Diagnosis of Cubital Tunnel Syndrome Using Deep Learning on Ultrasonographic Images

Abstract

Although electromyography is the routine diagnostic method for cubital tunnel syndrome (CuTS), imaging diagnosis by measuring cross-sectional area (CSA) with ultrasonography (US) has also been attempted in recent years. In this study, deep learning (DL), an artificial intelligence (AI) method, was used on US images, and its diagnostic performance for detecting CuTS was investigated. Elbow images of 30 healthy volunteers and 30 patients diagnosed with CuTS were used. Three thousand US images were prepared per each group to visualize the short axis of the ulnar nerve. Transfer learning was performed on 5000 randomly selected training images using three pre-trained models, and the remaining images were used for testing. The model was evaluated by analyzing a confusion matrix and the area under the receiver operating characteristic curve. Occlusion sensitivity and locally interpretable model-agnostic explanations were used to visualize the features deemed important by the AI. The highest score had an accuracy of 0.90, a precision of 0.86, a recall of 1.00, and an F-measure of 0.92. Visualization results show that the DL models focused on the epineurium of the ulnar nerve and the surrounding soft tissue. The proposed technique enables the accurate prediction of CuTS without the need to measure CSA.

Keywords: artificial intelligence; cubital tunnel syndrome; deep learning; ulnar nerve; ultrasonography.

Figure 1

(a) US probe placed on the medial epicondyle to visualize the ulnar nerve; (b) short-axis image of the ulnar nerve (red arrows) at the level of the medial epicondyle.

Figure 2

Flowchart of the proposed framework.

Figure 3

Images were randomly extracted by AI to be used as training data (light blue for control, orange for CuTS patients).

Figure 4

Block diagram of ResNet-50.

Figure 5

Block diagram of MobileNet_v2.

Figure 6

Block diagram of EfficientNet.

Figure 7

(a) A confusion matrix is a table of four combinations based on predicted and actual values and the presence or absence of disease; (b) diagnostic accuracy from the learning model is calculated from the confusion matrix created using testing data.

Figure 8

Area under the curve (AUC), based on the receiver operating characteristic (ROC) curve, was high for all learning models.

Figure 9

Confusion matrix of each learning model.

Figure 10

Visualization of the region of interest using occlusion sensitivity and image LIMEs. Learning models focus on neural interior and perineural tissues. The red circle is a cross section of the ulnar nerve in the original image. AI focused on hyperechoic changes in the ulnar nerve epithelium and hypoechoic changes in the ulnar nerve interior and surrounding tissue.