Tampere University of Technology

TUTCRIS Research Portal

MR image texture in Parkinson's disease: A longitudinal study

Research output: Contribution to journalArticleScientificpeer-review

Details

Original languageEnglish
Pages (from-to)97-104
Number of pages8
JournalActa Radiologica
Volume56
Issue number1
DOIs
Publication statusPublished - 2015
Publication typeA1 Journal article-refereed

Abstract

Background: Few of the structural changes caused by Parkinson's disease (PD) are visible in magnetic resonance imaging (MRI) with visual inspection but there is a need for a method capable of observing the changes beyond the human eye. Texture analysis offers a technique that enables the quantification of the image gray-level patterns. Purpose: To investigate the value of quantitative image texture analysis method in diagnosis and follow-up of PD patients. Material and Methods: Twenty-six PD patients underwent MRI at baseline and after 2 years of follow-up. Four co-occurrence matrix-based texture parameters, describing the image homogeneity and complexity, were calculated within clinically interesting areas of the brain. In addition, correlations with clinical characteristics (Unified Parkinson's Disease Ranking Scales I-III and Mini-Mental State Examination score) along with a comparison to healthy controls were evaluated. Results: Patients at baseline and healthy volunteers differed in their brain MR image textures mostly in the areas of substantia nigra pars compacta, dentate nucleus, and basilar pons. During the 2-year follow-up of the patients, textural differences appeared mainly in thalamus and corona radiata. Texture parameters in all the above mentioned areas were also found to be significantly related to clinical scores describing the severity of PD. Conclusion: Texture analysis offers a quantitative method for detecting structural changes in brain MR images. However, the protocol and repeatability of the method must be enhanced before possible clinical use.

Keywords

  • Follow-up, Image analysis, Magnetic resonance imaging (MRI), Parkinson's disease, Texture analysis

Publication forum classification

Field of science, Statistics Finland