Related Experiment Video
Updated: Jun 6, 2026
Magnetic Resonance Imaging of Multiple Sclerosis at 7.0 Tesla
Published on: February 19, 2021
Intracortical lesions: relevance for new MRI diagnostic criteria for multiple sclerosis
M Filippi1, M A Rocca, M Calabrese
1Neuroimaging Research Unit, Institute of Experimental Neurology, Division of Neuroscience, Scientific Institute and University Hospital San Raffaele, via Olgettina, 60, 20132 Milan, Italy. filippi.massimo@hsr.it
Objective:
To generate and validate new MRI diagnostic criteria for multiple sclerosis (MS) taking into account not only white matter lesions but also intracortical lesions (ICLs).
Methods:
Brain double inversion recovery and brain and cord T2- and postcontrast T1-weighted scans were acquired in a training (80 patients with clinically isolated syndromes [CIS], median follow-up = 55.3 months) and a validation (39 patients with CIS, median follow-up = 28.0 months) sample. In the training sample, regression analysis and Cox proportional hazard model were used to identify MRI variables independently predicting the evolution to clinically definite (CD) MS. The best criterion selected was then validated. The performance of the new and previously available MRI criteria for disease dissemination in space (DIS) and time (DIT) were tested.
Results:
The final multivariate model showed that ≥1 ICL (p < 0.001), ≥1 infratentorial (p = 0.03), and ≥ 1 gadolinium-enhancing or ≥1 spinal cord lesion (p = 0.004) were independent predictors of CDMS. The presence of at least 2 of these variables was the best DIS criterion in both samples. New ICLs had a poor sensitivity for DIT. The combination of the new DIS criterion with the MAGNIMS criteria for DIT yielded to an accuracy of 81%, which was higher than those of the other available criteria.
Conclusions:
The accuracy of MRI diagnostic criteria for MS is increased when considering the presence of ICLs on baseline scans from patients at presentation with CIS suggestive of MS. If confirmed by other studies, ICL detection might be considered in future diagnostic algorithms for MS.
