Dogs can sniff out hemangiosarcoma in blood samples, early study shows

A team from the Penn Vet Working Dog Center tested whether highly trained detection dogs could identify an odor signature associated with canine hemangiosarcoma. The work is a controlled, proof-of-concept laboratory study that used automated olfactometer line-ups and double blinding to reduce bias. Serum came from dogs with histologically confirmed hemangiosarcoma, from dogs with other diseases, and from healthy controls. In total, dogs completed 423 blinded trials.¹

Five biodetection dogs performed the task. Across all 423 trials the dogs’ overall accuracy was 70.0%, with individual dog performance ranging from 57.1% to 78.6%. On initial exposure, mean accuracy was also 70.0%, with first-trial results ranging from 58.3% to 83.3%. Sensitivity and specificity calculated from first-trial encounters were both approximately 70%.¹

The dogs alerted to hemangiosarcoma samples far more often than to controls. They signaled for HSA samples in 73.4% of presentations, compared with 21.3% for diseased controls and 17.1% for healthy controls. Odds ratios indicated dogs were more than 10 times as likely to alert to HSA versus diseased controls, and more than 13 times as likely to alert to HSA versus healthy controls. All comparisons were statistically significant.¹

For clinicians, the study is important because it provides evidence that hemangiosarcoma is associated with a detectable volatile organic compound profile in blood. That is a biologic signal that could, in principle, be captured by an engineered sensor or a standardized volatile organic compound (VOC) assay. At the same time this is an early laboratory result and not a validated clinical screening test. Accuracy near 70% and a relatively small set of hemangiosarcoma samples limit immediate clinical application. The testing environment used highly trained dogs and controlled presentation methods that are not directly transferable to the busy clinic.¹

Limitations noted by the authors and apparent from the methods include the modest number of HSA-positive samples and the need for external validation in larger, independent clinical cohorts. Future work should focus on identifying the specific VOCs responsible for the signal, standardizing sample handling, and developing objective sensor technology that can reproduce canine olfaction under clinical conditions.¹

For now this study is best read as a rigorous proof of concept. It supports further research into VOC-based diagnostics for hemangiosarcoma and other cancers, but it does not change diagnostic or screening recommendations at this time. Clinicians should monitor follow-up studies that report external validation and assay development.¹

Reference

Wilson C, Holden S, King J, et al. Trained dogs can detect the odor of hemangiosarcoma in canine blood samples. Vet J. Published online November 29, 2025. doi:10.1016/j.tvjl.2025.106522