Fluorescence in situ hybridisation in the cytological diagnosis of pancreatobiliary tumours

AIMS: To assess the sensitivity, specificity, positive and negative predictive values of fluorescence in situ hybridisation (FISH) and conventional cytology in identifying bile duct stricture malignancies. METHODS: Brushing samples were collected from 64 patients by means of endoscopic retrograde cholangiopancreatography, and assessed cytologically and by means of a multi-probe FISH set. The cytological diagnoses were: positive, negative and suspicious, whereas criteria for FISH positivity were: more than five polysomic cells or more than 10 trisomic cells for chromosomes 3 or 7. RESULTS: Forty-eight of the 64 patients showed histological or clinical signs of malignancy. The sensitivity of cytology was high (77%) if suspicious cases were considered positive, but was significantly lower than that of FISH if suspicious cases were considered negative (58% versus 90%; p < 0.05). The specificity of cytology was 81% (positive and suspicious) or 100% (negative and suspicious), and the specificity of FISH was 94% (p = 1). FISH yielded one false negative result (isolated chromosome 7 trisomy). FISH allowed a definite diagnosis of 9/12 cytologically inconclusive cases. CONCLUSIONS: Our findings suggest using FISH in the case of bile duct strictures cytologically negative or inconclusive; a FISH diagnosis of malignancy should only be made in the presence of polysomic pattern.