Related Experiment Video
Updated: Jun 8, 2026
08:11
Measuring Progressive Neurological Disability in a Mouse Model of Multiple Sclerosis
Published on: November 14, 2016
11.1K
Evaluation of Disability Progression in Multiple Sclerosis via Magnetic-Resonance-Based Deep Learning Techniques
Alessandro Taloni1, Francis Allen Farrelly1, Giuseppe Pontillo2,3
1Institute for Complex Systems, National Research Council (ISC-CNR), 00185 Rome, Italy.
International Journal of Molecular Sciences
|September 23, 2022
Summary
Predicting multiple sclerosis (MS) disability progression is possible using specific brain regions identified via 3D T1-weighted MRI scans and deep learning models. These findings may help forecast disease advancement in MS patients.
Area of Science:
- Neuroimaging
- Artificial Intelligence in Medicine
- Neurology
Background:
- Multiple sclerosis (MS) is a chronic neurological disease characterized by unpredictable disability progression.
- Early prediction of disability progression in MS is crucial for timely intervention and patient management.
- Current prediction methods often lack precision, highlighting the need for advanced analytical techniques.
Purpose of the Study:
- To investigate the potential of deep learning models applied to 3D T1-weighted MRI scans for predicting short-term disability progression in MS patients.
- To identify specific brain regions informative for predicting MS disability progression using MRI data.
- To evaluate the performance of fine-tuned ResNet50 models in forecasting disease advancement.
Main Methods:
- One hundred eighty-one MS patients underwent 3T-MRI scans at baseline and were followed for 2-6 years.
- 3D T1-weighted MRI data were processed, including bias correction, brain extraction, and spatial registration.
- Deep learning models (ResNet50 adaptations) were trained and validated on image slices to predict disability progression (EDSS increment).
Main Results:
- 34% of patients experienced disability progression during the follow-up period.
- Specific coronal, sagittal, and axial MRI slices, particularly in frontal grey matter areas, were identified as informative for prediction.
- Classifiers demonstrated significant predictive performance, with AUC values exceeding 0.72 for coronal, 0.81 for sagittal, and 0.69 for axial slices.
Conclusions:
- 3D T1-weighted MRI data, analyzed with deep learning, can predict short-term disability progression in MS patients.
- Specific frontal brain regions contain crucial information for forecasting MS disease advancement.
- This approach offers a promising non-invasive tool for predicting MS disability progression and guiding clinical decisions.
