Identification and validation of a multivariable prediction model based on blood plasma and serum metabolomics for...

Area of Science:

• Gastroenterology and Hepatology
• Metabolomics
• Biomarker Discovery

Background:

• Chronic pancreatitis (CP) is a serious condition characterized by inflammation and fibrosis, leading to organ dysfunction, pain, and increased pancreatic cancer risk.
• Current diagnosis relies on symptoms and insensitive imaging, hindering early detection and management due to a lack of specific biomarkers.

Purpose of the Study:

• To identify and validate a metabolomic signature for the early diagnosis of chronic pancreatitis.
• To develop a potential laboratory test for CP detection.

Main Methods:

• A multivariable prediction model was developed following TRIPOD guidelines.
• Gas chromatography-mass spectrometry and liquid chromatography-tandem mass spectrometry were used to analyze ethylenediaminetetraacetic acid (EDTA)-plasma.
• A Naive Bayes algorithm identified an eight-metabolite signature, validated in independent cohorts.

Main Results:

• The identified metabolomic signature accurately detected CP with an AUC of 0.85 in the initial cohort.
• External validation in two independent cohorts (n=502) confirmed high diagnostic accuracy in both EDTA-plasma (AUC 0.85) and serum (AUC 0.87).

Conclusions:

• This study presents the first independently validated metabolomic signature for CP diagnosis in large, prospective cohorts.
• The findings support the development of the first routine laboratory test for chronic pancreatitis.
• Early and accurate diagnosis of CP can lead to improved patient outcomes and management strategies.