Related Experiment Video
Updated: Jul 15, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
42.6K
Automated Segmentation and Diagnostic Measurement for the Evaluation of Cervical Spine Injuries Using X-Rays
Jae Hyuk Shim1, Woo Seok Kim2, Kwang Gi Kim3
1Department of Biomedical Engineering, Gil Medical Center, Gachon University College of Medicine, Incheon, Korea.
Journal of Imaging Informatics in Medicine
|February 21, 2024
Summary
Machine learning and computer vision automate cervical spine X-ray analysis, accurately measuring diagnostic metrics for injury assessment and surgical evaluation. This technology shows significant potential for clinical applications in diagnosing cervical spine conditions.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Orthopedics
Background:
- Accurate assessment of cervical spine X-rays is vital for diagnosing injuries and evaluating surgical outcomes.
- Traditional manual measurements can be time-consuming and prone to variability.
- Automated methods using machine learning and computer vision offer potential for improved efficiency and accuracy.
Purpose of the Study:
- To evaluate the efficacy of multiclass segmentation using deep learning models for automated measurement of diagnostic metrics on cervical spine X-rays.
- To compare the accuracy of automatically derived metrics with manually measured values.
Main Methods:
- Utilized a dataset of 852 cervical X-rays from Gachon Medical Center.
- Employed EfficientNetB4, DenseNet201, and InceptionResNetV2 architectures for multiclass segmentation of craniofacial bones and cervical spine (C1-C7).
- Compared automatically measured diagnostic metrics (e.g., McGregor's line, cervical sagittal vertical axis) with manually measured metrics using Pearson's correlation coefficient and paired t-tests.
Main Results:
- High Dice coefficient values achieved for cervical spine segmentation (0.93-0.96) and moderate values for craniofacial bones (0.69-0.81).
- Strong Pearson's correlation coefficients observed between manual and automatic measurements for key metrics like cervical sagittal vertical axis (r=0.99) and space available cord (r=0.94).
- No statistically significant differences (P<0.05) were found between manual and automatic measurement methods for any metric.
Conclusions:
- Multiclass segmentation effectively automates the measurement of diagnostic metrics on cervical spine X-rays.
- The developed automated methods demonstrate significant clinical potential for diagnosing cervical spine injuries and evaluating surgical outcomes.
- This approach offers a reliable and efficient alternative to manual measurements in clinical practice.
