Detection of Esophageal Adenocarcinoma (EAC) by a Four-Probe Fluorescence in situ Hybridization (FISH) Assay

Abstract

Purpose: To assess the viability and performance characteristics of a four probe FISH assay for detection of EAC among esophageal biopsy specimens. Methods: A four-probe FISH assay (MYC[8q24], p16[9p21.3], HER2[17q11.2], and ZNF217[20q13.2]) was performed on a training set of 36 esophageal biopsy specimens (18 normal squamous epithelium [NSE], 18 EAC). Assay results were quantified by the percentage of cells with multiple probe gains, single probe gains, homozygous 9p21 loss, and any probe gains with any 9p21 loss. Median assay results for NSE and EAC were compared using nonparametric tests. Thresholds for neoplasia detection were selected by receiver operating characteristic analysis to achieve 90% specificity and maximum sensitivity. These thresholds were assessed in a unique test set of samples with EAC (n = 17) or nondysplastic Barrett's esophagus (NDBE) (n = 15) who did not progress in 3 years. Results: The median age (63 years) among groups was similar while the EAC group was more likely than the NSE group to be male (71 vs 33%, p = 0.04). Median assay results were significantly greater for EAC compared to NSE for multiple gains (51 vs 1%), single gains (24 vs 6%), and combined any gain with any 9p21 loss (7 vs 1%) (see table) (p<0.0001 for each). The presence of neoplasia was detected with 100% sensitivity and 90% specificity in the training set above the following cutpoints: ≥4% multiple gains, ≥16% single gains, ≥9% homozygous 9p21 loss, and ≥4% combined any gain with any 9p21 loss. In the test set, EAC assay results were significantly greater than in NDBE specimens (see Table) (p<0.05 for each). Using the previously established cutpoints, neoplasia was detected with 100% sensitivity and 100% specificity in the test set. Multiple gains alone was 100% sensitive and 100% specific in both the training and test sets.TableConclusion: DNA abnormalities detected by FISH assay were more prevalent among specimens with EAC rather than NSE or NDBE. Neoplasia was detected with high sensitivity and specificity. Future study of this assay for the detection of intermediate grades of Barrett's esophagus is warranted. Disclosure: Grant D. Carlson BS - employee: NeoGenomics, Tiffany Chouinard BS, CG(ASCP), MB(ASCP) - employee: NeoGenomics, Robert P. Gasparini MS - employee: NeoGenomics, Melinda Lux MS, CG(ASCP) - employee: NeoGenomics, A. Scott Gasparini BS - employee: NeoGenomics, William J. Bulsiewicz MD MSc - none, Theresa S. Emory MD - none, Jackie M. Makapugay MD - none, Nicholas J. Shaheen MD MPH - none. This research was supported by an industry grant from NeoGenomics.