A Sequential Proteomic Relay Defines a Decade-long Pre-diagnostic Window for Pancreatic Cancer

Pancreatic adenocarcinoma is characterized by late-stage presentation and high mortality, largely due to the absence of biomarkers that signal disease during its prolonged preclinical phase. The aim of this study was to identify and temporally characterize circulating proteomic biomarkers that undergo systematic change years before clinical diagnosis, and to determine whether these trajectories define discrete pre-diagnostic risk windows that could enable earlier, biologically informed interception.

Using longitudinal proteomic data from the UK Biobank, we employed hinge-regression change-point modeling to identify temporal inflection points for circulating proteins. We partitioned the pre-diagnostic period into "Far" (5-10 years) and "Near" (0-5 years) windows to evaluate discriminatory performance.

We identified a sequential "relay" of protein trajectories. CTHRC1 serves as a primary early-warning signal with an inflection point 8.93 years prior to diagnosis. This is followed by a secondary rise in RELT at 2 years. The integrated proteomic model achieved an Adjusted R² of 0.434 and an area under curve (AUC) of 0.814. At an optimal probability threshold of 0.626, the panel distinguished between pre-diagnostic windows with 87.5% precision and 80.0% specificity.

Pancreatic cancer is characterized by a predictable, decade-long proteomic countdown. This staged relay model provides a biologically grounded framework for risk-stratified surveillance, extending the window for clinical action far beyond current standards.