Our overall goal is to develop next generation MRI markers of clinical disease progression to be used as primary outcome measures in phase 2 trials for progressive MS (PMS) in a fashion analogous to the use of Gadolinium-enhancement in relapsing MS trials. The underlying hypothesis is that disease progression in MS is detectable through MRI prior to its clinical expression, and that we will be able to infer measures or features of the MRI that may not be directly observable and that have the following characteristics:
i) sensitive to pathological changes over short time intervals consistent with efficient phase 2 trials for progressive MS (face validity),
ii) correlate with clinical disease progression over the same time interval (concurrent validity), and predictive of future disability (predictive validity)
iii) predictive of treatment effect on clinical outcomes.
To identify MRI patterns with the above characteristics, we will further develop and use: i) statistical approaches including algorithms to identify combinations of MRI features that maximize the change over time estimated by mixed models, and also ii) cutting-edge computer science tools, algorithms from the fields of computer vision/pattern recognition, and machine learning (including deep learning) techniques.