- Vetology expanded its AI validation dashboard to report 11 performance metrics per classifier, up from four.
- The update covers more than 89 classifiers used in canine and feline thoracic, abdominal, and musculoskeletal imaging.
- 31 classifiers were retrained and revalidated using updated consensus data from board-certified veterinary radiologists.
- New condition classifiers include obscuring pleural effusion, esophageal enlargement, thoracic intervertebral disc disease, and feline gastrointestinal conditions.
- The company says publishing comprehensive validation metrics is intended to improve transparency and clinician trust in veterinary AI diagnostics.
Vetology has expanded its publicly available artificial intelligence performance dashboard, increasing the number of validation metrics reported for each diagnostic classifier from four to eleven.
The update includes statistical profiles for more than 89 classifiers used to analyze canine and feline thoracic, abdominal, and musculoskeletal imaging. The dashboard now provides sensitivity, specificity, positive predictive value, negative predictive value, area under the curve, F1 score, accuracy, prevalence, confidence intervals, and radiologist agreement rate for each condition.
According to the company, the expanded reporting is designed to give veterinarians greater visibility into how the AI models perform and how their outputs should be interpreted in clinical decision-making.
Retraining and Revalidation of Existing Models
Of the classifiers currently listed on the dashboard, 31 are retrained models that were originally released and later revalidated using updated consensus readings from board-certified veterinary radiologists. Confusion matrices for the models have been regenerated using validation data as recently as February 2026.
Eric Goldman, president of Vetology, said that maintaining and revalidating existing models is as important as building new ones in a rapidly evolving AI environment.
He noted that publishing validation data for both new and retrained models reflects the company’s commitment to transparency with veterinary partners and the patients they serve.
New Condition Classifiers Added
The latest dashboard update also introduces several new classifiers designed to detect specific imaging findings and disease indicators.
These include classifiers for obscuring pleural effusion, esophageal enlargement, thoracic intervertebral disc disease, feline small intestine enlargement, feline diffuse colon distension, and a consolidated heart failure classifier for canine imaging.
Cory Clemmons, chief technology officer at Vetology, said that providing multiple performance metrics allows clinicians to better evaluate how reliable an AI-generated result may be in a real clinical setting.
He said that metrics such as positive predictive value, specificity, and confidence intervals help veterinarians determine how much weight to place on a model’s findings when reviewing imaging studies.
Data Transparency in Veterinary AI
Vetology reports that its validation dataset is built on approximately 300,000 multi-image patient cases. The company cited a 2026 Frontiers in Veterinary Science audit indicating that 63.3 percent of commercial veterinary AI vendors do not publicly disclose validation data.
By publishing expanded statistical profiles, Vetology says it aims to establish a higher standard of transparency within the veterinary AI field.
About Vetology
Vetology provides AI-generated radiology screening reports and on-demand teleradiology services delivered by board-certified veterinary radiologists, cardiologists, and a veterinary dentist. The company’s AI screening platform analyzes canine and feline imaging across thoracic, abdominal, and musculoskeletal studies and is designed to integrate with existing veterinary clinic workflows.
Information sourced from the company’s press release.
