Background The assessment of muscle mass is a key determinant of the diagnosis of sarcopenia. We introduce for the first time an ultrasound imaging method for diagnosing sarcopenia based on changes in muscle geometric proportions. Methods Vastus lateralis muscle fascicle length (Lf) and thickness (Tm) were measured at 35% distal femur length by ultrasonography in a population of 279 individuals classified as moderately active elderly (MAE), sedentary elderly (SE) (n = 109), mobility impaired elderly (MIE) (n = 43), and in adult young controls (YC) (n = 60). The ratio of Lf/Tm was calculated to obtain an ultrasound index of the loss of muscle mass associated with sarcopenia (USI). In a subsample of elderly male individuals (n = 76) in which corresponding DXA measurements were available (MAE, n = 52 and SE, n = 24), DXA-derived skeletal muscle index (SMI, appendicular limb mass/height2) was compared with corresponding USI values. Results For both young and older participants, USI values were found to be independent of sex, height and body mass. USI values were 3.70 ± 0.52 for YC, 4.50 ± 0.72 for the MAE, 5.05 ± 1.11 for the SE and 6.31 ± 1.38 for the MIE, all significantly different between each other (P < 0.0001). Based on the USI Z-scores, with reference to the YC population, the 219 elderly participants were stratified according to their muscle sarcopenic status. Individuals with USI values within a range of 3.70 < USI ≥ 4.23 were classified as non-sarcopenic (prevalence 23.7%), those with USI values within 4.23 < USI ≥ 4.76 were classified as pre-sarcopenic (prevalence 23.7%), those with USI values within 4.76 < USI ≥ 5.29 were classified as moderately sarcopenic (prevalence 15.1%), those with USI values within range 5.29 < USI ≥ 5.82 were classified as sarcopenic (prevalence 27.9%), and those with USI values >5.82 were classified as severely sarcopenic (prevalence 9.6%). The DXA-derived SMI was found to be significantly correlated with USI (r = 0.61, P < 0.0001). Notably, the USI cut-off value for moderate sarcopenia (4.76 a.u.) was found to coincide with the DXA cut-off value of sarcopenia (7.26 kg/m2). Conclusions We propose a novel, practical, and inexpensive imaging marker of the loss of muscle mass associated with sarcopenia, called the ultrasound sarcopenic index (USI), based on changes in muscle geometric proportions. These changes provide a useful ‘signature of sarcopenia’ and allow the stratification of individuals according to the presence and severity of muscle sarcopenia. We are convinced that the USI will be a useful clinical tool for confirming the diagnosis of sarcopenia, of which the assessment of muscle mass is a key-component.