Bastian Schönemeier1, Jochen Metzger, Julie Klein, Holger Husi, Birgit Bremer, Nina Armbrecht, Mohammed Dakna, Joost P Schanstra, Jonas Rosendahl, Johannes Wiegand, Mark Jäger, William Mullen, Benjamin Breuil, Ruben R Plentz, Ralf Lichtinghagen, Korbinian Brand, Florian Kühnel, Harald Mischak, Michael P Manns, Tim O Lankisch. 1. From the *Department of Gastroenterology, Hepatology and Endocrinology, Hannover Medical School; and †Mosaiques Diagnostics GmbH, Hannover, Germany; ‡Institute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, United Kingdom; §Institut National de la Santé et de la Recherche Médicale (INSERM), U1048, Institute of Cardiovascular and Metabolic Disease; and ∥Université Toulouse III Paul-Sabatier, Toulouse, France; ¶Department of Gastroenterology, University of Leipzig, Leipzig; and #Department of Visceral Surgery, Hannover Medical School, Hannover, Germany; **Plateau de Protéomique des Liquides Biologiques, Institute of Cardiovascular and Metabolic Disease, Toulouse, France; and ††Department of Internal Medicine I, Medical University Hospital, Tübingen; and ‡‡Institute of Clinical Chemistry, Hannover Medical School, Hannover, Germany.
Abstract
OBJECTIVES: Differentiation of pancreatic cancer (PCA) from chronic pancreatitis (CP) is challenging. We searched for peptide markers in urine to develop a diagnostic peptide marker model. METHODS: Capillary electrophoresis-mass spectrometry was used to search for peptides in urine of patients with PCA (n = 39) or CP (n = 41). Statistical different peptides were included in a peptide multimarker model. Peptide markers were sequence identified and validated by immunoassay and immunohistochemistry (IHC). RESULTS: Applied to a validation cohort of 54 patients with PCA and 52 patients with CP, the peptide model correctly classified 47 patients with PCA and 44 patients with CP (area under the curve, 0.93; 87% sensitivity; 85% specificity). All 5 patients with PCA with concomitant CP were classified positive. Urine proteome analysis outperformed carbohydrate antigen 19-9 (area under the curve, 0.84) by a 15% increase in sensitivity at the same specificity. From 99 healthy subjects, only four were misclassified. Fetuin-A was the most prominent peptide marker source for PCA as verified by immunoassay and IHC. In silico protease mapping of the peptide markers' terminal sequences pointed to increased meprin-A activity in PCA, which in IHC was associated with neoangiogenesis. CONCLUSIONS: Urinary proteome analysis differentiates PCA from CP and may serve as PCA screening tool.
OBJECTIVES: Differentiation of pancreatic cancer (PCA) from chronic pancreatitis (CP) is challenging. We searched for peptide markers in urine to develop a diagnostic peptide marker model. METHODS: Capillary electrophoresis-mass spectrometry was used to search for peptides in urine of patients with PCA (n = 39) or CP (n = 41). Statistical different peptides were included in a peptide multimarker model. Peptide markers were sequence identified and validated by immunoassay and immunohistochemistry (IHC). RESULTS: Applied to a validation cohort of 54 patients with PCA and 52 patients with CP, the peptide model correctly classified 47 patients with PCA and 44 patients with CP (area under the curve, 0.93; 87% sensitivity; 85% specificity). All 5 patients with PCA with concomitant CP were classified positive. Urine proteome analysis outperformed carbohydrate antigen 19-9 (area under the curve, 0.84) by a 15% increase in sensitivity at the same specificity. From 99 healthy subjects, only four were misclassified. Fetuin-A was the most prominent peptide marker source for PCA as verified by immunoassay and IHC. In silico protease mapping of the peptide markers' terminal sequences pointed to increased meprin-A activity in PCA, which in IHC was associated with neoangiogenesis. CONCLUSIONS: Urinary proteome analysis differentiates PCA from CP and may serve as PCA screening tool.
Authors: Gladys N Nangami; Amos M Sakwe; Michael G Izban; Tanu Rana; Philip E Lammers; Portia Thomas; Zhenbang Chen; Josiah Ochieng Journal: Cancer Med Date: 2016-11-23 Impact factor: 4.452