Deciphering Circulating Signatures of Infected Pancreatic Necrosis

Acute pancreatitis is a common disease that can progress to gland necrosis (necrotizing pancreatitis [NP]).

NP imposes significant risk of morbidity, mortality, and substantial healthcare costs, especially when infected (infected pancreatic necrosis [IPN)]. A preemptive approach with early IPN recognition is suggested to improve patient outcomes. However, existing diagnostic tools fail to distinguish IPN from sterile NP early in the disease course.

Furthermore, related studies on potential serum markers lack sufficient methods of assessment.

The aim of this project is to develop a simple, reliable, and cost-effective biomarker panel for accelerated and improved decision-making, validated on real patient results.

First, we use information theory and machine learning to select the best-performing combination of available serum cytokines for detecting IPN.

The accuracy and prediction abilities of the biomarker panel are then verified in an independent cohort using the standardized prospective specimen collection, retrospective blinded evaluation study design for pivotal evaluation of biomarkers.

Finally, the net benefit to patients and costs by incorporating effects of clinical decisions based on the panel is assessed through a multicenter randomized controlled trial.

The project creates a critical step in the improvement of care and clinical management in millions of patients worldwide. Moreover, the systematic phased approach facilitates open science and reproducibility. As the panel moves from research to application, it offers potential for commercial exploitation.

In the longer term, if clinically validated, the test can be adopted by the larger clinical community for further research and more personalized care.