External Validation of a Custom‐Trained Artificial Intelligence Model for Canine Ear Canal Disease and Comparison With Veterinary Professionals
Veterinary Dermatology · 2026-10 · 37 (5) · 850-857
DOI: 10.1111/vde.70105 ↗Kurz gefasst
Validierungsstudie in der Zeitschrift Veterinary Dermatology, online erschienen am 15.07.2026. Verfasst von Marley McInnis und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Veterinary Clinical Sciences, Louisiana State University School of Veterinary Medicine, Baton Rouge, Louisiana, USA. Hauptthemen laut MEDLINE: Dog Diseases, Artificial Intelligence, Otitis Externa. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 21 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.
Abstract (Originalsprache)
Background
Otitis externa (OE) is common in dogs, yet accurate otoscopic diagnosis remains challenging. Artificial intelligence (AI) may augment clinical decision‐making in otoscopy. Previous work suggested that dataset size and duplicate images affect model performance. Hypothesis/Objectives External validation of four YOLOv8n models using new images and comparison with six veterinary professionals of varying experience. We expected that models trained on larger, deduplicated datasets would demonstrate superior performance, and that the best‐performing model would outperform non‐specialist veterinary professionals while remaining comparable to board‐certified referral clinicians. Materials and Methods Four YOLOv8n models (A–D) were trained using different dataset compositions and externally validated on 300 previously unseen otoscopic images (100 healthy, 100 OE, 100 ear canal masses). Predictions were compared with reference labels assigned by a board‐certified referral clinician. Six veterinary professionals independently classified the same images, blinded to ground truth and AI outputs. Performance was assessed across multiple confidence thresholds (0.25–0.75) using overall agreement, Cohen's κ , sensitivity and specificity.
Results
Model D (largest deduplicated dataset) demonstrated the highest external performance, achieving 66.3% agreement and κ = 0.508 at a 0.25 confidence threshold. Classwise sensitivity/specificity for Model D (0.25) were: healthy (90.0%/69.0%), OE (59.0%/75.5%) and mass (50.0%/93.5%). Veterinary professionals outperformed all AI models (median agreement 77.2%, κ = 0.658), with the referral clinician achieving the highest concordance. Conclusions and Clinical Relevance Model D generalised best during external validation yet remained inferior to veterinary professionals. These results support the model's role as a decision‐support tool rather than a replacement for clinical expertise.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Veterinary Clinical Sciences, Louisiana State University School of Veterinary Medicine, Baton Rouge, Louisiana, USA.
- Office of Research and Graduate Education, School of Veterinary Medicine, Baton Rouge, Louisiana, USA.
ORCID-Profile: Marley McInnis, Chin-Chi Liu, Neoklis Apostolopoulos
Zitieren
APA
McInnis, M., Liu, C. C., Arbona, V., & Apostolopoulos, N. (2026). External Validation of a Custom‐Trained Artificial Intelligence Model for Canine Ear Canal Disease and Comparison With Veterinary Professionals. Veterinary Dermatology, 37(5), 850-857. https://doi.org/10.1111/vde.70105
Vancouver
McInnis M, Liu CC, Arbona V, Apostolopoulos N. External Validation of a Custom‐Trained Artificial Intelligence Model for Canine Ear Canal Disease and Comparison With Veterinary Professionals. Veterinary Dermatology. 2026;37(5):850-857. doi:10.1111/vde.70105
BibTeX
@article{McInnis2026,
author = {McInnis, Marley and Liu, Chin-Chi and Arbona, Veronica and Apostolopoulos, Neoklis},
title = {External Validation of a Custom‐Trained Artificial Intelligence Model for Canine Ear Canal Disease and Comparison With Veterinary Professionals},
journal = {Veterinary Dermatology},
year = {2026},
volume = {37},
number = {5},
pages = {850--857},
doi = {10.1111/vde.70105}
}
EUVIMED