Background and objective: Pancreatic cancer (PDAC) is characterized by very poor prognosis mostly caused by late diagnosis. Although modern imaging techniques have pushed the detection limit to lesions below 10mm. The combination of diagnostic imaging and biomarkers can distinguish between PDAC and chronic pancreatitis in only 67% (Carriere et al. J Eval Clin Pract. 2009). We have therefore conducted discovery and confirmation studies to identify metabolite plasma biomakers for the detection of pancreatic cancer and the differentiation from chronic pancreatitis. Material and method:The retrospective study was conducted in three phases: An initial pilot study on plasma samples was followed by analysis of a second plasma sample collection and a serum sample collection from pancreatic cancer, chronic pancreatitis and liver cirrhosis patients as well as blood donors. Metabolomic profiles of plasma and serum samples were generated applying a high throughput polar and lipid GC-MS and LC-MS/MS technology (MxP™ Broad Profiling). In addition, targeted platforms for steroids and lipids (MxP™ Steroid, MxP™ Lipids) were applied. Up to 477 metabolites were analyzed semi-quantitatively or quantitatively per study, 90% of them with an identified chemical structure. Statistical data analysis was done by linear models (ANOVA) on log10 transformed data considering age, gender, BMI and sample storage time as fixed effects. A panel of metabolites was selected for the creation of a diagnosis biomarker. The predictive ability of the biomarker was evaluated through the estimation of ROC characteristics and AUC values from Bootstrap-based Cross-Validation. Result within the three consecutive studies, sphingolipids (sphingomyelins and ceramides) were consistently and significantly increased in the pancreatic cancer group relative to the corresponding pancreatitis group whereas certain amino acids, amino acid related metabolites and coenzyme Q9 were consistently and significantly decreased. A multi-marker panel consisted of 10 metabolites and provided an AUC=0.85 when discriminating between pancreatic cancer and pancreatitis. When the CA19-9 data was included in the analysis, an AUC= 0.92 was reached. Conclusions: The results from the metabolomics study indicate that a plasma metabolite biomarker panel can be used to distinguish between pancreatic cancer and chronic pancreatitis with a high degree of accuracy. The most discriminating metabolites showed robustness with respect to transferability of their diagnostic potential from plasma to serum. A multicenter validation study has now been initiated to establish a diagnostic assay with a targeted negative predictive value (NPV) > 95%. Such a test would allow the exclusion of suspicious patients from further more invasive diagnostic procedures.

Citation Format: Julia Mayerle, Holger Kalthoff, Regina Reszka, Beate Kamlage, Christian Pilarsky, Robert Grützmann, Bodo Schniewind, F. Ulrich Weiss, Markus M. Lerch. Identification of plasma metabolites as biomarker candidates for the diagnosis of pancreatic ductal adenocarcinoma (PDAC). [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3501. doi:10.1158/1538-7445.AM2013-3501