Related Experiment Video
Updated: Jun 5, 2025
08:39
Assessing Cortical Cerebral Microinfarcts on High Resolution MR Images
Published on: November 20, 2015
13.4K
Evaluating CNN Architectures for the Automated Detection and Grading of Modic Changes in MRI: A Comparative Study
Li-Peng Xing1,2, Gang Liu2, Hao-Chen Zhang1
1State Key Laboratory of Reliability and Intelligence of Electrical Equipment, School of Health Sciences & Biomedical Engineering, Hebei University of Technology, Tianjin, China.
Orthopaedic Surgery
|December 5, 2024
Summary
This study developed a convolutional neural network (CNN) for grading Modic changes (MCs) on MRI. The YOLOv8 model demonstrated superior performance compared to junior doctors, enhancing diagnostic reliability.
Area of Science:
- Artificial Intelligence in Medical Imaging
- Spinal Imaging and Diagnostics
- Deep Learning for Radiomics
Background:
- Modic changes (MCs) classification is standard for vertebral marrow changes on MRI but is semiquantitative and sensitive to imaging variations.
- A quantitative MC grading method was proposed in 2021, yet automated grading tools are lacking.
- The need for reliable, automated MC grading systems is critical for consistent clinical interpretation.
Purpose of the Study:
- To develop and evaluate a convolutional neural network (CNN) for automated detection and grading of Modic changes (MCs) based on maximum vertical extent.
- To assess the generalization performance of the CNN models.
- To compare the CNN's performance against junior doctors and evaluate AI assistance's impact on diagnostic consistency.
Main Methods:
- Retrospective analysis of 139 patient MRIs with MCs, annotated by a spine surgeon.
- Development of YOLOv8 and YOLOv5 models using PyTorch, incorporating data enhancement and transfer learning.
- Performance evaluation using precision, recall, F1 score, and mAP50, with comparisons on a separate dataset and AI-assisted junior doctor grading.
Main Results:
- YOLOv8 achieved superior performance on the test set (precision 81.60%, recall 80.90%, mAP50 84.40%) compared to YOLOv5.
- On Dataset 2, YOLOv8 outperformed junior doctors (precision 95.1% vs. 72.5%, recall 68.3% vs. 60.6%).
- AI assistance significantly improved junior doctor agreement with senior spine surgeons (Cohen's kappa from 0.368 to 0.681).
Conclusions:
- The YOLOv8 model demonstrates significant superiority over YOLOv5 for detecting and grading Modic changes.
- YOLOv8 performance exceeds that of junior doctors, indicating its potential as a powerful diagnostic tool.
- AI assistance using YOLOv8 enhances junior doctors' capabilities and improves the overall reliability of spinal diagnoses.
More Related Videos
Related Concept Videos
01:24Magnetic Resonance Imaging
4.9K
Magnetic resonance imaging (MRI) is a noninvasive medical imaging technique based on a phenomenon of nuclear physics discovered in the 1930s, in which matter exposed to magnetic fields and radio waves was found to emit radio signals. In 1970, a physician and researcher named Raymond Damadian noticed that malignant (cancerous) tissue gave off different signals than normal body tissue. He applied for a patent for the first MRI scanning device in clinical use by the early 1980s. The early MRI...
4.9K
01:21Imaging Studies for Cardiovascular System IV: CMRI
5
Cardiovascular magnetic resonance imaging, or CMRI, is a non-invasive diagnostic test that employs a magnetic field and radiofrequency waves to create precise images of the heart and arteries. It provides comprehensive information about cardiac anatomy, function, perfusion, and tissue characterization without ionizing radiation.IndicationsCMRI diagnoses various heart conditions, including tissue damage from heart attacks, ischemic heart disease, myocarditis, aortic issues (tears, aneurysms,...
5