| Summary: | Soya bean meal-induced enteritis (SBMIE) is a well-documented condition observed in distal intestine (DI) on Atlantic salmon (Salmo salar) fed with unprocessed soya. Many of the current studies describing inflammation in DI of Atlantic salmon fed unprocessed soya bean meal (SBM) utilize a semi-quantitative scoring system for histological assessment, which is considered the gold standard. This thesis investigates if a quantitative assessment using image analysis with artificial intelligence (AI) can produce similar results as with semi-quantitative scoring systems. This study provides evidence that the results for assessing inflammation in the DI extracted from image analysis using AI correlate to results from similar studies utilizing semi-quantitative scoring systems. A standard diet not containing unprocessed SBM and a diet containing 25% SBM was used as control diets. Six experimental diets all containing 25% unprocessed SBM were supplemented with different inclusion levels of two phytogenic ingredients with anti-inflammatory characteristics. Atlantic salmon were distributed into 24 different experimental tanks and fed for 56 days. There was no significant effect from the experimental diets, as all the fish fed with SBM experienced severe morphological changes in the DI. The observed morphological changes indicated a decrease in the number of mucous, vacuoles and absorptive area fraction. An increase for lamina propria, EGC, IgD and apoptosis area fraction was also observed for fish fed with SBM. In total weight gained during the experimental trial, a significant difference was observed for the Atlantic salmon not receiving SBM as they gained 25.4% more weight.
|
|---|