Endoskopie heute 2012; 25 - FV21
DOI: 10.1055/s-0032-1308741

Urine proteomic analysis differentiates cholangiocarcinoma from primary sclerosing cholangitis and other benign biliary diseases

J Metzger 1, A Negm 2, R Plentz 2, J Wedemeyer 3, H Mischak 1, M Manns 2, T Lankisch 2
  • 1mosaiques diagnostics GmbH, Hannover
  • 2Medizinische Hochschule Hannover, Gastroenterologie, Hepatologie und Endokrinologie, Hannover
  • 3Klinikum Robert Koch, Gastroenterologie, Gehrden

Aims: Diagnosis of cholangiocarcinoma (CC) and subsequent curative treatment often comes too late due to a lack of reliable early tumor markers. Patients with primary sclerosing cholangitis (PSC) are at risk to develop CC. Thus, a convincing strategy for early CC detection during PSC surveillance is urgently needed. Endoscopic procedures to diagnose CC are limited due to a low sensitivity and specificity. We recently demonstrated that bile proteomic analysis distinguishes CC from non-malignant lesions. Nevertheless, bile collection is still invasive. Urine is an attractive diagnostic source because it is less complex than blood and sampling is non-invasive. We hypothesized, that urine proteomic analysis is able to differentiate CC from other biliary disorders.

Methods: We used capillary electrophoresis mass spectrometry (CE-MS) to establish a CC-specific peptide model in patients with CC (n=16), PSC (n=15), and Non-PSC benign bile duct strictures (BBDS, n=13). Peptides were characterized by sequencing.

Results: Accuracy of classification with the CC-specific urinary peptide model consisting of 42 peptides was evaluated on an independent set of 41 CC patients (including 11 CC on top of PSC), 50 PSC and 36 Non-PSC BBDS. The model differentiated CC from BBDS with an area under the curve value of 0.86 (95% confidence interval: 0.79 to 0.91, p=0.0001) resulting in correct classification of 34 from 41 CC and 67 from 86 benign strictures (83% sensitivity, 78% specificity). All 11 patients with CC on top of PSC were correctly classified. Specificity for 74 healthy controls was 89%. Amino acid sequencing revealed that the majority of peptides are interstitial collagen fragments with some of them also detected in blood suggesting an extra-renal origin. Staining of liver biopsies indicated increased matrix metalloproteinase 1 activity on the liver epithelial cell surface of CC patients.

Conclusion: The urinary peptide model differentiates CC from benign biliary diseases and may become a diagnostic non-invasive tool for PSC surveillance and CC detection.