Validation of Dual-Energy X-Ray Absorptiometry to Estimate the Carcass Tissue Composition of Nellore Cull Cows

Read full text
Authors: Wenderson Moura de Carvalho, Guilherme Marques Rodrigues, Ana Clara Antunes Luiz, M. C. Gonçalves, Vitória Marques de Souza Ferreira, Cris Luana de Castro Nunes, Simone Eliza Facioni Guimarães, M. L. Chizzotti
Cited by: 0
References: 17

Abstract

This study aimed to evaluate the use of dual-energy X-ray absorptiometry (DXA) to predict the physical composition of half carcasses from Nellore cull cows, comparing DXA estimates with data obtained by tissue dissection. Twenty-four half carcasses were cut into 5 standardized anatomical sections, and each section was individually scanned by DXA and subsequently fully dissected into lean, fat, and bone tissues. Dissection indicated that each carcass side consisted, on average, of 85.19 kg plus or minus 10.43 kg of lean, 22.93 kg plus or minus 7.89 kg of fat, and 21.31 kg plus or minus 2.75 kg of bone. DXA overestimated lean and fat values by 19.35% and 14.42%, respectively, while bone mineral content was underestimated by 66.82% compared to the total dissected bones. Linear regression models demonstrated high predictive accuracy for fat (coefficient of determination of cross-validation [R2 CV] = 0.96; root mean square error of cross-validation [RMSECV] = 1.53 kg), satisfactory prediction for lean (R2 CV = 0.84, RMSECV = 4.31 kg), and moderate performance for bone (R2 CV = 0.73, RMSECV = 1.46 kg). Regarding the contrasting of whole half-carcass tissue with each section, S2 provided the best prediction for total half-carcass fat (R2 CV = 0.88, RMSECV = 2.75 kg), S4 and S5 for total lean (R2 CV = 0.78, RMSECV = 5.07 kg; R2 CV = 0.76, RMSECV = 5.25 kg, respectively), and S1 for total bone (R2 CV = 0.80, RMSECV = 1.25 kg). Building on these section-level predictions, DXA estimates for total carcass lean and section-based scans also reliably predicted the overall dissected lean mass of the half carcass. These results demonstrate that DXA-based regression models accurately predict the lean, fat, and bone composition of Nellore cull cow carcasses and that scanning specific sections provides a practical, nondestructive alternative. This highlights the need to evaluate this method in other carcass types, such as those with higher fat or muscle content.

Fields of Study & Subjects

Body Composition Measurement Techniques Ruminant Nutrition and Digestive Physiology Meat and Animal Product Quality

Publication details

Journal / source
Meat and Muscle Biology
Volume
10
Issue
1
Publication date
29 January 2026