Submitted to: International Journal of Poultry Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: December 28, 2005
Publication Date: January 20, 2006
Citation: Windham, W.R., Heitschmidt, G.W., Smith, D.P., Berrang, M.E. 2006. Detection of ingesta on pre-chilled boiler carcasses by hyperspectral imaging. International Journal of Poultry Science. 4(12):959-964. Interpretive Summary: Undigested feed in the upper digestive tract of chickens, commonly referred to as ingesta, has been suggested as a source of bacteria. The two organs most likely to contain ingesta are the crop and gizzard and the crop contents have been observed to be a source of visible contamination on chicken carcasses when accidentally ruptured during commercial processing. The U. S. food safety performance regulations mandate that no visible fecal spots of any size can appear on carcasses at certain points in commercial processing. In this study, we used a camera and computer to detect visible ingesta from crop and gizzard contents on chicken carcasses and then determined the microbiological profile of the ingesta contamination. The camera system accurately detected all contamination on the carcasses. However, contamination of carcasses with ingesta did not change the microbiological profile. Based on the number of bacteria found, carcass contamination with visible ingesta does not appear to significantly increase total carcass bacteria.
Technical Abstract: The contents of the upper digestive tract (i.e. crop, proventriculus and gizzard) may serve as a source of carcass contamination during broiler processing. The crop as been identified as a source of Salmonella and Campylobacter on contaminated carcasses and is more likely to rupture than the ceca during commercial evisceration. The objective of this study was to determine the effectiveness of hyperspectral imaging for detecting ingesta contamination spots varying in mass from the crop and gizzard. Pre-chilled broiler carcasses were collected from a commercial processing plant. Crop and gizzard contents were also aseptically collected and enumerated for Campylobacter, coliforms, E. coli and total aerobic bacteria. Broiler carcasses were imaged and then contaminated with a spot of known mass (10, 50, or 100 mg) of crop or gizzard contents. Carcasses were then re-imaged. The imaging system correctly detected 100% of the crop and gizzard contents regardless of the mass or spot size. However, not every pixel associated with a given spot (contaminant ground truth) was detected. Detection of crop and gizzard content contaminant ground truth pixels averaged 72 and 53%, respectively. The mean number of bacteria in the crop contents were as follows: E. coli 4.0 log, coliforms 4.1 log, and total aerobic bacteria 5.7 log CFU/g of crop contents. Crop contents in the current study were Campylobacter negative. Applying crop contents in the amounts of about 9, 54, and 231 mg resulted in significant (P<0.05) increases in all bacterial population measured, with the biggest increase being noted for total aerobic bacteria. Gizzard contents contained only 4.6 log CFU/g of total aerobic bacteria. The total added bacterial load from contamination with known amounts of crop and gizzard contents did not significantly increase whole carcass counts of all bacteria enumerated. Based on these counts and numbers of bacteria found in gizzard, carcass contamination with visible ingesta does not appear to significantly increase bacterial load.