학술논문

Texture image analysis of normal appearing white matter areas in Clinically Isolated Syndrome that evolved in demyelinating lesions in subsequent MRI scans: Multiple sclerosis disease evolution
Document Type
Conference
Source
Proceedings of the 10th IEEE International Conference on Information Technology and Applications in Biomedicine Information Technology and Applications in Biomedicine (ITAB), 2010 10th IEEE International Conference on. :1-5 Nov, 2010
Subject
Bioengineering
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Signal Processing and Analysis
Robotics and Control Systems
Magnetic resonance imaging
Entropy
Correlation
Variable speed drives
Manuals
Lesions
Language
ISSN
2168-2194
2168-2208
Abstract
The purpose of this paper is to explore the use of texture features analysis for evaluating normal appearing white matter (NAWM) areas in brain Magnetic Resonance Imaging (MRI) scans of patients with a first demyelinating event (Clinically Isolated Syndrome-CIS) that have been subsequently converted (in follow up MRI scans) in demyelinating plaques. Texture features analysis was carried out in NAWM areas selected from transverse sections of T2-weighted MRI scans acquired from 10 untreated subjects that had developed their first neurological-demyelinating event (CIS). All NAWM areas and demyelinating lesions were manually segmented by an experienced multiple sclerosis (MS) neurologist. The patients were twice scanned with brain MRI with an interval of 6–12 months (time 0, initial MRI and time 6–12, repeated MRI scan). Areas in repeated MRI scan, where new lesions had been developed, were mapped back to their corresponding initial MRI scan NAWM areas. Following that, texture features were extracted from those NAWM areas that evolved to demyelinating lesions and compared to texture features extracted from NAWM areas of the initial MRI scan that didn't evolve to demyelinating lesions. It was shown that for the texture features investigated, no single feature could be used to differentiate between NAWM at 0 months (NAWM_C_0), and ROIS that will evolve into demyalating lesions also at 0 months (ROIS_0). Several features could be used to differentiate between ROIS_0 vs Lesions at 0 months, NAWM at 6 months and Lesions at 6 months, as well as between Lesions at 0 months vs Lesions at 6 months. Future work will investigate the proposed method on more subjects.