Related Experiment Video
Updated: Sep 29, 2025
Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
Published on: December 15, 2023
Advanced neural networks for classification of MRI in psoriatic arthritis, seronegative, and seropositive rheumatoid
Lukas Folle1, Sara Bayat2,3, Arnd Kleyer2,3
1Pattern Recognition Lab, Friedrich-Alexander-University Erlangen-Nürnberg.
Objectives:
To evaluate whether neural networks can distinguish between seropositive RA, seronegative RA, and PsA based on inflammatory patterns from hand MRIs and to test how psoriasis patients with subclinical inflammation fit into such patterns.
Methods:
ResNet neural networks were utilized to compare seropositive RA vs PsA, seronegative RA vs PsA, and seropositive vs seronegative RA with respect to hand MRI data. Results from T1 coronal, T2 coronal, T1 coronal and axial fat-suppressed contrast-enhanced (CE), and T2 fat-suppressed axial sequences were used. The performance of such trained networks was analysed by the area under the receiver operating characteristics curve (AUROC) with and without presentation of demographic and clinical parameters. Additionally, the trained networks were applied to psoriasis patients without clinical arthritis.
Results:
MRI scans from 649 patients (135 seronegative RA, 190 seropositive RA, 177 PsA, 147 psoriasis) were fed into ResNet neural networks. The AUROC was 75% for seropositive RA vs PsA, 74% for seronegative RA vs PsA, and 67% for seropositive vs seronegative RA. All MRI sequences were relevant for classification, however, when deleting contrast agent-based sequences the loss of performance was only marginal. The addition of demographic and clinical data to the networks did not provide significant improvements for classification. Psoriasis patients were mostly assigned to PsA by the neural networks, suggesting that a PsA-like MRI pattern may be present early in the course of psoriatic disease.
Conclusion:
Neural networks can be successfully trained to distinguish MRI inflammation related to seropositive RA, seronegative RA, and PsA.
Insights
Neural networks can differentiate inflammatory patterns in hand MRIs for rheumatoid arthritis (RA) and psoriatic arthritis (PsA). Psoriasis patients with subclinical inflammation were often classified as PsA, indicating early disease patterns.
Area of Science:
- Radiology
- Artificial Intelligence
- Rheumatology
Background:
- Distinguishing between rheumatoid arthritis (RA) and psoriatic arthritis (PsA) is crucial for effective treatment.
- Hand MRI inflammatory patterns offer potential biomarkers for differentiating these conditions.
- Subclinical inflammation in psoriasis patients may share features with PsA.
Purpose of the Study:
- To evaluate the capability of neural networks in distinguishing seropositive RA, seronegative RA, and PsA using hand MRI data.
- To assess the classification of psoriasis patients with subclinical inflammation within these patterns.
Main Methods:
- ResNet neural networks were trained on hand MRI data from RA and PsA patients.
- Various MRI sequences (T1, T2, contrast-enhanced) were utilized for classification.
- Network performance was evaluated using AUROC, with and without demographic/clinical data, and applied to psoriasis patients.
Main Results:
- Neural networks achieved 75% AUROC for seropositive RA vs. PsA, 74% for seronegative RA vs. PsA, and 67% for seropositive vs. seronegative RA.
- Contrast-enhanced sequences were relevant but not critical; demographic/clinical data offered minimal improvement.
- Psoriasis patients were predominantly classified as PsA, suggesting early PsA-like MRI patterns.
Conclusions:
- Neural networks can effectively differentiate MRI-based inflammatory patterns of seropositive RA, seronegative RA, and PsA.
- The findings suggest that psoriasis patients may exhibit early PsA-like inflammatory patterns on hand MRIs.
