Automated acute pain assessment in brachycephalic cats with ocular pain using the Feline Grimace Scale.

AI-powered Feline Grimace Scale pain assessment in brachycephalic cats requires breed-specific datasets before reliable clinical use.

This study evaluated the feasibility of an AI-powered Feline Grimace Scale (FGS) pipeline for automated acute pain assessment specifically in brachycephalic cats experiencing ocular pain. Using 200 facial images for model training and evaluation, the researchers employed a landmark localization model combined with XGBoost regression submodels corresponding to the five FGS action units: ear position, orbital tightening, muzzle tension, whiskers change, and head position. Three input variable configurations were tested to optimize pain prediction performance. Facial landmark detection accuracy was notably reduced in brachycephalic cats (NRMSE 19.96%) compared to previously reported performance in predominantly mesocephalic cats (NRMSE 16.76%), reflecting the anatomical challenges posed by flattened facial conformations. Pain prediction models demonstrated limited generalizability overall, with the best-performing configuration excluding muzzle tension and whiskers change yielding a mean absolute error of 0.168, R² of 0.115, and Pearson correlation of 0.339. These modest results underscore that the current AI-FGS pipeline, while technically applicable to brachycephalic cats, is not yet reliable enough for clinical deployment in this population. The authors attribute the poor predictive performance primarily to the limited size of the annotated dataset. The study concludes that conformation-specific datasets are essential before this technology can be validated for routine pain assessment in brachycephalic cats, representing an important step toward breed-specific AI pain assessment tools in feline medicine.

This summary was distilled by AI and may occasionally misinterpret data. Confirm critical details with the primary literature before clinical application.