AI-Powered Vaccine Shrinks Dog’s Cancer, Pioneering Personalized Treatment

A Sydney-based machine learning engineer, Paul Conyngham, has successfully used artificial intelligence (AI) to develop a personalized mRNA vaccine that significantly reduced a cancerous tumor in his dog, Rosie. This groundbreaking case highlights the potential of AI in revolutionizing veterinary and potentially human cancer treatment.

From Diagnosis to AI-Driven Solution

In 2024, Rosie, a Staffordshire Bull Terrier-Shar Pei cross, was diagnosed with mast cell tumors on her back leg. Conventional treatments like surgery, chemotherapy, and immunotherapy offered limited success, slowing the disease’s progression but failing to shrink the tumors. Veterinarians estimated Rosie had only one to six months to live [UNSW Sydney].

Leveraging AI for Personalized Medicine

Conyngham, with no formal biomedical training, turned to AI tools to explore potential solutions. He utilized ChatGPT to understand the molecular mechanisms of canine cancer and to devise a data pipeline. This pipeline involved sequencing the tumor’s DNA, comparing it to healthy cells, identifying mutations, and designing a vaccine targeting those specific proteins [aihola.com].

Collaboration with UNSW and AlphaFold’s Role

Conyngham collaborated with researchers at the UNSW Ramaciotti Centre for Genomics, where Associate Professor Martin Smith initially expressed skepticism about the request from a private individual. Still, Conyngham’s thorough analysis impressed the team. He used Google DeepMind’s AlphaFold to predict protein structures from the identified mutations and pinpoint potential drug targets [Let’s Data Science].

Vaccine Development and Remarkable Results

The analysis culminated in an mRNA sequence formula, which was then produced into a physical vaccine by Páll Thordarson, director of the UNSW RNA Institute, in under two months. After the first injection in December 2025, Rosie’s tumor shrank by 75% within a month [aihola.com]. Further observation in March 2026 showed continued tumor shrinkage [Let’s Data Science].

A Single Case, Significant Implications

Researchers emphasize that this is a single case involving one dog and one tumor, with no controlled trials conducted. Scientists at UNSW and the University of Queensland have confirmed the results, but acknowledge the limited scope of the study [aihola.com]. Despite this, the success represents a significant step towards personalized cancer treatment and demonstrates the potential of combining AI with traditional research methods. Conyngham is currently developing a second vaccine [Let’s Data Science].

The Future of AI in Cancer Research

This case raises the question of why similar approaches haven’t been explored in human patients. It exemplifies how a “citizen scientist” can leverage AI tools in collaboration with research institutions to achieve tangible medical outcomes, paving the way for innovative approaches to cancer treatment and research.