved cartilage visualization, and less satisfactory axial images with 3.5 mm/2.0 mm MPR than 2D-PAT. High levels of diagnostic agreement and accuracy were observed across all diagnoses.