IHC versus FISH for the detection of ERG rearrangements in prostate cancer.

Abstract

102 Background: ERG rearrangements are present in about 50% of all prostatic adenocarcinomas (CaP). Using ERG break-apart fluorescence in situ hybridization (FISH) probes, several classes of ERG rearrangement patterns have been described: Edel, Esplit, and 2+Edel. Class ERG 2+Edel was reported to be associated with lethal disease and significantly worse clinical outcome in CaP, while prognostic significance of other classes of ERG rearrangements and of ERG protein expression remains controversial. In this study, we found that IHC cannot discriminate between various forms of ERG rearrangements including 2+Edel, while FISH can. Methods: Formalin fixed paraffin embedded (FFPE) histological specimens from 52 Radical Prostatectomy (RP) cases were used. FISH analysis was performed using Abbott Molecular dual-color break-apart probes for the 5' and 3' regions of the ERG gene. Percent of overall ERG abnormalities and percent of Class 2+Edel per specimen were calculated. ERG expression was evaluated using DAKO FLEX Monoclonal Rabbit Anti-Human ERG Clone EP111. For each slide, H scores (H score =intensity X percentage of positive cells) from both tumor areas and benign prostate epithelium areas were assigned. The IHC score was defined as “H score from tumor minus H score from benign prostate epithelium”. IHC score > 0 was considered positive. Statistical analysis was conducted using JMP 10. Results: Of the 52 RP cases, 9 had the ERG 2+Edel rearrangement (cutoff >2%), 25 had Edel or Esplit rearrangements (cutoff >45%), and 27 had no ERG rearrangements by FISH. The overall Percent Agreement (OPA) between the ERG IHC and FISH (any class of ERG rearrangements) was 90% with 95% Confidence Interval (CI) (79%, 96%). However, for the ERG 2+Edel group, the OPA between the ERG IHC and FISH 2+Edel was 71%, and 62% of IHC positives were false positive. Thus IHC does not discriminate between the 2+Edel, Esplit or Edel forms of ERG while FISH identifies each class of rearrangement. Conclusions: IHC assay allows the detection of overall ERG abnormalities, however it does not discriminate ERG Class 2+Edel, which is consistently reported to be of prognostic significance. Further studies are required to confirm prognostic utility of ERG rearrangements and protein expression in CaP.