Aims Fluorescence in situ hybridisation (FISH) increases the sensitivity for detecting pancreatobiliary tract cancer over routine cytology. In this study, diagnostic accuracy and costs of cytology and FISH in detecting cancer in patients with jaundice with biliary strictures were assessed.

Methods Brushing specimens from 109 patients with jaundice were obtained during endoscopic retrograde cholangiopancreatography and examined by cytology and FISH. The specimens were considered FISH-positive for malignancy if at least five polysomic cells or 10 cells with homozygous or heterozygous 9p21/p16 deletion were detected. Definitive diagnosis of the stricture as benign or malignant relied on surgical pathology (45 cases) or clinical-radiological follow-up >18 months (64 cases). We calculated costs of cytology and FISH based on the reimbursement from the Piedmont region, Italy (respectively, €33 and €750).

Results Ninety of 109 patients had evidence of malignancy (44 pancreatic carcinomas, 36 cholangiocarcinomas, 5 gallbladder carcinomas, 5 other cancers), while 19 had benign strictures. Routine cytology showed 42% sensitivity, but 100% specificity for the diagnosis of malignancy, while FISH-polysomy showed 70% sensitivity with 100% specificity and FISH-polysomy plus homozygous or heterozygous 9p21/p16 deletion showed 76% sensitivity with 100% specificity. The cost per additional correct diagnosis of cancer obtained by FISH, in comparison with cytology, was €1775 using a sequential cytological approach (ie, performing FISH only in patients with negative or indeterminate cytology).

Conclusions FISH should be recommended as the second step in detecting cancer in patients with jaundice with pancreatobiliary tract strictures and cytology negative or indeterminate for malignancy.