Assessment of Ankle Fractures using Deep Learning Algorithms and Convolutional Neural Network

Abstract

Category: Ankle; Trauma Introduction/Purpose: Early and accurate detection of ankle fractures is crucial for reducing future complications. Radiographs are the most abundant imaging techniques for assessing fractures. We believe deep learning (DL) methods, through adequately trained deep convolutional neural networks (DCNNs), can assess radiographic images fast and accurate without human intervention. In this study, we aimed to assess the performance of two different DCNNs in detecting ankle fractures using radiographs compared to the ground truth. Methods: In this retrospective study, our DCNNs were trained using radiographs obtained from 1050 patients with ankle fracture and the same number of individuals with otherwise healthy ankles. Inception V3 and Renet50 pre-trained models were used in our algorithms. Danis-Weber classification method was used. Out of 1050, 72 individuals were labeled as occult fractures as they were not detected in the primary radiographic assessment. Using single-view radiographs was compared with 3-views (anteroposterior, mortise, lateral) for training the DCNNs. Results: Our DCNNs showed a better performance using 3-views images versus single-view based on greater values for accuracy, F-score, and area under the curve (AUC). The sensitivity and specificity in detection of ankle fractures using 3-views were 97.5% and 93.9% using Resnet50 compared to 98.7% and 98.6 using inception V3, respectively. Resnet50 missed 3 occult fractures while Inception V3 missed only one case. In cases that detected the fracture, the saliency map showed the location of the fracture ( Figure 1 ). Conclusion: The performance of our DCNNs showed a promising potential that can be considered in developing the currently used image interpretation programs or as a separate assistant to the clinicians to detect ankle fractures faster and more precisely.

Artificial intelligence and computer vision in orthopaedic trauma

“…5,6,8 Advances in the power of hardware and computing, the development of more accurate imaging techniques, and improvements in the capabilities of software by using computer vision, promise to increase the speed and accuracy of diagnosis and overcome concerns about reliability for the evaluation of images in trauma. 10,11 The widely used complex neural networks have several characteristic features and merits. Compared with conventional machine-learning methods such as decision tree, random forest, boosting, and support vector machines, which are typically used to solve problems in machine-learning on top of structured data, the convolutional filtering operations in a complex neural network can respond to local patterns in features of input which are spatially and temporally correlated.…”

Harnessing ResNet50 and SENet for enhanced ankle fracture identification

“…In comparison with previous literature on the use of deep learning for ankle fracture detection, such as the study in 2021 [ 30 ] that achieved high sensitivity and specificity using radiographs, our research contributes to the expanding body of knowledge by exploring the use of CT images and SENet-enhanced ResNet50. The referenced study demonstrated the effectiveness of employing Inception V3 and ResNet50 with radiographs, highlighting the potential of deep learning methods in accurately assessing fractures with high precision.…”

“…This approach underscores the diversity of methodologies and imaging modalities that can be leveraged to improve fracture detection. The high performance of DCNNs in the previous study, with sensitivity and specificity reaching up to 98.7% and 98.6% respectively using Inception V3, sets a benchmark for our work and others in the field [ 30 ]. Our findings contribute to this ongoing dialogue by suggesting that enhancements such as SENet can provide significant improvements, particularly in the context of CT-based fracture detection, which presents different challenges and opportunities compared to radiograph-based assessments.…”