Abstract
BACKGROUND AND AIMS: Preliminary single-institution data suggest that fluorescence in situ hybridization (FISH) may be useful for detecting high-grade dysplasia (HGD) and esophageal adenocarcinoma (EA) in patients with Barrett's esophagus (BE). This multicenter study aims to validate the measurement of polysomy (gain of at least two loci) by FISH as a way to discriminate degrees of dysplasia in BE specimens. METHODS: Tissue specimens were collected from four different hospitals and read by both the local pathology department ("Site diagnosis") and a single central pathologist ("Review diagnosis") at a separate institution. The specimens then underwent FISH analysis using probes 8q24 (MYC), 9p21 (CDKN2A), 17q12 (ERBB2), and 20q13 (ZNF217) for comparison. A total of 46 non-BE, 42 non-dysplastic specialized intestinal metaplasia (SIM), 23 indefinite-grade dysplasia (IGD), 10 low-grade dysplasia (LGD), 29 HGD, and 42 EA specimens were analyzed. RESULTS: We found that polysomy, as detected by FISH, was the predominant chromosomal abnormality present as dysplasia increased. Polysomy was also the best predictor for the presence of dysplasia or EA when comparing its area under the curve to that of other FISH abnormalities. We observed that if at least 10% of cells had polysomy within a specimen, the FISH probe was able to differentiate between EA/HGD and the remaining pathologies with a sensitivity of 80% and a specificity of 88%. CONCLUSIONS: This study demonstrates that using FISH to determine the percentage of cells with polysomy can accurately and objectively aid in the diagnosis of HGD/EA in BE specimens.