A Clinical Prediction Model to Assess Risk for Pancreatic Cancer Among Patients With Acute Pancreatitis

Area of Science:

• Oncology
• Gastroenterology
• Epidemiology

Background:

• Pancreatic cancer (PC) poses a significant health challenge, often diagnosed late.
• Acute pancreatitis (AP) is a risk factor for PC, necessitating early identification of high-risk individuals.
• Current screening strategies for PC lack targeted approaches for AP patients.

Purpose of the Study:

• To develop and validate a prediction model for identifying patients with AP at elevated risk of developing PC within two years.
• To establish a foundation for a sequential screening strategy for PC in AP populations.
• To utilize readily available clinical data for risk stratification.

Main Methods:

• A population-based retrospective cohort study was conducted using US Veterans Health Administration data.
• Individuals aged 40+ hospitalized for AP were included.
• Least absolute shrinkage and selection operator (LASSO) regression with cross-validation identified predictors for a logistic regression model.

Main Results:

• The final model included factors like gallstones, pancreatic cysts, alcohol use, smoking, and specific blood markers (bilirubin, triglycerides, etc.).
• The model achieved an area under the receiver operating curve of 0.70 (95% CI: 0.67-0.73).
• A 2% risk threshold identified 20.3% of AP patients for screening, capturing nearly 50% of AP-associated PC cases.

Conclusions:

• A validated prediction model was developed to identify patients with AP at high risk for PC.
• This model serves as the initial step in a sequential screening approach for PC.
• The model leverages common clinical factors, facilitating practical implementation in healthcare settings.