Automated classification of hip fractures using deep convolutional neural networks with orthopedic surgeon-level accuracy: ensemble decision-making with antero-posterior and lateral radiographs.

Background and purpose - Deep-learning approaches based on convolutional neural networks (CNNs) are gaining interest in the medical imaging field. We evaluated the diagnostic performance of a CNN to discriminate femoral neck fractures, trochanteric fractures, and non-fracture using antero-posterior (AP) and lateral hip radiographs. Patients and methods - 1,703 plain hip AP radiographs and 1,220 plain hip lateral radiographs were included in the total dataset. 150 images each of the AP and lateral views were separated out and the remainder of the dataset was used for training. The CNN made the diagnosis based on: (1) AP radiographs alone, (2) lateral radiographs alone, or (3) both AP and lateral radiographs combined. The diagnostic performance of the CNN was measured by the accuracy, recall, precision, and F1 score. We further compared the CNN's performance with that of orthopedic surgeons. Results - The average accuracy, recall, precision, and F1 score of the CNN based on both anteroposterior and lateral radiographs were 0.98, 0.98, 0.98, and 0.98, respectively. The accuracy of the CNN was comparable to, or statistically significantly better than, that of the orthopedic surgeons regardless of radiographic view used. In the CNN model, the accuracy of the diagnosis based on both views was significantly better than the lateral view alone and tended to be better than the AP view alone. Interpretation - The CNN exhibited comparable or superior performance to that of orthopedic surgeons to discriminate femoral neck fractures, trochanteric fractures, and non-fracture using both AP and lateral hip radiographs.

Effect of direct oral anticoagulant for acute major cerebral artery occlusion in cardioembolic stroke/transient ischemic attack patients with non-valvular atrial fibrillation.

BACKGROUND: Direct oral anticoagulants (DOACs) can reduce the frequency of cardioembolic stroke with non-valvular atrial fibrillation as well as or better compared to vitamin K antagonists (VKAs). However, whether taking DOACs prior to stroke can prevent acute major cerebral artery occlusion (MCAO) has not been fully elucidated. METHODS: We enrolled patients who underwent cardioembolic stroke or transient ischemic attack with non-valvular atrial fibrillation who were admitted to our hospital between April 2011 and February 2017. The patients were classified into four groups based on anticoagulant medications prior to stroke: no oral anticoagulant (No OAC), VKA below therapeutic range on admission, VKA within therapeutic range on admission, and the DOAC group. We compared clinical backgrounds, National Institutes of Health Stroke Scale (NIHSS) scores, and MCAO prevalence on admission. We identified those patients with MCAO and investigated factors related to MCAO. RESULTS: A total of 287 patients were enrolled in the study (200 No OAC; 49 VKA below therapeutic range; 21 VKA within therapeutic range; and 17 DOAC). Median and interquartile range of NIHSS scores for each group were 10.5 (4-22) for No OAC; 14 (4-22) for VKA below therapeutic range; 8 (6-17) for VKA within therapeutic range; and 3 (1-9) for DOAC (P = 0.041). The prevalence of MCAO in each group was 40% in No OAC; 35% in VKA below therapeutic range; 29% in VKA within therapeutic range; and 6% in DOAC (P = 0.040). In total, 103 patients were identified with MCAO on admission. Multivariate analysis revealed that taking DOACs prior to stroke was significantly associated with MCAO (OR, 0.09; 95% CI, 0.004-0.75; P = 0.023). CONCLUSIONS: DOACs were an independent factor negatively correlated with MCAO in acute cardioembolic stroke with non-valvular atrial fibrillation.