Quantitative neuromuscular ultrasound analysis as biomarkers in amyotrophic lateral sclerosis

Abstract

Objectives: To assess the differences in morphological and texture parameters of median nerve (MN) and abductor pollicis brevis (APB) between amyotrophic lateral sclerosis (ALS) patients and controls.

Methods: The cross-sectional area (CSA) of the MN and the muscle thickness (MTh) of APB were measured bilaterally in 59 recently diagnosed ALS patients and 20 matched healthy controls. Echointensity (EI), echovariation (EV) and grey-level co-occurrence matrix (GLCM) texture features of both structures were also analysed. Correlations between these parameters and clinical variables (muscle strength and disability) were analysed.

Results: The CSA of MN was significantly lower in ALS patients (MD = - 1.83 mm2 [95% CI = 2.89; - 0.77 mm2]; p = 0.01). ALS patients showed significantly lower MTh (- 2.23 mm [3.16; - 1.30 mm]; p < 0.001) and EV (- 7.40 [11.5; - 3.33]; p = 0.004) and higher EI (21.2 [11.9; 30.6]; p < 0.001) in the APB muscle. No relevant differences were detected in GLCM features for this muscle. The model including all parameters (CSA for MN and MTh, EI and EV for APB) showed an AUC of 82% (sensitivity 87%; specificity 42%). Muscle strength and disability correlated with APB muscle ultrasound parameters but not with those of the MN.

Conclusions: APB muscle ultrasound biomarkers (especially MTh and EI) showed better discrimination capacity and correlation with clinical variables than MN biomarkers. However, the combination of both biomarkers increased their ability to detect LMN impairment, suggesting that both biomarkers could be used in a complementary manner for the diagnosis and progression monitoring in ALS.