The aim of this study was to create artificial intelligence (AI) software with the purpose of providing a second opinion to physicians to support distal radius fracture (DRF) detection, and to compare the accuracy of fracture detection of physicians with and without software support. The dataset consisted of 26,121 anonymized anterior-posterior (AP) and lateral standard view radiographs of the wrist, with and without DRF. The convolutional neural network (CNN) model was trained to detect the presence of a DRF by comparing the radiographs containing a fracture to the inconspicuous ones. A total of 11 physicians (six surgeons in training and five hand surgeons) assessed 200 pairs of randomly selected digital radiographs of the wrist (AP and lateral) for the presence of a DRF. The same images were first evaluated without, and then with, the support of the CNN model, and the diagnostic accuracy of the two methods was compared.Aims
Methods
This study identified variables which influence the outcome of surgical management on 126 ununited scaphoid fractures managed by internal fixation and non-vascular bone grafting. The site of fracture was defined by a new method: the ratio of the length of the proximal fragment to the sum of the lengths of both fragments, calculated using specific views in the plain radiographs. Bone healing occurred in 71% (89) of cases. Only the site of nonunion (p = 1 × 10−6) and the delay to surgery (p = 0.001) remained significant on multivariate analysis. The effect of surgical delay on the probability of union increased as the fracture site moved proximally. A prediction model was produced by stepwise logistic regression analysis, enabling the surgeon to predict the success of surgery where the site of the nonunion and delay to surgery is known.
