Development and Validation of a Nomogram Model for Severe Acute Pancreatitis Risk Stratification: A Retrospective...

Area of Science:

• Gastroenterology
• Clinical Prediction Modeling
• Biostatistics

Background:

• Early prediction of acute pancreatitis (AP) severity is vital for patient management.
• Current methods may not fully leverage readily available laboratory data for accurate risk stratification.

Purpose of the Study:

• To develop and validate a nomogram for predicting severe acute pancreatitis (SAP) using laboratory markers.
• To identify key laboratory variables predictive of SAP using LASSO regression.

Main Methods:

• Retrospective analysis of 965 adult AP patients.
• Least Absolute Shrinkage and Selection Operator (LASSO) regression for variable selection.
• Nomogram construction and performance evaluation (AUC, calibration, DCA).

Main Results:

• Identified eight independent predictors: C-reactive protein, lactate dehydrogenase, hematocrit, monocyte percentage, prothrombin time, D-dimer, glucose, and total cholesterol.
• The nomogram achieved an AUC of 0.809 (training) and 0.768 (validation).
• Model demonstrated good calibration and clinical utility via decision curve analysis.

Conclusions:

• The developed nomogram is a valuable tool for predicting AP severity.
• This model aids in clinical decision-making and patient risk stratification.