In Reply:

Fred R. Hirsch, Marileila Varella‐Garcia, Paul A. Bunn et al. · Journal of Clinical Oncology · 2004

We read with interest the recently published study by Hirsch et al 1 on the correlation between the expression of epidermal growth factor receptor (EGFR) and prognosis in non-small-cell lung carcinoma.A number of observations are worth mentioning.In the technical part of the study, the authors chose to use tissue microarrays (TMA).Given the real and perceived heterogeneity of EGFR distribution in tumor sections, as we routinely encounter in our practice, the use of an entire tissue section may have been more accurate in scoring the results, particularly because no study has yet validated the applicability of TMA for EGFR testing by immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH).This is especially relevant given that published studies already documented significant heterogeneity of EGFR expression when examined on tissue sections.2 Also, only a single antibody in this study was used on a relatively small number of specimens.The authors could have easily performed the studies on entire sections without significantly lengthening the interpretation time of IHC.In terms of the discussion of correlation with survival, a much larger cohort of patients is needed to address this point, especially given the uncertainty of IHC results on the TMA used in this study.This is critical given that no significant correlation has been previously found between the status of the EGFR gene and its protein product.3,4 Finally, the authors have misquoted a study that they claim shows superiority of FISH over IHC in predicting treatment response to trastuzumab.In fact, the quoted article is from a study done by Pauletti et al, 5 in which trastuzumab was not even employed.Instead, that study addressed the question of outcome analysis on patients who largely received surgical treatment alone.Pauletti et al conclude that FISH predicts outcome better than IHC.Unfortunately, those authors used profoundly flawed IHC methodology that we discussed in detail elsewhere.6 Therefore, we warn that a potentially similar problem might present itself in EGFR testing unless investigators validate their methodology prior to conducting the study, and unless the studies are conducted on large cohorts of patients.

Cited by 1 works. doi:10.1200/jco.2004.88.130 Open-access PDF

Authors

References