Epstein-Barr virus (EBV) has long been implicated in Hodgkin's disease (HD) etiology based on epidemiologic evidence. In the last decade, molecular techniques (EBER in situ hybridization (ISH) and LMP-1 immunohistochemistry (IHC)) have been developed that can detect EBV gene products directly in tumor cells. These morphologic assays, now widely employed in clinical and epidemiologic studies, have identified EBV latent infection in a proportion of HD tumors. Yet, although this proportion varies considerably among studies, and assay interpretation is subjective, assay reliability has never been evaluated. This project examined inter- and intra-rater reliability of EBER ISH and LMP-1 IHC in EBV detection in 40 previously assayed HD tumors, based on first and second reviews by each of four hematopathologists of slides prepared using standard assay methods at each of two laboratories. Raters, blinded to slide identity and prior EBV status, determined EBV status of all 640 slides reviewed, and ranked interpretive difficulty in preparation for developing interpretation guidelines. Overall, kappa statistics for inter-rater agreement indicated very good (kappa=0.7-0.9) reliability for both assays. However, for EBER ISH, at least one reviewer dissented in approximately one-fifth of reviews, and extent of disagreement varied by histologic subtype. For LMP1 IHC, disagreement was less frequent and less variable. Interpretation difficulty was highly correlated with inter-observer disagreement but ranged widely among reviewers. Additional kappa statistics to address intra-rater agreement and other factors affecting interpretation will be presented. Implications of variation in interpretation accuracy to understanding extant and future studies will be discussed.
Learning Objectives: 1. To understand the extent of inter- and intra-observer reliability of the commonly used assays to detect EBV in Hodgkin's disease tumors 2. To appreciate the importance of understanding assay reliability to the interpretation of study results
Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.