The epidermal growth factor receptor (EGFR; HER 1; ErbB1) has been at the center of an explosion of translational research over the past 20 years. The EGFR is expressed or overexpressed in a large number of malignancies, including non–small-cell lung (NSCLC), head and neck, esophageal, gastric, colorectal, breast, prostate, bladder, renal, pancreatic, and ovarian cancers. EGFR expression is often found in association with increased expression of its ligands, most notably epidermal growth factor (EGF), tumor growth factor-alpha, or amphiregulin. EGFR activation and signaling affects many aspects of cell biology relevant to malignant transformation, including cell cycle progression, migration and invasion, differentiation, and cell survival. The antiapoptotic effects of EGFR activation may be particularly relevant to the assessment of treatment outcomes in patients afflicted with EGFR expressing tumors have been ascribed to regulation of Bcl-2 family members. In this issue, Bentzen et al examine EGFR expression status in 304 patients with available pretreatment tumor biopsy material among 918 patients randomized to Continuous Hyperfractionated Accelerated Radiotherapy (CHART) versus conventionally fractionated radiotherapy. The EGFR index was defined as the proportion of tumor cells with EGFR membrane staining. Significant benefit in locoregional tumor control from CHART was seen in patients with head and neck squamous cell carcinoma with high EGFR expression (2P .010) but not in patients with low EGFR expression (2P .85). By contrast, EGFR expression status was not linked to patient survival or rate of distant metastases. Bentzen et al used immunohistochemical analysis to assess EGFR expression levels, which raises some technical concerns. First, manual methods of immunohistochemical scoring of antigen expression as used in this report are fraught with variability. This variability may explain the lack of association of EGFR expression with other biomarkers evaluated in this study (Ki-67 index, Ki-67 pattern, p53 index, p53 intensity, bcl-2 expression, or cyclin D1 index). Another concern centers on the retrospective analysis of archived tissue stored for different periods of time. The CHART head and neck phase III trial accrued patients over 5 years from March 1990 to April 1995. During this extended period, oxidation of antigens/DNA/RNA may have occurred. A recent editorial in the Journal of Clinical Oncology by Meropol refers to work by Atkins et al in support of the notion that extended storage of tissue sections over time affects the sensitivity of EGFR immunostaining. However, despite these concerns, the study by Bentzen et al provides an interesting new perspective on intratumoral EGFR expression as a variable with obvious relevance to the design of fractionated radiotherapy. Several reports have illuminated molecular mechanisms leading to enhanced expression of the EGFR in epithelial malignancies. In addition to gene amplification, sequence analysis has revealed polymorphisms of intron 1 of the EGFR gene characterized by (CA)n dinucleotide repeats of different lengths that regulate transcription rates of the EGFR in cell lines and in tumor tissues in patients. The length of the (CA)n dinucleotide repeat tract is inversely correlated with the transcriptional activity of the EGFR gene. Further investigation of EGFR sequences expressed by tumor cells has also revealed alterations that primarily affect the activation state of the receptor rather than the expression levels. For example, in patients with colorectal cancer, a polymorphic variant of the EGFR gene (HER-1 R497K) has JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 23 NUMBER 24 AUGUST 2