Objective: We assessed changes in quantitative muscle ultrasound data in boys with Duchenne muscular dystrophy (DMD) and healthy controls to determine whether ultrasound can serve as a biomarker of disease progression. Two approaches were used: gray scale level (GSL), measured from the ultrasound image, and quantitative backscatter analysis (QBA), measured directly from the received echoes. Methods: GSL and QBA were obtained from 6 unilateral arm/leg muscles in 36 boys with DMD and 28 healthy boys (age = 2–14 years) for up to 2 years. We used a linear mixed effects model with random intercept and slope terms to compare trajectories of GSL, QBA, and functional assessments. We analyzed separately a subset of boys who initiated corticosteroids. Results: Compared to healthy boys, increasing GSL in DMD boys >7.0 years old was first identified at 6 months (eg, anterior forearm slope difference of 1.16 arbitrary units/mo, p = 0.004, 95% confidence interval [CI] = 0.38–1.94); in boys ≤ 7 years old, differences in GSL first appeared at 12 months (0.82 arbitrary units/mo, p = 0.04, 95% CI = 0.075–1.565, in rectus femoris). QBA performed similarly to GSL (eg, DMD boys > 7 years old: 0.41dB/mo, p = 0.01, 95% CI = 0.096–0.72, in anterior forearm at 6 months). Ultrasound identified differences earlier than functional measures including 6-minute walk and supine-to-stand tests. However, neither QBA nor GSL showed an effect of corticosteroid initiation. Interpretation: QBA performs similarly to GSL, and both appear more sensitive than functional assessments for detecting muscle deterioration in DMD. Additional studies will be required to determine whether quantitative muscle ultrasound can detect therapeutic efficacy. Ann Neurol 2017;81:633–640.