Resectional surgery offers a curative intent and a survival benefit for patients with hilar cholangiocarcinoma, but is associated with high morbidity. Since morphological imaging cannot solve differential diagnosis preoperatively, in order to exclude patients inappropriate to this aggressive surgery, we evaluated the impact of functional imaging using fluorodeoxyglucose positron emission tomography (FDG PET) in the detection of cholangiocarcinoma and its usefulness in the differentiation from benign Klatskin tumour-mimicking lesions. Fifteen consecutive patients aged 47-78 years underwent standardized whole-body FDG PET with attenuation correction before potentially curative surgery using a conventional full-ring PET scanner with an axial field-of-view of 16.2 cm. FDG PET was evaluated visually and semiquantitatively using tumour-to-background ratios (T/B) ratios. All lesions were evaluated histopathologically. FDG PET presumed to be indicative for carcinoma was positive in 12 of 15 patients, true positive in 10 (T/B ratio, 3.2+/-1.9) and false positive in two of them (T/B ratios, 2.1 and 2.8) with Klatskin tumour-mimicking lesions. While all true positive PET results were seen in the tubular type of cholangiocarcinoma with a high amount of tumour cells and only low production of mucus, a false negative FDG PET in three patients was observed in mucinous adenocarcinoma. Additionally, FDG PET detected locoregional lymph nodes in two patients and distant metastases in a further three patients. Due to false positive results FDG PET does not allow the differentiation of benign from malignant lesions, and FDG PET should be avoided in patients with mucinous cholangiocarcinoma. However, FDG PET may have significant influence on the treatment strategy in as much as 20% of the patients, since it may detect distant metastases.