Feasibility of high-resolution deep learning reconstruction synthetic T2 maps contrast images in diagnosis of joint degeneration

Motivation: Combing T2 mapping with routine knee MRI increase diagnostic efficacy especially early identification of joint degeneration. Goal(s): To explore the feasibility of high-resolution deep-learning reconstruction (DLR) synthetic MRI with equivalent to higher diagnosis performance using conventional routine knee MRI as reference. Approach: To test reliability of T2 mapping for in vitro phantom and in vivo human knees and explore the image quality of DLR synthetic contrast MR images in comparison with conventional MR images. Results: DLR synthetic MRI offer reliable T2 mapping and provide sufficient image quality for diagnosis of knee. Impact: There has been always a demand for high-resolution knee MRI acquisition and straightforward diagnosis in clinics. DLR synthetic MRI is a rapid-acquisition and high-image-quality contrast images and quantitative maps and may improves diagnosis, prognosis and follow-ups.