To the Editor:

I read with interest the recently published article by Saltz et al¹ in which they reported poor correlation between the epidermal growth factor receptor (EGFr) score and response to cetuximab treatment in patients with refractory colorectal cancer. In that study,¹ one of 17 patients with 1+ EGFr status, four of 30 with EGFr 2+ status, and zero of 10 with EGFr 3+ status, were responders. The authors stated in the Discussion that immunohistochemistry for EGFr expression is a poor indicator of which tumors are most treatable by the EGFr-targeted therapy used in the study. Before accepting this conclusion, we need the authors to provide important methodological details that were lacking in their article. Answering the following questions will help us understand some possible reasons for the unexpected results they reported:

(1) The authors should name the “single reference laboratory” who performed and measured the expression of EGFr by immunohistochemistry. This is a customary practice in most publications.

(2) The authors need to clarify how EGFr status was scored. The grading method mentioned in their Methods section is certainly NOT according to the manufacturer's (DakoCytomation, Carpinteria, CA) instructions that are included in the recently released EGFrpharmDx kit. The product sheet recommends an “observational” approach to reporting the results of EGFr staining in which the overall percent of positive cancer cells, the percent of cancer cells with membranous staining, the percent of cancer cells with cytoplasmic staining, the intensity of the membranous staining, and the intensity of the cytoplasmic staining are all reported. There is no suggestion of a combined score. Therefore, the authors should explain how did they come up with the 0, 1+, 2+, and 3+ scores for EGFr expression, and what these scores mean.

(3) It will be helpful to know what percent of cancer cells were positive for membranous EGFr expression, and whether or not there were correlation between this percentage and treatment response.

(4) The instructions included with Dako EGFr pharmDx kit specifically state that “Only tissues preserved in neutral buffered formalin for routine processing and paraffin embedding are suitable for use.” Unexpected results in previous studies on other markers were later found to be due to authors’ using alcohol-based tissue fixatives or postfixation instead of neutral buffered formalin and routine processing. The authors did not provide any information on the type of tissue fixatives used, whether postfixatives were used, or whether the same tissue fixatives, reagents, and processing protocols were used in all participating institutions.

Author's Disclosures of Potential Conflicts of Interest

The author indicated no potential conflicts of interest.