Comparison of traditional and artificial intelligence methods for detection of enteric pathogens in feces of dogs with diarrhea

Giardia duodenalis, the primary species infecting domestic dogs, is a common enteric pathogen that sheds infectious and environmentally resistant cysts in dog feces, presenting a common diagnostic challenge in veterinary clinics. The cysts can be detected using a number of diagnostic techniques, including several fecal flotation methods that have varying degrees of sensitivity and specificity. While one of the most sensitive cyst detection methods is direct fluorescent antibody (DFA), it requires special equipment and so is not performed in most veterinary clinics. The Vetscan Imagyst® (Zoetis), designed for in-clinic use, uses artificial intelligence (AI) enabled algorithms to detect Giardia cysts within approximately 10 minutes and has similar agreement to traditional centrifugal fecal flotation methods. The primary objective of this study was to use feces from shelter dogs with diarrhea to compare Giardia cyst detection results amongst traditional fecal flotation techniques including: zinc sulfate centrifugation (with expert read), sugar centrifugation (with expert read), and direct fluorescent assay (DFA) (with expert read) with AI-enabled Giardia cyst assay and AI-enabled fecal ova/oocyst assay. A secondary study objective using the same samples tested for Giardia was to verify results from prior published studies for other common gastrointestinal parasites using AI-enabled parasite detection compared with sugar centrifugation with expert read.. The sensitivity and specificity between DFA and AI-enabled Giardia cyst assay for detection of Giardia were 71.4% and 94.8%, respectively. The sensitivity and specificity between DFA and Zinc sulfate centrifugation (with expert read) were 71.4% and 100%, respectively. The sensitivity and specificity of Zinc sulfate centrifugation (with expert read) and AI-enabled Giardia cyst assay were 80.0% sensitivity and 89.4%, respectively. There were 3 samples that were DFA positive/ AI-enabled Giardia cyst assay negative and PCR confirmed that these three samples were falsely negative by the AI-enabled as say. There were three samples that were DFA negative/ AI-enabled Giardia cyst assay positive, and PCR confirmed two of the samples were false negative by DFA. Overall results from the three methods were similar and it was determined not to use the AI-enabled fecal ova/oocyst assay for Giardia detection due to having the lowest sensitivity. Results for detection of other common GI parasites were similar to prior published studies. The results suggest that the AI-enabled Giardia cyst assay could be considered for use in small animal clinics for Giardia and other parasite detection, in particular, if a parasite expert and/or DFA is not routinely available.