Characterization of the urinary metabolic profile of cholangiocarcinoma in a United Kingdom population

Munirah Alsaleh,1 Thomas A Barbera,1 Helen L Reeves,2 Matthew E Cramp,3 Stephen Ryder,4,5 Hani Gabra,1,6 Kathryn Nash,7 Yi-Liang Shen,1,8 Elaine Holmes,1 Roger Williams,9 Simon D Taylor-Robinson11Division of Surgery and Cancer, Imperial College London, London; 2Northern Institute for Cancer Research, Medical School, University of Newcastle, Newcastle upon Tyne, UK; 3Liver Unit, Derriford Hospital, Plymouth, Devon , UK; 4Nottingham Digestive Diseases Centre, University of Nottingham, Nottingham, UK; 5NIHR Biomedical Research Unit, Nottingham University Hospitals NHS Trust, Queen’s Medical Centre, Nottingham, UK; 6Early Clinical Development, IMED Biotech Unit, AstraZeneca, Cambridge, UK; 7Liver Unit, Southampton General Hospital, Southampton, Hampshire, UK; 8Department of Radiation Oncology, Chang Gung Memorial Hospital and Chang Gung University, Taoyuan, Taiwan; 9Institute of Hepatology, UKBackground: Outside South-East Asia, most cases of cholangiocarcinoma (CCA) have an obscure etiology. There is often diagnostic uncertainty. Metabolomics using ultraperformance liquid chromatography mass spectrometry (UPLC-MS) offers the portent to distinguish disease-specific metabolic signatures. We aimed to define such a urinary metabolic signature in a patient cohort with sporadic CCA and investigate whether there were characteristic differences from those in patients with hepatocellular carcinoma (HCC), metastatic secondary liver cancer, pancreatic cancer and ovarian cancer (OCA).Methods: Spot urine specimens were obtained from 211 subjects in seven participating centers across the UK. Samples were collected from healthy controls and from patients with benign hepatic disease (gallstone, biliary strictures, sphincter of Oddi dysfunction and viral hepatitis) and patients with malignant conditions (HCC, pancreatic cancer, OCA and metastatic cancer in the liver). The spectral metabolite profiles were generated using a UPLC-MS detector and data were analyzed using multivariate and univariate statistical analyses.Results: The greatest class differences were seen between the metabolic profiles of disease-free controls compared to individuals with CCA with altered acylcarnitine, bile acid and purine levels. Individuals with benign strictures showed comparable urine profiles to patients with malignant bile duct lesions. The metabolic signatures of patients with bile duct tumors were distinguishable from patients with hepatocellular and ovarian tumors, but no difference was observed between CCA cases and patients with pancreatic cancer or hepatic secondary metastases.Conclusion: CCA causes subtle but detectable changes in the urine metabolic profiles. The findings point toward potential applications of metabonomics in early tumor detection. However, it is key to utilize both global and targeted metabonomics in a larger cohort for in-depth characterization of the urine metabolome in hepato-pancreato-biliary disease.Keywords: cholangiocarcinoma, metabolomics, diagnostic biomarker