# Predictive Models for Necrotizing Soft Tissue Infections: Are the Available Scores Trustable?

## TL;DR

This paper reviews the reliability of various scoring systems for predicting outcomes in necrotizing soft tissue infections and finds that none are fully trustworthy.

## Contribution

The paper evaluates five scoring systems for necrotizing soft tissue infections and suggests the need for a more reliable integrative model.

## Key findings

- The LRINEC score lacks high sensitivity and needs additional clinical parameters.
- The NECROSIS score is promising but lacks external validation.
- The POTTER score uses AI and machine learning but is not yet valid for NSTI patients.

## Abstract

Background: Necrotizing soft tissue infections (NSTIs) remain a significant source of in-hospital morbidity and mortality in the U.S. and around the world, yet the need for a reliable tool to assess prognosis early in treatment remains unaddressed in the current medical literature. Many scoring systems have been developed; however, none have proven to be entirely reliable for use in patients with NSTIs. Methods: Using collected data through a PubMed and Google Scholar search, this review provides an overview of five scoring systems—LRINEC, platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), NECROSIS, and POTTER—while highlighting potential areas for further improvement of these scoring systems or the conception of a novel, more effective system. Results: The most widely used scoring tool, the Laboratory Risk Indicator for Necrotizing Fasciitis Score (LRINEC), lacks high sensitivity and requires supplementation of other clinical parameters. The NECROSIS score offers a potentially improved system, though it lacks necessary external validation. NLR and PLR provide reliable measurements for immune response; however, they lack specificity for NSTI and require further research to determine parameters like cutoff values. The POTTER score, though not valid for use in patients with NSTI, poses a novel system utilizing AI technology and machine learning. Conclusions: This review concludes that further development of a reliable scoring system that accounts for the many factors involved in NSTI is required and may benefit from an integrative model like the POTTER score.