# Clinical significance of elevated tumor markers in patients with biliary pancreatitis

## TL;DR

This study examines how elevated tumor markers in biliary pancreatitis patients can help predict disease severity and improve treatment through a machine learning model.

## Contribution

The study introduces a machine learning-based prediction model to guide treatment adjustments in mild biliary pancreatitis.

## Key findings

- Tumor markers like CA50, CA19-9, and CA125 are significantly higher in severe and moderately severe biliary pancreatitis cases.
- Hospitalization duration is an independent risk factor for disease progression.
- A machine learning model is proposed to screen variables and guide treatment for mild biliary pancreatitis.

## Abstract

This study explores the clinical significance of elevated tumor markers in patients with biliary pancreatitis. It aims to develop a machine learning-based clinical prediction model to facilitate early intervention and improve outcomes in acute biliary pancreatitis (ABP).

We collected data from patients admitted with biliary pancreatitis to the Department of General Surgery at Jiangsu University Hospital from January 1, 2016, to December 31, 2023. We recorded general patient information.

Markers including Carbohydrate Antigen (CA) 50, CA19-9, CA125, CA724, CA242, ferritin, leukocyte count, high-sensitivity C-reactive protein (HS-CRP), total bilirubin, direct bilirubin, alanine aminotransferase, and aspartate aminotransferase were significantly higher in the severe acute pancreatitis (SAP) and moderately severe acute pancreatitis (MSAP) groups compared to the mild acute pancreatitis (MAP) group (P < 0.05). Univariate logistic regression analysis identified white blood cell count, HS-CRP, CA50, CA19-9, CA125, urinary amylase, total bilirubin, aspartate aminotransferase, and hospitalization duration as risk factors for progression to MSAP or SAP. Multivariate logistic regression analysis confirmed hospitalization duration as an independent risk factor.

Elevated tumor markers have clinical significance in biliary pancreatitis. We propose a clinical prediction model based on machine learning to screen variables and guide treatment adjustments for MAP.