AI-Assisted, Literature-Informed Development and Retrospective Validation of a Point-Based Surgical Site Infection...

Area of Science:

• Neurosurgery
• Infectious Disease Epidemiology
• Artificial Intelligence in Medicine

Background:

• Postoperative surgical site infections (SSI) are a significant complication in spine surgery, leading to increased morbidity and healthcare costs.
• Existing risk calculators often lack applicability and accuracy for the specific patient population undergoing spine surgery.
• There is a need for improved tools to identify patients at high risk of SSI for targeted interventions.

Purpose of the Study:

• To introduce and validate a literature-informed, point-based SSI risk calculator for spine surgery patients.
• To utilize AI (GPT-4) for literature synthesis and variable weighting in calculator development.
• To assess the calculator's performance in discriminating and calibrating SSI risk.

Main Methods:

• Retrospective cohort study of adult patients undergoing spine surgery (2019-2025) at a single academic center.
• Development of a 29-variable AI-assisted risk calculator incorporating demographic, clinical, laboratory, and surgical factors.
• Evaluation of calculator performance using ROC analysis, bootstrap validation, calibration plots, and decision curve analysis.

Main Results:

• The AI-assisted calculator demonstrated strong discrimination (AUC 0.7978) and excellent calibration.
• An optimal threshold of 17 points yielded 61.0% sensitivity and 84.2% specificity for SSI prediction.
• Patients scoring ≥17 had a significantly higher risk of infection (OR 8.31).

Conclusions:

• The internally validated, AI-assisted SSI calculator shows robust performance in predicting SSI risk in spine surgery.
• The tool can complement standard preoperative assessments by identifying high-risk individuals.
• Prospective external validation is recommended to confirm clinical utility.