Mathematical Modeling of The Challenge to Detect Pancreatic Adenocarcinoma Early with Biomarkers

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive tumor type and is usually detected at late stage. Here, mathematical modeling is used to assess the feasibility of two-step early detection with biomarkers, followed by confirmatory imaging. A one-compartment model of biomarker concentration in blood was parameterized and analyzed. Tumor growth models were generated from two competing genomic evolution models: gradual tumor evolution and punctuated equilibrium. When a biomarker is produced by the tumor at moderate-to-high secretion rates, both evolutionary models indicate that early detection with a blood-based biomarker is feasible and can occur approximately one and a half years before the limit of detection by imaging. Early detection with a blood-based biomarker is at the borderline of clinical utility when biomarker secretion rates by the tumor are an order of magnitude lower and the fraction of biomarker entering the blood is also lower by an order of magntidue. Regardless of whether tumor evolutionary dynamics follow the gradual model or punctuated equilibrium model, the uncertainty in production and clearance rates of molecular biomarkers is a major knowledge gap, and despite significant measurement challenges, should be a priority for the field. The findings of this study provide caution regarding the feasibility of early detection of pancreatic cancer with blood-based biomarkers and challenge the community to measure biomarker production and clearance rates.