Reproducibility of immunohistochemical scoring for epidermal growth factor receptor expression in non-small cell lung cancer: round robin test.

CONTEXT: The addition of cetuximab to first-line chemotherapy substantially prolonged survival in patients with advanced non-small cell lung cancer whose tumors expressed high levels of epidermal growth factor receptor (EGFR; immunohistochemistry score of ≥200 on a scale of 0-300).
OBJECTIVE: To evaluate the interobserver reproducibility of this EGFR immunohistochemistry scoring system, based on both the tumor cell membrane staining intensity (graded 0-3+) and the percentage of cells staining at each intensity.
DESIGN: In parts 1 (initial feasibility study) and 2 of this 2-part round robin test, sections of different non-small cell lung cancer tissue microarrays were stained in a central reference laboratory. Following reference evaluation, EGFR expression in 30 selected tumor cores was characterized in serial sections by lung cancer pathology specialists. The reproducibility of scoring by different raters was assessed. Analysis of between-rater agreement was based on the allocation of EGFR immunohistochemistry scores into low- (<200) and high- (≥200) EGFR expression groups.
RESULTS: After discussion with raters of the issues impacting reproducibility identified in part 1 and following adjustment of processes, part 2 of the round robin test showed a high interobserver agreement in EGFR immunohistochemistry scoring, with an overall concordance rate of 90.9% and a mean κ coefficient of 0.812. Specimens with a reference EGFR immunohistochemistry score of lower than 200 and of 200 or higher showed mean concordance rates of 94.7% and 85.6%, respectively.
CONCLUSIONS: After appropriate training, assessing EGFR expression by this immunohistochemistry-based method allowed a highly reproducible allocation of non-small cell lung cancers into clinically relevant high- or low-EGFR expression groups.