Related Experiment Video
Updated: Sep 26, 2025
07:00
Diffusion Tensor Magnetic Resonance Imaging in Chronic Spinal Cord Compression
Published on: May 7, 2019
9.0K
Deep learning algorithm to evaluate cervical spondylotic myelopathy using lateral cervical spine radiograph.
Gun Woo Lee1, Hyunkwang Shin2, Min Cheol Chang3
1Department of Orthopedic Surgery, Yeungnam University College of Medicine, Yeungnam University, Medical Center, 170 Hyonchung-ro, Namgu, Daegu, 42415, South Korea.
BMC Neurology
|April 21, 2022
Summary
A deep learning model using cervical spine X-rays accurately detects cervical spondylotic myelopathy (CSM). This convolutional neural network (CNN) shows promise for diagnosing CSM in clinical settings.
Area of Science:
- Medical imaging analysis
- Artificial intelligence in healthcare
- Spine diagnostics
Background:
- Deep learning (DL) is a machine learning technique excelling in image analysis.
- Convolutional neural networks (CNNs) are effective for image recognition and classification.
- Cervical spondylotic myelopathy (CSM) is a condition affecting the cervical spine.
Purpose of the Study:
- To develop a CNN model for detecting cervical spondylotic myelopathy (CSM).
- To utilize lateral cervical spine radiographs as input for the CNN model.
Main Methods:
- Retrospective study of 207 patients (96 with CSM, 111 without).
- A CNN algorithm was developed and trained on 70% of lateral cervical spine radiographs.
- Model performance was evaluated on the remaining 30% of images.
Main Results:
- The CNN model achieved an accuracy of 87.1% in detecting CSM.
- The area under the curve (AUC) for CSM detection was 0.864.
Conclusions:
- A CNN model utilizing lateral cervical spine radiographs can aid in diagnosing CSM.
- This AI approach offers a potential tool for improving CSM detection.
