Introduction The aim of this study was to compare the reproducibility of epidermal growth factor receptor (EGFR) immunohistochemistry (IHC), EGFR gene amplification analysis, and EGFR and KRAS mutation analysis among different laboratories performing routine diagnostic analyses in pathology in The Netherlands, and to generate normative data.
Methods In 2008, IHC, in-situ hybridisation (ISH) for EGFR, and mutation analysis for EGFR and KRAS were tested. Tissue microarray sections were distributed for IHC and ISH, and tissue sections and isolated DNA with known mutations were distributed for mutation analysis. In 2009, ISH and mutation analysis were evaluated. False-negative and false-positive results were defined as different from the consensus, and sensitivity and specificity were estimated.
Results In 2008, eight laboratories participated in the IHC ring study. In only 4/17 cases (23%) a consensus score of ≥75% was reached, indicating that this analysis was not sufficiently reliable to be applied in clinical practice. For EGFR ISH, and EGFR and KRAS mutation analysis, an interpretable result (success rate) was obtained in ≥97% of the cases, with mean sensitivity ≥96% and specificity ≥95%. For small sample proficiency testing, a norm was established defining outlier laboratories with unsatisfactory performance.
Conclusions The result of EGFR IHC is not a suitable criterion for reliably selecting patients for anti-EGFR treatment. In contrast, molecular diagnostic methods for EGFR and KRAS mutation detection and EGFR ISH may be reliably performed with high accuracy, allowing treatment decisions for lung cancer.
- breast cancer
- cancer genetics
- cancer research
- comparative genomic hybridisation
- EGFR mutation
- familial cancers
- gene amplification
- image analysis
- in-situ hybridisation
- lung cancer
- molecular pathology
- proficiency testing
Statistics from Altmetric.com
Competing interests None.
Ethics approval Ethics approval was obtained.
Provenance and peer review Not commissioned; externally peer reviewed.