Diagnostic Accuracy of Artificial Intelligence Models in Computed Tomography Interpretation of Chronic Rhinosinusitis: A Systematic Review

Key Points

• To evaluate the diagnostic accuracy of AI models in interpreting CT scans for chronic rhinosinusitis (CRS).
• Conducted a systematic review of AI models based on ResNets for CT interpretation.
• Assessed the methodological quality and heterogeneity across studies.
• Reviewed the integration of clinical data to improve model generalizability.
• AI models showed promising diagnostic accuracy for CRS.
• High variability in study methodologies limited comparability of results.
• Need for standardized multicenter validation was identified to enhance reliability.