Knee Osteoarthritis DTI using Deep Learning Reconstruction: A Prospective Study on Quantitative Accuracy and Diagnostic Efficacy

Motivation: Diffusion-tensorimaging(DTI) has the potential to serve as a marker of joint degeneration and can elucidate the integrity of articular cartilage with good reproducibility. Goal(s): The use of DTI as a marker in the clinical setting is challenging due to the difficulty in balancing signal-to-noise ratio, resolution, and acceptable scan time. Approach: A deep learning reconstruction (DLR) technique breaks the tradeoffs between signal-to-noise ratio, spatial resolution, and scan time. Results: High-resolution and fast DTI was potentially achieved to identify cartilage injury or degeneration has good test-retest reproducibility, and may be accurate in discriminating healthy subjects from subjects with OA. Impact: The application of deep learning reconstruction technology to select the appropriate acceleration factor can significantly shorten the acquisition time of DTI sequence images of the knee joint with good test-retest repeatability.