Dominik S. Meier1, Charles R.G. Guttmann1
1Radiology, Brigham & Women's Hospital - Harvard Medical School, Boston, MA, USA
Global brain parenchymal atrophy, assessed from structural brain MRI, has become established as reliable metric of progression in many neurodegenerative diseases. Uncertainty remains to what extent this measure expresses nonlinearity over time and to what extent current methods are sensitive to capture nonlinear behavior. We tested nonlinearities in BPF change in multiple sclerosis (MS) over time and the impact on predicting disease progression. Linear models tended to overestimate atrophy at 3-5 year follow-up. Nonlinear models achieved significantly better fits. A greater atrophy rate in the first year was associated with a higher number of attacks.