DX47
The Topographical Model of Multiple Sclerosis: A New Visualization of Disease Course
Objectives: To develop a new unified MS clinical course model.
Methods: The model design encapsulates five factors: topographical distribution of lesions and relapses they cause; relapse frequency, severity, and recovery; and progression rate. The model visually depicts the interplay of factors utilizing 3D animated renderings.The CNS is represented as a pool, with a shallow end and a deep end. Spinal cord and optic nerves occupy the shallow end, the posterior fossa comprises the middle, and the cerebral hemispheres constitute the deep end, each with increasing amounts of functional reserve. MS lesions rise up as topographical peaks from the pool floor. The water's surface depicts the clinical threshold: lesions below the surface are clinically silent; those that cross it cause clinical relapses. Progression is depicted as the slowly dropping water level, representing depletion of functional capacity. Thus, like symptom recrudescence in Uhthoff's phenomenon and pseudoexacerbations, progression clinically recapitulates the form of prior relapses/lesions, incrementally revealing above the surface the underlying lesion topography.
Results: A dynamic, 3D-rendered visualization of the topographical model will be shown. Clinical and MRI correlates of the depicted factors will be elucidated.
Conclusions: This model depicts MS disease course in a clinically nuanced way that complements the clinical course categories. It illuminates several well-described but poorly reconciled phenomena, including the clinical/MRI paradox, impact of MRI lesion burden on late (but not early) disease outcomes, and prognostic importance of brainstem and cord lesions. The model is congruent with emerging data on brain atrophy, and the protective effect of large baseline brain volume. The topographical model can be used as an educational tool for patients and professionals, and predictive implications can be empirically tested by application to clinical trial datasets.