AI in Veterinary Cytology

Cytology examines cells from a sample, such as an aspirate or fluid, under a microscope. AI research applies image-analysis models to digitized slides to locate cells or classify selected patterns. That scope matters: a model trained for one specimen and one distinction is not a general system for diagnosing every tumor, blood disorder or infection. A 2026 study evaluated an object-detection model on canine lymph-node cytology images for a defined task: identifying lymphocyte categories that can support assessment of lymphoma. The study used images from a limited set of aspirates and tested cross-device configurations; the authors still called for prospective, workflow-integrated and external validation across more lymph-node diseases. Its results are evidence about that dataset and task, not proof of universal diagnostic performance. Digital slide transfer is another workflow. Zoetis describes its VETSCAN Imagyst Digital Cytology service as sending whole-slide images to board-certified clinical pathologists for review. Scanning a slide for a remote human reader is not the same thing as an AI classifier. The platform also lists separate AI tests for other sample types, such as fecal parasite screening, which is not cytology. A veterinarian uses findings alongside sample quality, history, physical examination and other tests. Uncertain, atypical or consequential cases may need specialist review or additional testing. AI can help organize or screen image information, but it does not replace a clinical pathologist’s interpretation or the treating veterinarian’s decision.