Image analysis for MBM characterization : preliminary results of comparison between bovine and poultry lacunae

Introduction: The reintroduction of non-ruminant PAP in aquafeed was voted and adopted on 18th July 2012 at the European Commission level. To that purpose, PCR was added to the microscopic one, as official method for detecting the species origin of PAP in a new version of annex VI of regulation 152/2009 (Reg 51/2013). Nevertheless none of the method is able on its own to fit all requirements for the accurate identification of prohibited ingredients of animal origin (e.g. differentiate between authorized and prohibited ingredients), leading to propose a combinatory approach in which all methods can be used, implemented, and eventually merged. Accordingly the aim of this work was to combine the official analytical method with the image analysis measurements for the detection of constituents of animal origin in feedstuffs, in order to discriminate between bovine and poultry lacunae. Materials and Method:10 feed samples contaminated with 0.5% pure bovine meal and 12 samples containing 0.5% pure poultry meal were processed to obtain sediments, according to Annex VI of Regulation 152/2009. Sediment fractions of each sample were observed with a compound microscope at X40. 430 bone fragment lacunae images (215 from bovine and 215 from poultry) were obtained, processed and elaborated using Image Pro Plus® software, in order to acquire 8 geometric variable measurements for each lacuna (area, axis minor, diameter min, radius min, roundness, size width, area polygon and feret min). Mean, median and standard deviation values were calculated for each geometric variable measured. Results: Results obtained in the present study have indicated that all variables measured in bovine resulted on average 60% higher than in poultry. In both bovine and poultry, median for the same variables was always lower than mean value. When different variables have been considered bovine and poultry bone lacunae were overlapping, except for diameter min and size width which resulted the most promising as potential markers in distinguishing between the two materials tested. Conclusions: The use of microscopic methods in association with computer image analysis to identify bovine vs. avian material appears promising although there are several limitations, particularly when trying to distinguish material at taxonomic level of species. However present study is based on a limited number of samples and additional observations/studies are needed, in order to get an indepth analysis.