Purpose To assess the multiple texture features of skeletal muscles in neurogenic and myogenic diseases by using ultrasonography (US). Materials and Methods After institutional review board approval, muscle US studies of the medial head of the gastrocnemius were performed prospectively in patients with neurogenic diseases (n = 25 [18 men]; mean age, 66.0 years ± 12.3 [standard deviation]), in patients with myogenic diseases (n = 21 [12 men]; mean age, 68.3 years ± 11.5), and in healthy control subjects (n = 21 [11 men]; mean age, 70.5 years ± 8.4) between January 2013 and May 2016. Written informed consent was obtained. Muscle texture parameters were obtained, and five algorithms were used to classify the groups. Results The neurogenic and myogenic disease groups showed higher echo intensities than the control subjects. The histogram-derived texture parameters had overlaps between the neurogenic and myogenic groups and thus had a low discrimination rate. With assessment of more classes of texture parameters, three groups were correctly classified (100% correct, according to four of five classification algorithms). Tenfold cross validation showed 93.5%-95.7% correct classification between the neurogenic and myogenic groups. The run-length matrix, autoregressive model, and co-occurrence matrix were particularly useful in distinguishing the neurogenic and myogenic groups. Conclusion Texture analysis of muscle US data can enable differentiation between neurogenic and myogenic diseases and is useful in noninvasively assessing underlying disease mechanisms. RSNA, 2017 Online supplemental material is available for this article.

Texture analysis characterizes regions in an image by their texture content and has been utilized to infer the underlying structures of medical images such as skeletal muscles. Although potentially useful in tissue diagnosis and assessing disease progression of neuromuscular diseases, the use of texture analysis in such purposes are limited, due to lack of information such as effects of aging. Thus, we performed texture analysis of medial gastrocnemius in healthy individuals form their 20s to late 80s. Among the 283 texture features in 6 classes, the features related to histogram, co-occurrence matrix, absolute gradient, and wavelet were correlated to age in 17-40% % of the parameters, while none of the features related to run-length matrix and autoregressive model had significant correlation to age. This study showed that age-dependency in many texture features are present and need to be taken into account in elucidating the clinical significance. By contrast, the features related to runlength matrix and autoregressive model could have clinical utility. J. Med. Invest. 65 : 274-279, August, 2018

Given the recent technological advent of muscle ultrasound (US), classification of various myopathic conditions could be possible, especially by mathematical analysis of muscular fine structure called texture analysis. We prospectively enrolled patients with three neuromuscular conditions and their lower leg US images were quantitatively analyzed by texture analysis and machine learning methodology in the following subjects : Inclusion body myositis (IBM) [N=11] ; myotonic dystrophy type 1 (DM1) [N=19] ; polymyositis/dermatomyositis (PM-DM) [N=21]. Although three-group analysis achieved up to 58.8% accuracy, two-group analysis of IBM plus PM-DM versus DM1 showed 78.4% accuracy. Despite the small number of subjects, texture analysis of muscle US followed by machine learning might be expected to be useful in identifying myopathic conditions. J. Med. Invest. 237Abbreviations DM = dermatomyositis, DM1 = myotonic dystrophy type 1, EI = echointensity, GLCM = gray level co-occurrence matrix, GLNU = graylevel non-uniformity, GLRLM = gray-level run length matrix, GLZLM = gray-level zone length matrix, IBM = inclusion body myositis, NGLDM = neighborhood gray-level different matrix, PM = polymyositis, ROI = region of interest, US = ultrasound

Preferential high echogenicity in the medial gastrocnemius and deep finger flexors is suggestive of DM1. Muscle echogenicity is not generally related to functional dysfunction in DM1.

No abstract