A mixture model for differentiating longitudinal courses of multiple sclerosis.

Area of Science:

• Neurology
• Clinical Research
• Data Science

Background:

• Multiple sclerosis (MS) presents with diverse clinical courses, necessitating early identification of high-risk patients.
• Predicting disability accumulation is crucial for managing MS progression.

Purpose of the Study:

• To identify distinct patient subgroups in MS using a mixture model.
• To predict the long-term clinical trajectories of these identified subgroups.

Main Methods:

• Retrospective analysis of 629 MS patients from 1994-2019.
• Latent profile and transition analyses were performed using eight prognostic factors.
• Key parameters included age, disease duration, EDSS, relapses, and specific lesion/symptom types.

Main Results:

• Three distinct patient groups were identified: No Disability Regardless Disease Duration (NDRDD), Disability within a Short Disease Duration (DSDD), and Disability within a Long Disease Duration (DLDD).
• Initial class distinction was strongest for EDSS, disease duration, and absence of sphincter impairment.
• A significant probability (52.1%) exists for patients in the NDRDD group to transition to the DLDD group over time.

Conclusions:

• Three distinct clinical presentation groups and their temporal evolution in MS were identified.
• The primary observed transition is from the NDRDD group to the DLDD group, highlighting a potential progression pathway.