Identifying necrotizing soft tissue infection using infectious fluid analysis and clinical parameters based on...

Area of Science:

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

Background:

• Necrotizing soft tissue infection (NSTI) diagnosis is challenging.
• Artificial intelligence (AI) and machine learning (ML) offer efficient, precise disease identification.
• AI/ML adoption in diagnostics is increasing.

Purpose of the Study:

• To develop and evaluate ML models for NSTI diagnosis.
• To compare the diagnostic performance of different ML algorithms.
• To identify key features for an optimized NSTI diagnostic model.

Main Methods:

• Trained four ML models (random forest, KNN, SVM, logistic regression) on data from 13 NSTI and 12 cellulitis patients.
• Utilized 28 distinct features identified through statistical analysis.
• Developed a refined random forest model using the 6 most influential features.

Main Results:

• Random forest model achieved 89.6% sensitivity and 92.9% specificity.
• The optimized random forest model showed 90.2% sensitivity and 92.2% specificity.
• Random forest outperformed other ML models and fluid lactate levels in diagnostic accuracy.

Conclusions:

• A random forest-based ML model offers an efficient, cost-effective, and rapid diagnostic tool for NSTI.
• This AI-driven approach can enhance clinical decision-making for NSTI management.
• The developed model provides healthcare practitioners with improved efficacy in identifying and treating NSTI.