Applying Supervised Machine Learning to Effusion Analysis for the Diagnosis of Feline Infectious Peritonitis.

Area of Science:

• Veterinary Medicine
• Machine Learning
• Immunopathology

Background:

• Feline infectious peritonitis (FIP) is a fatal disease in cats caused by an aberrant immune response to feline coronavirus.
• Current FIP diagnosis is complex, relying on non-specific clinical signs and biomarkers, especially for effusive FIP.
• Accurate diagnosis of effusive FIP is challenging due to difficulties in interpreting fluid analysis results.

Purpose of the Study:

• To investigate the application of machine learning in diagnosing effusive FIP using fluid analysis data.
• To develop and validate a machine learning model for predicting FIP status from veterinary laboratory records.
• To assess the potential of machine learning to improve diagnostic accuracy and standardization in veterinary settings.

Main Methods:

• A dataset of 718 historical veterinary laboratory records for suspected effusive disease was compiled.
• The dataset included 336 confirmed FIP cases and 382 non-FIP cases.
• An ensemble machine learning model was trained using clinical observations and laboratory features to predict disease status.

Main Results:

• The machine learning model achieved high diagnostic performance.
• Accuracy: 96.51%
• Area Under the Receiver Operator Curve (AUC): 96.48%
• Sensitivity: 98.85%
• Specificity: 94.12%

Conclusions:

• Machine learning can be effectively applied to interpret fluid analysis results for accurate effusive FIP detection.
• This method shows significant potential for standardizing and enhancing diagnostic services in veterinary laboratories.
• The developed model offers a promising tool for improving the diagnosis of feline infectious peritonitis.