Evaluation of circulating tumor DNA as a prognostic biomarker for metastatic pancreatic adenocarcinoma

PurposePancreatic cancer is an aggressive solid tumor with a severe prognosis. Although tumor biomarkers are often used to identify advanced pancreatic cancer, this is not accurate, and the currently used biomarkers are not indicative of prognosis. The present study evaluated circulating tumor DNA (ctDNA) as a biomarker for prognosis prediction and disease monitoring in metastatic pancreatic adenocarcinoma (PAC). MethodsFrom 2017 to 2018, 40 patients with metastatic PAC were enrolled, and tumor tissue and blood samples were collected from 40 and 35 patients, respectively. CtDNA was sequenced by next-generation sequencing (NGS) with a 425-gene capture panel. The association of clinical characteristics, laboratory indicators, and dynamic ctDNA with patient outcomes was analyzed. ResultsMutations in KRAS (87.5%, N = 35) and TP53 (77.5%, N = 31) were most common in 40 tumor tissue. Patients' ECOG score, CA19-9, CEA, neutrophil-lymphocyte ratio (NLR), platelet- lymphocyte ratio (PLR) levels and mutations in >= 3 driver genes were strongly correlated with patients' overall survival (OS). Patients' gender, ECOG score, CA19-9, and CEA levels were associated with progression-free survival (PFS) (P<0.05). In 35 blood samples, univariate analysis showed a significant association between ECOG score, CA19-9, KRAS or CDKN2A mutation in ctDNA and OS and between CA19-9, CDKN2A or SMAD4 mutation in ctDNA and PFS. Cox hazard proportion model showed that patients' CDKN2A mutation in ctDNA (HR=16.1, 95% CI=4.4-59.1, P<0.001), ECOG score (HR=6.2, 95% CI=2.4-15.7, P<0.001) and tumor location (HR=0.4, 95% CI=0.1-0.9, P=0.027) were significantly associated with OS. Patients' CDKN2A mutation in ctDNA (HR=6.8, 95% CI=2.3-19.9, P=0.001), SMAD4 mutation in ctDNA (HR=3.0, 95% CI=1.1-7.9, P=0.031) and metastatic organ (HR=0.4, 95% CI=0.2-1.0, P=0.046) were significantly associated with PFS. Longitudinal changes in gene mutation allelic frequency (MAF) value were evaluated in 24 patients. Detection of progression disease (PD) by ctDNA was 0.9 months earlier than by radiological imaging (mean PFS: 4.6m vs 5.5m, P=0.004, paired t-test). ConclusionsThe ctDNA has the potential as a specific survival predictive marker for metastatic PAC patients. Longitudinal ctDNA tracking could potentially help identify disease progression and be a valuable complement for routine clinical markers and imaging.