To compare the diagnostic accuracy and predictive value of the 2023 radiologically isolated syndrome (RIS) and the 2024 McDonald multiple sclerosis (MS) criteria for the first symptomatic clinical event.

The 2024 McDonald criteria classify individuals with RIS who have dissemination in space or oligoclonal bands (OCB) to be diagnosed with MS. However, whether this revision improves the prediction of the first clinical event compared with the 2023 RIS definition has not been validated.

We analyzed a multicenter cohort of 216 individuals with incidental MRI findings suggestive of demyelination. Each individual was classified according to both the 2023 RIS and 2024 McDonald criteria. We calculated the time to first clinical event, sensitivity, specificity, accuracy, and predictive values for each criterion. Covariates included age at index MRI, OCB status, spinal cord and corpus callosum lesion counts, and new T2 lesions on follow-up imaging.

The 2023 RIS criteria were fulfilled by 63.4% of the cohort, whereas 57.9% fulfilled the 2024 McDonald criteria; 27.3% did not meet either. Overall, 50% of RIS (2023) had their first clinical event within a median of 5.8 years (95% CI 3.0–8.7), compared to 4.1 (95% CI 2.7–5.8)  years for those meeting the 2024 criteria. In Cox regression analysis, hazard ratios were 3.6 (p = 0.002) for the 2023 RIS and 4.9 (p < 0.001) for the 2024 MS group. The 2024 criteria improved specificity (0.53 vs 0.45) and overall accuracy (0.62 vs 0.56) while maintaining sensitivity (0.89 vs 0.87).

The 2024 McDonald criteria demonstrate a higher diagnostic precision and predictive accuracy for early clinical conversion compared to the 2023 RIS definition. Integrating OCB positivity with dissemination in space identifies a clinically meaningful subset of individuals with RIS who may benefit from early initiation of disease-modifying therapy.