Hemispheric asymmetry measured by texture analysis and diffusion tensor imaging in two multiple sclerosis subtypes

Abstract

Background This paper addresses two subtypes of multiple sclerosis (MS), primary progressive multiple sclerosis (PPMS) and relapsing-remitting multiple sclerosis (RRMS). The separation of PPMS and RRMS is challenging in certain cases. Purpose To quantitatively determine MS subtypes using texture analysis (TA) and diffusion tensor imaging (DTI). Material and Methods T1-weighted (T1W) magnetic resonance imaging (MRI) and DTI of the left and right brain hemispheres of 17 patients with PPMS and 19 patients with RRMS were studied. Areas of the caudate nucleus and thalamus were investigated as normal appearing gray matter (NAGM), and areas of the cerebral peduncle and centrum semiovale were investigated as normal appearing white matter (NAWM). The described locations were symmetrical and were accurately marked. TA was performed on the T1W images, and the fractional anisotropy and apparent diffusion coefficient were determined from the DTI data. Results Hemispherical differences were found with both TA and DTI. Several texture and diffusion tensor parameter values calculated for the left and right hemispheres of the patients showed statistically significant differences. The patients with RRMS had greater significant differences ( P < 0.01) in the thalamus between the hemispheres than did the patients with PPMS. The TA classification accuracy of the PPMS and RRMS subtypes was above 80%. Conclusion TA can be helpful when distinguishing between PPMS and RRMS, while DTI appears to reveal the hemispherical asymmetry of RRMS patients.