Clinical and biological markers predictive of treatment response associated with metastatic pancreatic adenocarcinoma

Abstract

Background: Chemotherapy for metastatic pancreatic adenocarcinoma (PDAC) offers limited benefits, but survival outcomes vary. Reliable predictive response biomarkers to guide patient management are lacking. Methods: Patient performance status, tumour burden (determined by the presence or absence of liver metastases), plasma protein biomarkers (CA19-9, albumin, C-reactive protein and neutrophils) and circulating tumour DNA (ctDNA) were assessed in 146 patients with metastatic PDAC prior to starting either concomitant or sequential nab-paclitaxel + gemcitabine chemotherapy in the SIEGE randomised prospective clinical trial, as well as during the first 8 weeks of treatment. Correlations were made with objective response, death within 1 year and overall survival (OS). Results: Initial poor patient performance status, presence of liver metastases and detectable mutKRAS ctDNA all correlated with worse OS after adjusting for the different biomarkers of interest. Objective response at 8 weeks also correlated with OS (P = 0.026). Plasma biomarkers measured during treatment and prior to the first response assessment identified ≥10% decrease in albumin at 4 weeks predicted for worse OS (HR 4.75, 95% CI 1.43-16.94, P = 0.012), while any association of longitudinal evaluation of mutKRAS ctDNA with OS was unclear (β = 0.024, P = 0.057). Conclusions: Readily measurable patient variables can aid the prediction of outcomes from combination chemotherapy used to treat metastatic PDAC. The role of mutKRAS ctDNA as a tool to guide treatment warrants further exploration.