Validation of live animal ultrasonic measurements of body composition in market lambs.

Market lambs from the state fair of Virginia (n = 172) were ultrasonically evaluated by 4 scan technicians and 3 image interpreters to determine accuracy of ultrasonic estimates of loin muscle area (ULMA), backfat thickness (UBF), and body wall thickness (UBW). Lambs were initially scanned at the preferred magnification setting of each technician; 2 chose 1.5x and 2 chose 2.0x. Lambs were then scanned a second time for ULMA and UBF with machine magnification settings changed from 1.5 to 2.0x, or vice versa, midway through the second scan. Lambs were then slaughtered, and analogous measurements [carcass loin muscle area, carcass backfat thickness, and carcass body wall thickness (CBW)] were recorded on chilled carcasses. Pooled, residual correlation coefficients within technicians and interpreters between ultrasonic measurements from the first scan and carcass measurements were 0.66 for loin muscle area, 0.78 for backfat thickness, and 0.73 for body wall thickness, but were reduced to 0.43, 0.69, and 0.50, respectively, by inclusion of linear effects of carcass weight in the model. Mean bias for technicians and interpreters ranged from -1.30 to -2.66 cm(2) for loin muscle area, -0.12 to -0.17 cm for backfat thickness, and 0.14 to -0.03 cm for body wall thickness; prediction errors ranged from 1.86 to 2.22 cm(2), 0.12 to 0.14 cm, and 0.35 to 0.38 cm, respectively. Pooled correlations between repeated measures were 0.67 for ULMA, 0.79 for UBF, and 0.68 for UBW at the same magnification and 0.73 for ULMA and 0.76 for UBF across different magnification settings. Mean differences between repeated measures were more variable among technicians and interpreters than statistics comparing ultrasound to carcass measures. Standard errors of repeatability ranged from 1.61 to 2.45 cm(2) for ULMA, 0.07 to 0.11 cm for UBF, and 0.36 to 0.42 cm for UBW. The effect of changing magnification setting on technician and interpreter repeatability was small for UBF and ULMA. The accuracy of prediction of CBW from UBW was similar to that achieved for backfat thickness; further assessment of the value of ultrasonic measurements of body wall thickness in lambs is warranted. These results indicate that ultrasound scanning can reliably predict carcass loin muscle area and backfat thickness in live lambs and, accordingly, has value in selection programs to improve composition. Development of certification standards for US lamb ultrasound technicians based on results of this study and others is proposed.