Relationship Between CT-Scanned and Dissected Subcutaneous and Intermuscular Fat

Abstract

In this study 6 whole beef rumps (3.67 to 5.05 kg) were purchased to evaluate the accuracy of estimating subcutaneous and intermuscular fat content from Computer-aided Tomography (CT)-scanned images. A full dissection of subcutaneous and intermuscular fat was used to evaluate the accuracy of the CT prediction. The methodology reported in this paper will be used to estimate the amount of subcutaneous and intermuscular fat in primal cuts, collected from a Beef CRC study investigating marbling and fat distribution in Angus, Hereford, and Wagyu × Angus steers within a serial slaughter experiment. The rumps were scanned using a Picker Ultra Z Spiral CT scanner (Philips Medical Imaging Australia, Sydney NSW) at the University of New England Meat Science CT unit. Voltage and current were set at 130 kV and 100 mAs, respectively. A pitch of 1.5, field of view (FOV) of 480 mm, slice thickness of 5mm and distance between slices of 15mm were used. One hundred and ten axial slices from 6 primal cuts were created. AutoCAT (Jopson et al 1995) and CT-CALC programme were then used to calculate total carcass fat (TCF) and lean in each rump. Boundaries for fat and lean were set at 10 to 128 and 129 to 210 grey scale units, respectively, and image diameter of 487mm was used. Java code in ImageJ was developed (M.J. McPhee pers. comm.) to remove the lean from each rump image slice, and intermusucular fat was removed from each slice using the elliptical tool in ImageJ. Therefore, it was hypothesised that subcutaneous fat was the only remaining fat depot. Subcutaneous fat images were then analysed to estimate the amount fat (kg) in that depot. Intermuscular fat (total fat - subcutaneous fat) was calculated by difference. A full dissection of each rump was performed by 2 operators. Subcutaneous and intermuscular fat depots were weighed and vacuum packed. Bags were then scanned to determine the amount of fat and lean in each dissected fat depot. Difficulties were encountered in completely removing all muscle from dissected fat, hence estimation of residual muscle was an important step to allow comparison between dissected fat (i.e. fat minus remaining muscle) and estimated fat from the CT images. The ratio of subcutaneous and intermuscular fat to total scanned (fat + lean) weight (kg) was determined to account for errors in weight estimation. The ratio for each fat depot was then multiplied by the weight of the primal cut to determine the amount of dissected fat in each depot. A regression of scanned images (dependent variable) versus dissected fat data (independent variable) for total, subcutaneous, and intermuscular fat were analysed using the lm procedure in the statistical package R. ... The results (Figure 1) illustrate a strong linear relationship between scanned and dissected fat. Subcutaneous fat had a higher adjusted r² (0.99) than intermuscular fat (0.83). The lower r² for intermuscular fat is attributed to the higher degree of difficulty in dissecting intermuscular fat. These results demonstrate that estimation of the amount of fat in different depots in beef primal cuts from scanned data is feasible and can be done with a high degree of accuracy.