Distal radius fractures (DRFs) are one of the most common types of fracture and one which is often treated surgically. Standard X-rays are obtained for DRFs, and in most cases that have an intra-articular component, a routine CT is also performed. However, it is estimated that CT is only required in 20% of cases and therefore routine CT's results in the overutilisation of resources burdening radiology and emergency departments. In this study, we explore the feasibility of using deep learning to differentiate intra- and extra-articular DRFs automatically and help streamline which fractures require a CT. Retrospectively x-ray images were retrieved from 615 DRF patients who were treated with an ORIF at the Royal Brisbane and Women's Hospital. The images were classified into AO Type A, B or C fractures by three training registrars supervised by a consultant. Deep learning was utilised in a two-stage process: 1) localise and focus the region of interest around the wrist using the YOLOv5 object detection network and 2) classify the fracture using a EfficientNet-B3 network to differentiate intra- and extra-articular fractures. The distal radius region of interest (ROI) detection stage using the ensemble model of YOLO networks detected all ROIs on the test set with no false positives. The average intersection over union between the YOLO detections and the ROI ground truth was The proposed DRF classification framework using ensemble models of YOLO and EfficientNet achieved satisfactory performance in intra- and extra-articular fracture classification. This work demonstrates the potential in automatic fracture characterization using deep learning and can serve to streamline decision making for axial imaging helping to reduce unnecessary CT scans.
Intramedullary nailing is the standard fixation method for displaced diaphyseal fractures of the tibia in adults. Anecdotal clinical evidence indicates that current nail designs do not fit optimally for Asian patients. This study aimed to develop a method to quantitatively assess the fitting of two nail designs for Asian tibiae. We used 3D models of two different tibial nail designs (ETN (Expert Tibia Nail) and ETN-Proximal-Bend, Synthes), and 20 CT-based 3D cortex models of Japanese cadaver tibiae. The nail models were positioned inside the medullary cavity of the intact bone models. The anatomical fitting between nail and bone was assessed by the extent of the nail protrusion from the medullary cavity into the cortical bone, which in a real bone would lead to axial malalignments of the main fragments. The fitting was quantified in terms of the total surface area, and the maximal distance of nail protrusion. In all 20 bone models, the total area of the nail protruding from the medullary cavity was smaller for the ETN-Proximal-Bend (average 540 mm2) compared to the ETN (average 1044 mm2). Also, the maximal distance of the nail protruding from the medullary cavity was smaller for the ETN-Proximal-Bend (average 1.2 mm) compared to the ETN (average 2.7 mm). The differences were statistically significant (p < 0.05) for both the total surface area and the maximal distance measurements. For all bone models, the nail protrusion occurred on the posterior side in the middle third of the tibia. For 12 bones the protrusion was slightly lateral to the centre of the shaft, for seven bones it was centred, and for one bone it was medial to the shaft. The ETN-Proximal-Bend shows a statistical significantly better intramedullary fit with less cortical protrusion than the original ETN. The expected clinical implications of an improved anatomical nail fit are fewer complications with malreduction and malalignments, a lower likelihood for fracture extension and/or new fracture creation during the nail insertion as well as an easier handling for the nail insertion. By utilising computer graphical methods we were able to conduct a quantitative fit assessment between implanted nail and bone geometry in 3D. In addition to the application in implant design, the developed method could potentially be suitable for pre-operative planning enabling the surgeon to choose the most appropriate nail design.
