Gadolinium-enhanced magnetic resonance imaging (MRI) and measures of brain volume have been extensively applied in large-scale studies to assess disease activity and irreversible tissue damage in multiple sclerosis (MS). Although histopathological studies of MS demonstrated that axonal transection occurs at sites of inflammatory changes, the correlation between brain tissue loss and gadolinium enhancement was found to be either absent or poor in virtually all in vivo MRI studies. This review discusses the reasons of this "inflammation/neurodegeneration mismatch" in MS and proposes possible strategies for a better in vivo characterization of the complex pathological process of this disease.