Ultrasound-based Artificial Intelligence for Classification of Carpal Tunnel Syndrome

Carpal tunnel syndrome (CTS) is one of the most prevalent peripheral neuropathies, impacting approximately 4% of the general population. It is typically classified into three degrees: mild, moderate, and severe. Accurate grading of carpal tunnel syndrome (CTS) is essential for determining appropriate treatment options, thereby playing a crucial role in optimizing patient outcomes. Electrophysiological testing (EST) is a key parameter for grading carpal tunnel syndrome (CTS). However, it is limited by several factors, including its invasive nature, poor reproducibility, and reduced sensitivity for detecting early-stage disease. Recently, ultrasound has gained widespread acceptance among clinicians for the assessment and grading of CTS. Nonetheless, radiologists often encounter challenges in this process due to the variability in image quality, differences in experience, and inherent subjectivity. To address these issues, artificial intelligence presents a promising solution. Therefore, this study aims to develop a deep learning model for grading CTS by leveraging multimodal imaging features, including B-mode ultrasound, superb microvascular imaging (SMI), and elastography. Additionally, the investigators intend to validate the model's effectiveness by testing it with images from various clinical centers, ensuring its generalizability across different clinical settings.