Targeted multiprobe fluorescence in situ hybridization analysis for elucidation of inconclusive pancreatobiliary cytology

BACKGROUND: Endoscopic fine-needle aspiration (FNA) and brush cytology are standard methods for the diagnosis of pancreatobiliary malignancies. Although the majority of cytological diagnoses are straightforward, there remains a difficult category of inconclusive cytology. This study explored the utility of fluorescence in situ hybridization (FISH) to improve the diagnostic stratification between reactive and malignant cells in cases of inconclusive cytology. METHODS: The multiprobe FISH assay UroVysion was used for copy number assessment of chromosomes 3, 7, 17, and the 9p21 locus on Papanicolaou-stained specimens with a diagnosis of inconclusive cytology (n=50), adenocarcinoma (n531) and no evidence of malignancy (n=9). The target cells were photographed and their coordinates saved on an automated stage prior to hybridization. A positive test was defined as increased copy number (> 2) of at least 2 chromosomes (3, 7, or 17) in at least 4 atypical cells, or loss of 9p21 in at least 12 cells. RESULTS: FISH confirmed all 31 cytological diagnoses of pancreatobiliary adenocarcinomas, and was negative in the 9 patients with negative cytology. Among the 50 cases with inconclusive cytology, FISH detected 19 of 31 cases with a final diagnosis of adenocarcinoma, and was negative in all 19 cases with no final evidence of malignancy (sensitivity of 61.3%, specificity of 100%, positive predictive value of 100%, negative predictive value of 61.3%). Loss of 9p21 was found in 43 (86%) of all 50 FISH-positive cases. CONCLUSIONS: Multiprobe FISH combined with automated relocation of atypical cells is a powerful technique to clarify inconclusive cytology of the pancreatobiliary tract, allowing for a better distinction between reactive atypia and malignancy.