Abstract 5048
Background
Due to the dismal prognosis of the pancreatic ductal adenocarcinoma (PDAC) biomarkers are needed to facilitate the early and preferably noninvasive diagnosis. The use of circulating tumor DNA (ctDNA) was already examined in the metastatic situation but not at all in a resectable stage with lower tumor load. Currently, CA19-9 is the only validated biomarker for PDAC but inherits poor performance. Recently, it has been reported that elevated levels of thrombospondin-2 (THBS2) protein are detectable in PDAC. Here we aimed at improving sensitivity and specificity of THBS2 based detection of early PDAC by combining THBS2 analysis with further markers.
Methods
39 patients with proven PDAC, enrolled to the NEONAX trial (identifier: NCT02047513), were selected for this study-independent retrospective translational analysis. 15 patients with benign pancreatic disease (IPMN) served as controls. cfDNA concentration was measured fluorometrically. KRAS genotyping of ctDNA was done by digital droplet PCR targeting the 7 most frequently occurring KRAS mutations in PDAC. CA 19-9 (Roche, cut-off 55 U/ml) and THSB2 levels (Quantikine, R&D Systems, cut-off 42 ng/ml) were determined by ELISA.
Results
PDAC patients had significantly more cfDNA (12.6 ng/ml) than IPMN patients (5.5 ng/ml, p = 0.0006). Only 5 % of PDAC patients and 7 % of IPMN patients had detectable KRAS mutations in ctDNA. CA19-9 was elevated in 56 % and THBS2 in 44 % of PDAC patients. Both markers were not elevated in any of the IPMN patients. Therefore, the assessment of THBS2 and CA19-9 levels was most suitable to discriminate the PDAC cohort from the IPMN cohort, with a sensitivity of 77 % and a specificity of 100 %.
Conclusions
THBS2 and CA19-9 panel assessed in human blood using a conventional ELISA assay may improve the diagnosis of pancreatic lesions as PDAC at an early stage. While total cfDNA amount differs between patients with benign and malignant pancreatic lesions, ctDNA genotyping for KRAS mutations failed to improve non-invasive diagnostic strategies in resectable PDAC most likely due to a low tumor load.
Clinical trial identification
NCT02047513.
Legal entity responsible for the study
Ulm University.
Funding
Has not received any funding.
Editorial Acknowledgement
Disclosure
All authors have declared no conflicts of interest.