By Cory Penn, DVM; Kristin Owens, DVM, DACVP; Eric Morissette, DVM, DACVP

Artificial intelligence (AI) holds vast potential to improve care for companion animals, from streamlining routine tasks to enabling accurate diagnostics.^1-3 AI's ability to quickly and accurately analyze complex datasets is already driving advancements in areas such as imaging, hematology, cytology, parasitology, and record-keeping, leading to its increasing adoption by veterinary professionals.⁴

An area that shows particularly strong promise is cancer diagnostics—a critical need, as one in four dogs,⁵ and one in five cats⁶ are diagnosed with cancer in their lifetime. As the leading cause of death in pets beyond middle age,⁵ cancer is a significant concern for pet owners,⁷ many of whom experience anxiety and depression when faced with a diagnosis.⁸

By increasing access to care for diagnostic detection, AI may help ease this emotional burden while improving outcomes for pets.

Diagnosing cancer in canine and feline patients

Cancer is a significant concern in veterinary medicine. The most commonly diagnosed cancer in both cats and dogs are lymphoma (accounting for up to 24 percent of all new canine cancers), osteosarcoma (the most prevalent primary bone tumor, making up 85 percent of skeletal tumors), and mast cell tumors, which are the leading type of skin tumor in dogs.⁵

Diagnosing cancer involves identifying neoplasia—any uncontrolled abnormal growth of cells or tissues within the body. While the terms "tumor" and "mass" are often used to describe the physical appearance of neoplasms, only malignant neoplasms qualify as true cancers.

The key to diagnosis is identifying neoplastic cells, which are central to understanding the nature and behavior of a tumor. For most tumors, cytology can quickly provide information about the lesion and provide individualized medical or surgical direction for the veterinarian, specific to that patient.⁹

This article explores how digital and AI-driven capabilities can be utilized to identify potentially neoplastic cells in lymph node and subcutaneous lesions. The faster we can accurately diagnose cancer, the earlier we can alert pet owners and start a treatment plan, improving outcomes and enhancing the quality of life for affected animals.

Understanding AI classification

The familiar saying "work smarter, not harder" is especially relevant in 2025, as the veterinary industry continues to face a shortage of professionals, workforce complexities, and burnout.¹⁰ If we are to maintain high-quality care for patients, especially those with cancer, it is essential we look inward to improve efficiencies and manage workflows. This is where the exponential growth in technology, particularly AI, can be tapped into.

The latest diagnostic tools use advanced digital connectivity and AI integration. These systems rely on different learning methods to perform tasks, each with their own strengths and limitations. In veterinary diagnostics, the choice of AI learning techniques can have an impact on how we ensure accuracy and safety.¹¹

Two primary techniques are machine learning (ML) and its subset, deep learning (DL):

1. Superficial machine learning. Based on human-provided training data, these algorithms evaluate limited features, such as abnormal cell attributes. However, these systems are constrained by the trainer's ability to identify and teach observable features, restricting their scope.¹¹
2. Deep learning. Deep learning is a type of machine learning that uses artificial neural networks — computational models inspired by the structure and function of the brain. These layered networks learn to recognize complex patterns in data by adjusting connections between "neurons," much like how the brain strengthens neural pathways through experience.

Unlike superficial learning, deep learning employs advanced convolutional neural networks (CNNs) to analyze data at microscopic, pixel-level detail, uncovering thousands of features and relationships beyond human awareness. This enables deeper insights and a more precise recognition, such as identifying abnormal cells.¹¹

Unlike superficial learning, deep learning systems improve with frequent, high-quality input, refining their algorithms to deliver enhanced, data-driven analyses.¹¹ However, prior to release, any new algorithm should be validated compared to a gold standard.

Cytological innovations

Cytology, or cytopathology, is the microscopic evaluation of individual cells or groups of cells to identify their origin and detect signs of disease.¹² It is a key diagnostic method used to identify cancer as it can be applied to a variety of tissues, including, but not limited to, lymph nodes, visceral organs, musculoskeletal structures, skin, reproductive organs, and even brain and
eye tissues.

Additionally, fluids, such as cerebrospinal fluid (CSF), blood, and body cavity effusions, can be analyzed. Its minimally invasive nature, lower complication risk, and quicker turnaround make it indispensable in veterinary oncology for detecting malignancies early and planning treatment strategies.¹²

However, cytology's effectiveness depends on the quality of sample collection, preparation, and interpretation — factors that can affect diagnostic reliability.¹²

Historically, veterinary cytology diagnostics have included both in-house and external laboratory evaluations. In-house analyses offer speed but are often limited by expertise and technology, while external labs provide higher diagnostic accuracy at the cost of extended turnaround times.¹³

In the digital age, technology has evolved to refine cytological processes further. Digital cytology (combines whole slide imaging (WSI) and cloud-based networks to enable remote evaluation by a clinical pathologist, eliminating logistical delays and providing fast, high-resolution diagnostic capabilities.

Mass identification with AI

In diagnostic imaging, AI enhances accuracy and reproducibility, reliably differentiating between cancerous and non-cancerous tissues.¹⁴ Machine learning algorithms can segment tumor areas from non-tumor regions and classify subsets of tumors with levels of accuracy comparable to experienced pathologists.¹⁵ Advanced models can also distinguish healthy tissues from cancerous ones with impressive precision, even when tumor cell infiltration is minimal.¹⁶

Further, by leveraging morphological and textural features, AI systems provide a robust framework for automated diagnostic tools, enabling histologists to refine diagnostic processes and reduce misdiagnosis rates.¹⁷

Recently available capabilities, such as AI-powered image recognition, hold promise for streamlining cytological workflows and helping improve diagnostic outcomes in veterinary practice.

Advancing diagnostic horizons

As cytology evolves with these innovative tools, the integration of digital platforms and AI technologies marks a pivotal shift in the field. By addressing challenges such as sample rejection rates and increasing diagnostic accessibility, these advancements empower veterinary professionals to deliver faster, more reliable care while ensuring high diagnostic standards.

Integration between practice management systems (PIMS) and advanced tools reduces administrative burdens, empowering veterinary professionals to do more with less.¹⁸

With more AI-powered diagnostics emerging, we foresee more veterinarians with an increased ability to quickly and accurately identify tumors, assess malignancies, and develop targeted treatment plans with greater confidence.

Eric Morissette, DVM, DACVP, earned his DVM from the University of Montreal and completed a clinical pathology residency at the University of Florida. At Zoetis Global Diagnostics, he works with a multidisciplinary team to advance AI-driven veterinary diagnostics, contributing to the development of Imagyst applications that improve diagnostic capabilities worldwide.

Kristin Owens, DVM, DACVP, earned her DVM from North Carolina State University and completed a Clinical Pathology residency at the University of Pennsylvania. She leads innovation in veterinary diagnostics through AI application development with the Imagyst platform at Zoetis, advancing diagnostics and transforming clinical pathology practice.

Cory Penn, DVM, received his DVM from the University of Missouri College of Veterinary Medicine after receiving his Bachelor of Science from Eastern Illinois University. Before joining Zoetis, Dr. Penn worked as a small animal veterinarian and medical director in Central Illinois.

References

1. Morissette, E., Penn, C. D., Hall Sedlak, R. A., Rhodes, A. J., Tippetts, D. S., Loenser, M., & Goldstein, R. (2025). Point-of-care platform integrated with deep-learning, convolutional neural network algorithms effectively evaluates canine and feline peripheral blood smears. American Journal of Veterinary Research, 86(2), ajvr.24.08.0226.
2. Nagamori Y, Scimeca R, Hall-Sedlak R, Blagburn B, Starkey LA, Bowman DD, Lucio-Forster A, Little SE, Cree T, Loenser M, Larson BS, Penn C, Rhodes A, Goldstein R. Multicenter evaluation of the Vetscan Imagyst system using Ocus 40 and EasyScan One scanners to detect gastrointestinal parasites in feces of dogs and cats. J Vet Diagn Invest. 2024 Jan;36(1):32-40.
3. Steuer A, Fritzler J, Boggan S, Daniel I, Cowles B, Penn C, Goldstein R, Lin D. Validation of Vetscan Imagyst, a diagnostic test utilizing an artificial intelligence deep learning algorithm, for detecting strongyles and Parascaris spp. in equine fecal samples. Parasit Vectors. 2024 Nov 12;17(1):465.
4. Gabor, S., Danylenko, G., and Voegeli, B. 2025. "Familiarity with Artificial Intelligence Drives Optimism and Adoption Among Veterinary Professionals: 2024 Survey." American Journal of Veterinary Research 86 (S1): ajvr.24.10.0293.
5. Veterinary Cancer Society. n.d. "Pet Owner Resources." Accessed April 14, 2025. https://vetcancersociety.org/resources/pet-owners/pet-owner-resources/.
6. Flint Animal Cancer Center. 2019. "Common Cancers in Cats." Accessed April 14, 2025. https://www.csuanimalcancercenter.org/2019/10/20/common-cancers-in-cats/.
7. Prescott, Elizabeth. 2024. "Dog Cancer Grief: During and After." Accessed April 14, 2025. https://www.dogcancer.com/articles/hospice-and-end-of-life/dog-cancer-grief-during-and-after/.
8. Nakano, Yuko, Masato Matsushima, Azusa Nakamori, Junshiro Hiroma, Eiji Matsuo, Hidetaka Wakabayashi, Shuhei Yoshida, Hiroko Ichikawa, Makoto Kaneko, Rieko Mutai, Yoshifumi Sugiyama, Eriko Yoshida, and Tetsuya Kobayashi. 2019. "Depression and Anxiety in Pet Owners after a Diagnosis of Cancer in Their Pets: A Cross-Sectional Study in Japan." BMJ Open 9 (2): e024512. doi:10.1036/bmjopen-2018-024512.
9. American Veterinary Medical Association. n.d. "Cancer in Pets." Accessed April 14, 2025. https://www.avma.org/resources/pet-owners/petcare/cancer-pets.
10. Gitter, Robert J., and Bill LaFayette. 2024. Demand for and Supply of Veterinarians in the U.S. to 2032. American Association of Veterinary Medical Colleges. Accessed April 17, 2025. https://www.aavmc.org/wp-content/uploads/2024/06/Demand-for-and-Supply-of-Veterinarians-in-the-U.S.-to-2032-New.pdf.
11. Zoetis. 2023. "How Artificial Intelligence Is Changing Veterinary Medicine." Accessed April 17, 2025. https://www.zoetisus.com/petcare/blog/how-artificial-intelligence-is-changing-veterinary-medicine.
12. Monti, Paola, and Francesco Cian. 2016. "Diagnostic Cytology." In BSAVA Manual of Canine and Feline Clinical Pathology, 3rd ed., edited by Elizabeth Villiers and Jelena Ristić, 398–434. Gloucester: British Small Animal Veterinary Association. doi:10.22233/9781910443255.chap21.
13. Skeldon, N., and E. Dewhurst. 2009. "The Perceived and Actual Diagnostic Utility of Veterinary Cytologic Samples." Journal of Small Animal Practice 50 (4): 180–185. doi:10.1010/j.1748-5827.2009.00739.
14. Alshammari, Aya Hasan, Takuya Oshiro, Umbhorn Ungkulpasvich, Junichi Yamaguchi, Masayo Morishita, Sura Abbas Khdair, Hideyuki Hatakeyama, Takaaki Hirotsu, and Eric di Luccio. 2025. "Advancing Veterinary Oncology: Next-Generation Diagnostics for Early Cancer Detection and Clinical Implementation." Animals 15 (3): 389. doi:10.3390/ani15030389
15. Fragoso-Garcia, Marco, Frauke Wilm, Christof A. Bertram, Sophie Merz, Anja Schmidt, Taryn Donovan, Andrea Fuchs-Baumgartinger, et al. 2023. "Automated Diagnosis of 7 Canine Skin Tumors Using Machine Learning on H&E-Stained Whole Slide Images." Veterinary Pathology 60 (6): 865–875. doi:10.1077/03009858231089205.
16. Rupšytė, G., R. Raišutis, M. Maciulevičius, I. M.-v. Houtana, and M. Tamošiūnas. 2024. "Diagnosis of Neoplasms in Veterinary Medicine Samples Using Artificial Intelligence." In Proceedings of the 2024 IEEE 14th International Conference on Nanomaterials: Applications & Properties (NAP), Riga, Latvia, pp. 1–8. doi:10.1009/NAP62956.2024.10739756.
17. Maciułevičius, Martynas, Greta Rupšytė, Blaž Cugmas, Renaldas Raisutis, and Mindaugas Tamošiūnas. 2024. "The Differentiation of (Sub-)Cutaneous Canine and Feline Tumors via Machine Learning." SSRN. Accessed April 14, 2025. https://ssrn.com/abstract=4914680.
18. Basran, Parminder S., and Ryan B. Appleby. 2022. "The Unmet Potential of Artificial Intelligence in Veterinary Medicine." American Journal of Veterinary Research 83 (5). doi:10.2460/ajvr.22.03.0038.