A joint project of the Graduate School, Peabody College, and the Jean & Alexander Heard Library

Title page for ETD etd-08102017-110606


Type of Document Master's Thesis
Author Xie, Yuan
Author's Email Address yuan.xie@vanderbilt.edu
URN etd-08102017-110606
Title Quantitative texture analysis of T2- weighted MR images in polymyositis and dermatomyositis patients
Degree Master of Science
Department Biomedical Engineering
Advisory Committee
Advisor Name Title
Bruce M. Damon Committee Chair
E. Brian Welch Committee Chair
Keywords
  • Dermatomyositis
  • Polymyositis
  • muscle imaging
  • T2 weighted MR images
  • Texture analysis
Date of Defense 2017-04-19
Availability unrestricted
Abstract
Dermatomyositis (DM) and polymyositis (PM) patients experience intramuscular inflammation and necrosis, eventually progressing to fat infiltration. The gold standard for MRI assessment of fat tissue infiltration is quantitative fat-water MRI. Fat tissue is also detectable using standard contrast-based clinical MRI sequences; however, typical analyses of these data are qualitative. Texture analysis is a quantitative method for analyzing signal variations in contrast-based images. The goals of this study were to determine which MRI and tissue parameters explain variations in texture parameters and to use texture analysis of contrast-based MR images to predict the fat fraction (Ffat), as determined by quantitative fat-water MRI. Fat signal-suppressed (FS) T1 and T2 maps, Ffat maps, and T2-weighted MR images were acquired from 5 DM patients, 8 PM patients, and 13 control subjects. Images were acquired at mid-thigh. The Grey Level Co-occurrence Matrix (GLCM) and Grey Level Run-length Matrix (GLRM) were calculated and used to derive 11 texture features. Regression analysis focused on the log(Energy) parameter, derived from the GLCM, and the High Gray-level Run-length Emphasis (HGRE), derived from the GLRM. 57.4% of the variance in log(Energy) was explained by Ffat variations. For HGRE, 68.6% of its variance was explained by Ffat variations. Finally, using HGRE, Low Gray level run emphasis, and Homogeneity as predictors, we were able to explain 70.3% of the variance in Ffat. These data show that HGRE primarily reflects fat tissue infiltration. Also, texture analysis can be used to predict Ffat from T2-weighted clinical MR images.
Files
  Filename       Size       Approximate Download Time (Hours:Minutes:Seconds) 
 
 28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access 
  XieY.pdf 511.49 Kb 00:02:22 00:01:13 00:01:03 00:00:31 00:00:02

Browse All Available ETDs by ( Author | Department )

If you have more questions or technical problems, please Contact LITS.