The Diagnostic Utility of Automated Thresholding Ultrasound Analysis in Adults With Hereditary Neuromuscular Disorders.

INTRODUCTION/AIMS: The diagnosis of neuromuscular diseases in adults can be challenging and novel ultrasound methods could facilitate the process. This study investigated the diagnostic value of automated thresholding in identifying hereditary neuromuscular disorders (HNMDs) in adults and in distinguishing between neurogenic and myopathic processes.

METHODS

Thirty-one patients with neurogenic HNMD, 37 with myopathic HNMD, and 68 healthy controls underwent ultrasound examination based on a four-muscle scanning protocol, including biceps brachii, flexor carpi radialis, rectus femoris, and tibialis anterior. The hyperechoic fraction of Otsu (Otsu-HF) and Triangle (Triangle-HF) thresholding algorithms, grayscale value (GSV), and echovariation (EV) were measured. Global estimates were calculated by averaging individual muscle values. Diagnostic accuracy of echogenicity measures and correlations with muscle strength were investigated.

RESULTS

Global and individual Otsu-HF, GSV and EV parameters showed high diagnostic accuracies in identifying HNMDs. All areas under receiver operating characteristic curve were > 0.9 for global parameters and > 0.8 for individual muscle estimates. Global cut-offs of the suggested protocol showed a specificity of 88.2%-98.5% and a sensitivity of 83.8%-89.7%. Furthermore, global Otsu-HF differentiated echogenicity between neurogenic and myopathic HNMDs, yielding a higher accuracy than grayscale analysis (71.6% vs. 63.1%, p = 0.0065). Significant relationships were shown between muscle strength, global and individual Otsu-HF, GSV, and EV.

DISCUSSION: Our quantitative muscle ultrasound (QMUS) screening protocol was shown to be a useful tool for assessing adult subjects with suspected HNMDs. In the era of next-generation sequencing, QMUS could be a valuable neurophysiological biomarker in the initial diagnostic workflow, guiding targeted further testing.