Related Experiment Video
Updated: May 8, 2026
07:00
Diffusion Tensor Magnetic Resonance Imaging in Chronic Spinal Cord Compression
Published on: May 7, 2019
Spinal Cord Radiomics-Driven Machine Learning Predicts Meaningful Clinical Improvement After Surgery for Degenerative
Ramesh M Arnest1, Kevin M Koch2, Matthew D Budde1
1Department of Neurosurgery, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI, 53226, USA.
Journal of Imaging Informatics in Medicine
|May 6, 2026
Summary
Machine learning models using MRI radiomics and clinical data can predict patient recovery after degenerative cervical myelopathy surgery. This approach offers improved prognostication for functional and quality of life improvements.
Area of Science:
- Neurosurgery
- Radiology
- Artificial Intelligence
Background:
- Predicting surgical outcomes in degenerative cervical myelopathy (DCM) is challenging due to limitations of conventional MRI and clinical scores.
- Radiomics quantifies MRI-derived tissue heterogeneity, offering potential imaging biomarkers for recovery.
- Machine learning models may enhance prediction of functional recovery and quality of life post-surgery.
Purpose of the Study:
- To evaluate machine learning models incorporating radiomic features from preoperative MRI and clinical variables for predicting minimum clinically important difference (MCID) achievement in DCM patients.
- To assess the predictive performance of models using radiomics alone, clinical data alone, or a combination of both.
Main Methods:
- A prospective observational cohort study included 46 DCM patients undergoing surgery.
- Preoperative 3D T2-weighted MRI was analyzed for radiomic features (Shape3D, First-Order, GLCM, GLSZM) using PyRadiomics.
- Clinical variables (age, sex, symptom duration, T2 hyperintensity, mJOA, SF-36 PCS) were collected. Predictive models were developed and validated.
Main Results:
- The combined radiomics-clinical model showed the best performance for predicting mJOA MCID (AUC = 0.88 ± 0.13).
- For SF-36 PCS MCID, the combined model achieved an AUC of 0.78 ± 0.17 and an AUCPR of 0.82 ± 0.14.
- SHapley Additive exPlanations identified texture-based radiomic features and age as key predictors for mJOA MCID, and first-order features and baseline SF-36 PCS for SF-36 PCS MCID.
Conclusions:
- MRI-based spinal cord radiomics significantly improves the prediction of meaningful postoperative recovery in DCM patients compared to clinical data alone.
- Radiomic features show potential as imaging biomarkers for individualized prognostication in DCM.
- Integrating radiomics into predictive models can enhance surgical outcome prediction for better patient management.