Classification of Tissues for Detecting An Inflammatory Disease in Brain MRI

Detection of an inflammatory disease on magnetic resonance image (MRI) is most important as a disease activity and surgical purpose in medical field. We propose an approach to provide tissue segmentation while appearing an inflammatory disease. The two stage of classification process uses in this method 1)a Bayesian classifier that performs a brain tissue classification  at each voxel of reference and follow-up scans using intensities and intensity differences, and 2) a random forest based lesion-level classifier provides a identification of an inflammatory diseases. The method is evaluated on sequential brain MRI of 160 subjects from a separate multi-center clinical trial. The proposed method is compared to the manual identification and gives better performance, sensitivity with fault detecting rate than manual identification. For new lesions greater than 0.15 cc in size, the classifier has near perfect performance (99% sensitivity, 2% false detection rate), as compared to ground truth. The proposed method was also shown to exceed the performance of any one of the nine expert manual identifications.