Image processing and analysis of muscle of frail elderly using MRI

• Main Author: Lip, Kenny Kang Ming
• Other Authors: Poh Chueh Loo
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Bioengineering
• Online Access: http://hdl.handle.net/10356/61543
• Institution: Nanyang Technological University

Sarcopenia is a syndrome characterised by progressive and generalised loss of skeletal muscle mass and strength, possibly resulting in a variety of adverse outcomes such as physical disability and poorer quality of life. Sarcopenia has always been highly associated to decreased muscle mass. However, defining sarcopenia only in terms of muscle mass is too restrictive and may be of limited clinical value. Studies have shown that various new indicators could correlate strength and function, such as fat infiltrates and inter-muscular fat. The aim of this project was to perform texture analysis using MRI images of muscles so as to identify parameters that could differentiate between healthy and frail subjects. To this end, texture parameters such as mean, variance, kurtosis, skewness, entropy, and grey level non-uniformity were derived and analysed. Texture analysis was conducted using results derived previously from volumetric quantification of muscle and fat, which was also part of a two-tier process aimed at differentiating healthy and frail subjects though analysis of muscles through MRI images, the other being texture analysis itself. Outliers were also identified and analysed to identify patterns that demonstrate early signs of frailty. In doing so, Sarcopenia could be detected earlier and controlled in its initial stages. Finally, relationships between volumetric quantification and texture analysis were drawn to assess viable connections that could simplify the entire two-tiered analysis in the future.