Age-dependent texture features in skeletal muscle ultrasonography

Area of Science:

• Medical imaging analysis
• Biomedical engineering
• Skeletal muscle physiology

Background:

• Texture analysis quantifies image texture to infer underlying structures, particularly in medical imaging like skeletal muscles.
• Its application in diagnosing neuromuscular diseases and monitoring progression is limited by a lack of age-related data.

Purpose of the Study:

• To investigate the age-dependency of texture features in the medial gastrocnemius muscle of healthy individuals.
• To determine which texture analysis parameters are correlated with aging.

Main Methods:

• Performed texture analysis on medial gastrocnemius muscle images from healthy individuals aged 20s to 80s.
• Evaluated 283 texture features across 6 classes, including histogram, co-occurrence matrix, absolute gradient, wavelet, run-length matrix, and autoregressive models.

Main Results:

• 17-40% of histogram, co-occurrence matrix, absolute gradient, and wavelet features showed significant correlation with age.
• No significant age correlation was found for run-length matrix or autoregressive model features.
• Age-dependency was evident in a substantial portion of the analyzed texture features.

Conclusions:

• Age-dependency must be considered when interpreting the clinical significance of many texture features in skeletal muscle.
• Texture features derived from run-length matrix and autoregressive models may offer independent clinical utility, unaffected by age.