Computed tomography (CT) is a non-destructive technique, based on X-rays, that has been used in several livestock species to evaluate carcass composition. The objective of this study was to construct predictive equations to estimate carcass and viscera composition for preweaning calves using CT. For this purpose, 24 Holstein male calves (4 ± 0.9 d of age; 40 ± 2.2 kg of body weight) were fed a milk replacer (MR; 23% CP; 15% fat) either 4 L/d or 8 L/d of MR at the rate of 125 g/L of water to ensure different levels of fat and protein accretion and generate sufficient variation to obtain the equations of calibration. Then, at 30 ± 2.4 d of age, 3 calves from each feeding program, and at 50 ± 1.9 d of age, 9 calves from each feeding program were CT-scanned, and humanly sacrificed. Carcasses were also CT scanned 24 h post mortem. Images from CT were analysed and used to predict content of protein and fat of carcasses, red and white viscera. The models rendered a residual predictive deviation between 1.1 (protein red viscera) and 2.6 (fat white viscera) in live animal images and between 1.1 (carcass moisture) and 4.5 (fat white viscera) in carcass images. The root mean square error of prediction relative to the mean ranged between 1.32 (carcass moisture) and 17.3% (fat white viscera) in live animal images and between 1.38 (carcass moisture) and 17.3 (fat red viscera) in carcass images. The coefficient of determination ranged between 0.19 (protein red viscera) and 0.88 (fat white viscera) in images from live calves and between 0.26 (carcass protein) and 0.98 (fat white viscera) in carcass images. In conclusion, it is possible to predict body composition of calves using a non-destructive technique by means of computed tomography images and this prediction could be used in studies were the estimation of this content would be relevant.

计算机断层扫描（CT）是一种基于X射线的非破坏性技术，已用于多种牲畜物种中以评估car体成分。这项研究的目的是建立预测方程，以使用CT估算断奶前犊牛的car体和内脏成分。为此，给24头荷斯坦公犊牛（4±0.9 d龄； 40±2.2 kg体重）喂奶代乳品（MR; CP的23％；15％的脂肪）以125 g / L的水的速率添加4 L / d或8 L / d的MR，以确保不同水平的脂肪和蛋白质积聚，并产生足够的变化量以获得校准方程式。然后，在30±2.4 d的年龄，每个饲喂程序的3头犊牛，以及在50±1.9 d的年龄，每个饲喂程序的9头小牛，进行CT扫描，并处死。尸体死后24小时也进行CT扫描。分析来自CT的图像，并将其用于预测car体，红色和白色内脏的蛋白质和脂肪含量。该模型在活体动物图像中的残差预测偏差在1.1（蛋白红色内脏）和2.6（脂肪白内脏）之间，在屠体图像中残差预测偏差在1.1（car体水分）和4.5（胖白内脏）之间。相对于平均值的预测均方根误差在活动物图像中介于1.32（car体水分）和17.3％（脂肪白内脏）之间，在尸体图像中介于1.38（car体水分）和17.3（脂肪红内脏）之间。测定系数在活牛犊图像中介于0.19（蛋白质红色内脏）和0.88（脂肪白色内脏）之间，在car体图像中介于0.26（car体蛋白质）和0.98（脂肪白色内脏）之间。结论，