Department of Trauma and Emergency Surgery, Chang Gung Memorial Hospital, Linkou ; Chang Gung University, Taoyuan, Taiwan.
Department of Surgery, Chang Gung Memorial Hospital, Linkou; Chang Gung University, Taoyuan, Taiwan.
Br J Radiol. 2023 Apr 1;96(1145):20220924. doi: 10.1259/bjr.20220924. Epub 2023 Mar 17.
To identify the feasibility and efficiency of deep convolutional neural networks (DCNNs) in the detection of ankle fractures and to explore ensemble strategies that applied multiple projections of radiographs.Ankle radiographs (AXRs) are the primary tool used to diagnose ankle fractures. Applying DCNN algorithms on AXRs can potentially improve the diagnostic accuracy and efficiency of detecting ankle fractures.
A DCNN was trained using a trauma image registry, including 3102 AXRs. We separately trained the DCNN on anteroposterior (AP) and lateral (Lat) AXRs. Different ensemble methods, such as "sum-up," "severance-OR," and "severance-Both," were evaluated to incorporate the results of the model using different projections of view.
The AP/Lat model's individual sensitivity, specificity, positive-predictive value, accuracy, and F1 score were 79%/84%, 90%/86%, 88%/86%, 83%/85%, and 0.816/0.850, respectively. Furthermore, the area under the receiver operating characteristic curve (AUROC) of the AP/Lat model was 0.890/0.894 (95% CI: 0.826-0.954/0.831-0.953). The sum-up method generated balanced results by applying both models and obtained an AUROC of 0.917 (95% CI: 0.863-0.972) with 87% accuracy. The severance-OR method resulted in a better sensitivity of 90%, and the severance-Both method obtained a high specificity of 94%.
Ankle fracture in the AXR could be identified by the trained DCNN algorithm. The selection of ensemble methods can depend on the clinical situation which might help clinicians detect ankle fractures efficiently without interrupting the current clinical pathway.
This study demonstrated different ensemble strategies of AI algorithms on multiple view AXRs to optimize the performance in various clinical needs.
确定深度卷积神经网络（DCNN）在踝关节骨折检测中的可行性和效率，并探索应用 X 射线多投影的集成策略。
踝关节 X 射线（AXR）是诊断踝关节骨折的主要工具。在 AXR 上应用 DCNN 算法可以提高检测踝关节骨折的诊断准确性和效率。
使用创伤图像登记处训练 DCNN，包括 3102 张 AXR。我们分别在前后位（AP）和侧位（Lat）AXR 上训练 DCNN。评估了不同的集成方法，如“求和”、“分裂-OR”和“分裂-Both”，以使用不同的视图投影来合并模型的结果。
AP/Lat 模型的个体敏感性、特异性、阳性预测值、准确性和 F1 评分分别为 79%/84%、90%/86%、88%/86%、83%/85%和 0.816/0.850。此外，AP/Lat 模型的受试者工作特征曲线下面积（AUROC）为 0.890/0.894（95%CI：0.826-0.954/0.831-0.953）。求和方法通过应用两个模型生成平衡的结果，AUROC 为 0.917（95%CI：0.863-0.972），准确率为 87%。分裂-OR 方法的敏感性提高到 90%，分裂-Both 方法的特异性达到 94%。
经过训练的 DCNN 算法可以识别 AXR 中的踝关节骨折。集成方法的选择可以取决于临床情况，这有助于临床医生在不中断当前临床路径的情况下高效地检测踝关节骨折。
本研究展示了 AI 算法在多视图 AXR 上的不同集成策略，以优化各种临床需求下的性能。
