A UK NEQAS ring study evaluating observer variation in the diagnosis of HER2 amplification using the Kreatech™ HER2 FISH probe.

Abstract

Abstract Abstract #3005 Background: Whilst HER2 testing is well established as a key requirement for all newly diagnosed breast cancers the quality and consistency of HER2 testing remains a point of controversy. Most countries use FISH or ISH based testing for HER2 gene amplification in equivocal (IHC 2+) cases. Data on reproducibility and portability of novel and even established HER2 methods is sparse. We report on a multicentre evaluation of intra and inter-observer/site variation in scoring HER2 results from 45 breast cancers measured in 8 laboratories using the Kreatech™ HER2 FISH probe.
 Methods: A commercial tissue micro-array (TMA; Stretton Scientific UK) containing 2 replicate cores from 45 breast cancers was circulated to 8 laboratories with varying experience in FISH methodology. Five laboratories (1-5) were UK NEQAS reference labs, 3 further experienced routine diagnostic labs (A-C) participated in the scheme designed to test the site to site reproducibility of the Kreatech™ FISH probe. All laboratories FISHed a 90 core TMA and 7/8 reported results for all assessable cores, 1 lab reported 1 result per evaluable cases (as per the original study design). Results were collated centrally and analyzed for intra and inter-observer between both “reference” laboratories (labs 1-5) and non-reference laboratories (A-C).
 Results: Overall 81% of cores (range 70-94%) were successfully analysed by Kreatech™ FISH. Intra-observer variation was lower than expected at 4.7% (range 3.5-8.4%) for HER2 ratio. Mean intersite variation in HER2:Chr 17 ratio was significantly lower (8.5%) for reference vs non-reference (13.4%) laboratories (p=0.0013). Of 546 reported cores only 7 (1.3%) were discordant (with 4 borderline). Analysis of site to site variation highlighted possible systematic differences in scoring approaches in 3/8 laboratories including a “reference laboratory”.
 Conclusions: Intra- and inter-observer variation of absolute HER2 ratios appears to be tightly controlled across 8 laboratories using Kreatech™ HER2 FISH probes. Detailed comparisons of scoring results suggests some systematic differences between a few laboratories in the interpretation of FISH. However the majority of laboratories show extremely tight reproducibility of numeric HER2 FISH results. This study highlights the potential for extremely robust and quantitatively reproducible FISH results from both “diagnostic” and “reference” laboratories. However it also suggests continued quality assessments are essential to good performance. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 3005.