Detection of Shoulder Rotator Cuff Tears from X-Ray Images Using Convolutional Neural Network

Abstract

Shoulder rotator cuff tears are common shoulder joint disorder that can be caused by daily activities, sports injuries, or aging. Magnetic resonance imaging (MRI) is widely used for the diagnosis of rotator cuff tears, while radiography is simpler than MRI as a quick and widespread imaging modality. Thus, the present study developed a shoulder rotator cuff tear detection method using deep learning on shoulder X-ray images. In a two-class classification by the severity of a tear, the detection accuracies of five shoulder regions of interest and two deep learning model combinations were compared, with those results showing a maximum accuracy of 79.3%.