Multiple Sclerosis (MS) is a complex neurodegenerative disease characterized by heterogeneous progression patterns. Traditional clinical measures like the Expanded Disability Status Scale (EDSS) inadequately capture the full spectrum of disease progression, highlighting the need for advanced Disease Progression Modeling (DPM) approaches. This study harnesses cutting-edge neuroimaging and deep learning techniques to investigate deviations in subcortical volumes in MS patients. We analyze T1-weighted and Fluid-attenuated inversion recovery (FLAIR) Magnetic Resonance Imaging (MRI) data using advanced DL segmentation models, \(SynthSeg^{+}\) and SynthSeg-WMH, which address the challenges of conventional methods in the presence of white matter lesions. By comparing subcortical volumes of 326 MS patients to a normative model from 37,407 healthy individuals, we identify significant deviations that enhance our understanding of MS progression. This study highlights the potential of integrating DL with normative modeling to refine MS progression characterization, automate informative MRI contrasts, and contribute to data-driven DPM in neurodegenerative diseases.