Fully automated patellofemoral segmentation from mri using holistically nested networks: implications for evaluating patellofemoral osteoarthritis, pain, injury, pathology, and adolescent development

Our understanding of the relationship between 3D bone shape and knee osteoarthritis (OA), as well as our ability to investigate potential causative factors of OA, has been hampered by the time intensive nature of manually segmenting bone from MR images. Thus, we aim to develop and validate a fully automated deep learning framework, based 2D holistically nested networks (HNN) architecture, for segmenting the patella and distal femur cortex, in both adults and actively growing adolescents. Data from 93 subjects, obtained from on IRB-approved study, formed the database. Our HNN with multi-feature architecture segmentation provides a fully automatic technique capable of delineating the often indistinct interfaces between the bone and other joint structures with an accuracy better than nearly all previous techniques, even when active growth plates are present.