Characterization of Multiple Sclerosis on the Brain Magnetic Resonance Images Using Texture Analysis

Abstract

Multiple Sclerosis (MS) is the most common chronic autoimmune demyelinating inflammatory disease of the central nervous system, which can be diagnosed by magnetic resonance imaging (MRI) by evidence of multiple patches Wight of scar tissue in different parts of the central nervous system on T1 weighted images, T2 weighted image and FLAIR. Texture analysis evaluates interpixel relationships that generate characteristic organizational patterns in an image, many of which are beyond the ability of visual perception. The aim of this study was to characterize MS plaques in MR images using Texture analysis which facilitate pattern recognition that might not visible to human eye. This study is an analytical study, which conducted at Modern medical Centre and Omdurman military hospital in a period from December 2015 to March 2018. The sample of this study was consisted from 200 MR brain (T1, T2 and FLAIR) images selected conveniently from patient with MS. Computer based software Interactive Data Language (IDL) and stepwise linear discriminant analysis were used to generate a classification score and to select the most discriminant features that can be used in the classification of normal and abnormal brain tissues. The results reveal that the MS areas were very different from the CSF, bones, white matter and gray matter. However, plaques can be identified and classified using textural analysis with high sensitivity of 90.9% for first order statistics and 96.9% using higher order statistics. In conclusion, textural feature can be used with some confidences to pin point the areas of MS in MRI brain images. Generation of image processing unit in each hospital is recommended to decrease the misdetection rate.