Abstract

Motivation: Bony structures of the human knee play import roles in the stability and normal function and show alterations in morphology with diseases such as osteoarthritis. Goal(s): Advance MR imaging of the bone. Approach: Three techniques utilizing delta ultrashort echo time (UTE), field echo (FE), and FE with deep learning reconstruction (DLR) with simple post-processing was developed to image knees of volunteers. SNR and CNR were determined. Results: We demonstrate that FE DLR, compared to UTE, was advantageous for greater signal to noise ratio for the cortical bone, as well as higher resolution that depicted trabecular structures. Impact: This study advances MR imaging of the bone, for specific use cases such as morphologic evaluation. The FE technique in particular is immediately translatable using existing sequences to help evaluate knee bone in the clinical settings.