Quantitative Muscle Ultrasonography Using Textural Analysis in Amyotrophic Lateral Sclerosis

Abstract

The purpose of this study was to analyze differences in gray-level co-occurrence matrix (GLCM) parameters, as assessed by muscle ultrasound (MUS), between amyotrophic lateral sclerosis (ALS) patients and healthy controls, and to compare the diagnostic accuracy of these GLCM parameters with first-order MUS parameters (echointensity [EI], echovariation [EV], and muscle thickness [MTh]) in different muscle groups. Twenty-six patients with ALS and 26 healthy subjects underwent bilateral and transverse ultrasound of the biceps/brachialis, forearm flexor, quadriceps femoris, and tibialis anterior muscle groups. MTh was measured with electronic calipers, and EI, EV, and GLCM were obtained using Image J (v.1.48) software. Sensitivity, specificity, likelihood ratios, and area under the curve (AUC) were performed by logistic regression models and receiver operating characteristic curves. GLCM parameters showed reduced granularity in the muscles of ALS patients compared with the controls. Regarding the discrimination capacity, the best single diagnostic parameter in forearm flexors and quadriceps was GLCM and in biceps brachialis and tibialis anterior was EV. The respective combination of these two parameters with MTh resulted in the best AUC (over 90% in all muscle groups and close to the maximum combination model). The use of new textural parameters (EV and GLCM) combined with usual quantitative MUS variables is a promising biomarker in ALS.

Ultrasound in the Diagnosis and Monitoring of Amyotrophic Lateral Sclerosis: A Review

“…Martínez‐Payá et al . (2017) also explored muscle textural analysis using the gray‐level co‐occurrence matrix (GLCM) . They used the same patients and data set as in their previously mentioned study, applying a new analysis to the images.…”

“…Martínez-Payá et al (2017) also explored muscle textural analysis using the gray-level co-occurrence matrix (GLCM). 32 They used the same patients and data set as in their previously mentioned study, applying a new analysis to the images. GLCM parameters were reduced in patients with ALS, but the authors concluded that combining this measure with echovariance and muscle thickness provided the best discrimination between ALS and control groups.…”

Age‐dependent texture features in skeletal muscle ultrasonography

“…Given the recent advent of radiomics that utilizes quantitative imaging features including texture features in oncology and related fields (5,13), there would be little doubt that this methodology of quantitative imaging analysis can be applicable to neuromuscular ultrasound. Indeed, there have been several studies to assess the texture characteristics of skeletal muscles ( 6,12,14,8,(15)(16)(17). Among these, texture analysis of muscle ultrasound was reported to be useful in identifying the underlying etiology (i.e., myogenic vs. neurogenic in nature) (6) and in being potentially utilized as a disease biomarker, such as in amyotrophic lateral sclerosis (17).…”

“…Indeed, there have been several studies to assess the texture characteristics of skeletal muscles ( 6,12,14,8,(15)(16)(17). Among these, texture analysis of muscle ultrasound was reported to be useful in identifying the underlying etiology (i.e., myogenic vs. neurogenic in nature) (6) and in being potentially utilized as a disease biomarker, such as in amyotrophic lateral sclerosis (17). However, before reliably using muscle texture features in any clinical settings, these should be carefully assessed in the confounding factors, such as gender and aging.…”

Local Texture Anisotropy as an Estimate of Muscle Quality in Ultrasound Imaging

“…This parameter is highly dependent on the ultrasound device and settings (Pillen and Van Alfen 2015;Zaidman et al 2008). In contrast to mean EI and other first-order descriptors (e.g., standard deviation, skewness, kurtosis and entropy), higher-order texture features (e.g., Haralick features, Galloway features, local binary pattern) have been identified as promising tools for diagnosis, characterization and follow-up of muscle alterations (Konig et al 2015;Martinez-Paya et al 2017;Molinari et al 2015;Sogawa et al 2017;Weng et al 2017). These descriptors may have the advantage of being less affected by variables such as intensity.…”