Detecting pediatric wrist fractures using deep-learning-based object detection.
Researchers
Journal
Modalities
Models
Abstract
Missed fractures are the leading cause of diagnostic error in the emergency department, and fractures of pediatric bones, particularly subtle wrist fractures, can be misidentified because of their varying characteristics and responses to injury.This study evaluated the utility of an object detection deep learning framework for classifying pediatric wrist fractures as positive or negative for fracture, including subtle buckle fractures of the distal radius, and evaluated the performance of this algorithm as augmentation to trainee radiograph interpretation.We obtained 395 posteroanterior wrist radiographs from unique pediatric patients (65% positive for fracture, 30% positive for distal radial buckle fracture) and divided them into train (nā=ā229), tune (nā=ā41) and test (nā=ā125) sets. We trained a Faster R-CNN (region-based convolutional neural network) deep learning object-detection model. Two pediatric and two radiology residents evaluated radiographs initially without the artificial intelligence (AI) assistance, and then subsequently with access to the bounding box generated by the Faster R-CNN model.The Faster R-CNN model demonstrated an area under the curve (AUC) of 0.92 (95% confidence interval [CI] 0.87-0.97), accuracy of 88% (nā=ā110/125; 95% CI 81-93%), sensitivity of 88% (nā=ā70/80; 95% CI 78-94%) and specificity of 89% (nā=ā40/45, 95% CI 76-96%) in identifying any fracture and identified 90% of buckle fractures (nā=ā35/39, 95% CI 76-97%). Access to Faster R-CNN model predictions significantly improved average resident accuracy from 80 to 93% in detecting any fracture (Pā<ā0.001) and from 69 to 92% in detecting buckle fracture (Pā<ā0.001). After accessing AI predictions, residents significantly outperformed AI in cases of disagreement (73% resident correct vs. 27% AI, Pā=ā0.002).An object-detection-based deep learning approach trained with only a few hundred examples identified radiographs containing pediatric wrist fractures with high accuracy. Access to model predictions significantly improved resident accuracy in diagnosing these fractures.Ā© 2023. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.