Texture in Quantitative Viscoelastic Response (QVisR) Images Differentiates Dystrophic from Control Skeletal Muscles in Boys, In Vivo

Quantitative Viscoelastic Response (QVisR) ultrasound is a new machine learning-based elasticity imaging method in which elastic and viscous moduli are estimated from tissue deformation in response to two consecutive acoustic radiation force (ARF) excitations. The estimated moduli are rendered into two-dimensional parametric images of elastic and viscous modulus. From QVisR images, the spatial distribution of elastic and viscous properties may be evaluated using established computational texture analysis tools. In this study, computational texture analysis by the omnidirectional gray-level run-length matrix (GLRLM) was applied to QVisR images. The images were obtained in the vastus lateralis (VL) muscles of 11 boys with Duchenne muscular dystrophy, aged 5-12 years, and 8 age-matched boys with no known neuromuscular disorders, who served as controls. GLRLM-derived elastic entropy in QVisR images was statistically higher (Wilcoxin, p<; 0.05) in the VL muscles of boys with DMD than control for ages between 5.5 and 7 years. This result is consistent with expected heterogeneous distribution of inflammation, necrosis, fibrosis, and fat in early stages of dystrophic degeneration. The findings suggest that texture in QVisR images may be a relevant biomarker for DMD progression and response to treatment, particularly at young ages when interventions are likely to be most impactful.