Automated segmentation of MS lesions in MR

Objective Quantitative analysis of MR images is becoming increasingly important in assessing the progression of multiple sclerosis (MS) and in monitoring the effect of a drug therapy. In clinical trials, manual analysis of the MR images by human experts is prohibitively time-consuming because of the large amounts of data typically involved. The interand mtra-observer variability associated with manual delineations. and the unclear decision making process when multi-spectral MR data are presented to a human expert. further complicate matters. The goal of our work is to develop a fast, fully-automated method for MS lesion segmentation that analysis multi-spectral MR data in an objective and reproducible way.