Machine Learning Approach to Classify Vibrio vulnificus Necrotizing Fasciitis, Non-Vibrio Necrotizing Fasciitis and...

Area of Science:

• Medical Informatics
• Machine Learning in Healthcare
• Infectious Disease Diagnostics

Background:

• Artificial intelligence (AI) and machine learning (ML) are increasingly used for disease identification.
• AI offers efficiency, objectivity, and accuracy, addressing diagnostic challenges in clinical practice.
• Focus on difficult diagnoses such as necrotizing fasciitis and Vibrio vulnificus infections.

Purpose of the Study:

• To develop and validate a machine learning model for classifying soft tissue infections.
• To predict the occurrence of cellulitis, non-Vibrio necrotizing fasciitis (NF), or Vibrio vulnificus NF.
• To interpret model predictions using SHapley Additive exPlanations (SHAP).

Main Methods:

• A multi-class categorization model was developed using Light Gradient Boosting Machine (LightGBM).
• 180 inpatients with soft tissue infections were categorized into cellulitis, non-Vibrio NF, or V. Vulnificus NF groups.
• 5-fold cross-validation and SHAP methodology were employed for model development and interpretation.

Main Results:

• The model demonstrated strong predictive performance with a weighted-average AUC of 0.86.
• Key performance metrics included sensitivity of 87.2%, specificity of 74.5%, NPV of 81.6%, and PPV of 85.4%.
• A low Brier score (weighted mean of 0.084) indicated high predictive accuracy and reliability.

Conclusions:

• A reliable multiclassification model was developed to forecast cellulitis, non-Vibrio NF, or V. Vulnificus NF in soft tissue infections.
• The SHAP algorithm was successfully used to explain the model's predictions.
• The study highlights the potential of AI in improving the diagnosis of complex infections.