DNA mutations in EGFR as detected by polymerase chain reaction (PCR) can occur in regions corresponding to the extracellular or intracellular portions of the protein. In non-small cell lung cancer, overexpression of EGFR or mutations in intracellular EGFR have been observed in 43-89% of cases (5). Others report that one quarter of NSCLC had mutations in the EGFR tyrosine kinase domain and these were associated with increased receptor expression in 75% of cases (35,36). Of the known EGFR tyrosine kinase domain mutations, greater than 90% occur as short in-frame deletions in exon 19 or as point mutations in exon 21, the latter resulting in arginine replacing leucine at codon 858 (L858R) (6). These mutations can result in constitutive activation of signal transduction pathways, leading to cell proliferation or anti-apoptosis, regardless of the presence of extracellular ligand. Two less common mutations occur at exons 18 and 21. Of note, EGFR and KRAS mutations appear to be mutually exclusive (7).

Subsequently, three landmark trials in 2004 showed that lung adenocarcinomas with mutated EGFR had significant responses to gefitinib and erlotinib (9,10,11). This was supported by other phase III clinical trials with chemotherapy plus erlotinib or gefitinib versus chemotherapy plus placebo in NSCLC patients that showed a survival improvement for EGFR mutated cases irrespective of the treatment they received (38,39). Overall, response rates to TKIs in adenocarcinomas with mutations in EGFR are in the range of 65-90% (6). Despite the increased response rates to these drugs in NSCLC cases with an EGFR mutation, they may not have an overall survival benefit (22). One study showed that presence of EGFR mutations was not associated with survival, although a trend for shorter survival was observed in the subgroup of patients harboring an exon 19 deletion (37). One earlier study reported that treatment of NSCLC patients with gefitinib improved symptoms and caused a radiologic response in only 10% of cases (44), suggesting that EGFR activation may be a minor component of the tumerogenic process.

The utility of detecting EGFR overexpression by immunohistochemistry (IHC) is controversial. In some studies, IHC has been shown to have value in predicting response to TKIs, while others have not been able to show its predictive value (14,21,22,23). There has been intense interest in developing novel antibodies that are able to identify abnormal EGFR or more reliably detect overexpression, and thus better predict response to targeted therapies (21).

Several studies have investigated how best to evaluate EGFR abnormalities. A meta-analysis and systemic review of the literature involving nearly 5000 patients with lung adenocarcinoma from 27 studies was recently undertaken in order to determine the value to EGFR tests in predicting response to targeted therapy (21). IHC, FISH, and PCR were assessed and it was concluded that all three methods significantly correlate with response to TKIs. Positive predictive values of IHC, FISH, and PCR were 6.5-82%, 11-89%, and 7-100%, respectively. The authors of this review note that there was significant variation in study methodology, and highlighted the importance of standardizing these methods. Further studies are required in order to identify the best method for selecting patients who will benefit from TKIs.

Direct sequencing of mutations by PCR has obvious advantages in that it can detect specific mutations, and thus identify the mutations known to be associated with better responses to TKIs. For example, it has been observed that tumors with EGFR exon 19 mutations have better overall survival than those with exon 21 muta tions when treated with TKIs (24,25). In addition, PCR can identify mutations that are thought to carry resistance to TKIs, including the exon 20 insertion that confers primary resistance, as well as the acquired resistance mutation T790M. Identification of a KRAS mutation also indicates primary resistance to TKIs, as these mutations are mutually exclusive with EGFR mutations.

On the other hand, not all tumors with the susceptible EGFR mutations respond to TKIs, and some tumors respond to TKIs that do not show mutations by direct sequencing. There are several potential reasons for this. It has been postulated that mutations in the tyrosine kinase domain of EGFR are early events in lung carcinogenesis, since these mutations have been found in about 50% of atypical adenomatous hyperplasia, as well as in normal lung tissue surrounding a tumor (26). Second, it is known that mutations may be missed if the sample submitted for direct sequencing contains less than 25% tumor cells (6). Thus, although PCR is able to identify the exact mutation status in many cases, it is not a perfect method for predicting TKI response, and DNA sequencing is yet to be available for routine clinical use in most laboratories.

EGFR gene copy number, as evaluated by FISH or CISH, has also been associated with response to TKIs (27). However, other studies indicate that no difference in survival is observed regarding EGFR gene status measured by FISH analysis (40,41,42,43). There continues to be ongoing debate regarding the relative importance of mutation status versus gene amplification status. A certain degree of co-existence occurs, in that about 50% of tumors with EGFR mutations show increased gene copy number, while about 75% of tumors with increased copy number contain mutations (6). In accordance with the concept that EGFR mutations occur early in lung carcinogenesis, it has been suggested that EGFR gene amplification is a later event (20,28). One recent study found that among 65 cases of never-smoking Asian women with adenocarcinoma, 80% had mutations in EGFR, of which nine showed gene amplification by FISH (20). All of the amplified cases had the exon 19 deletion mutation. Patients with tumors containing amplified EGFR were significantly more likely to have a pattern of solid growth on histology, show a strong staining pattern by IHC, and have a worse overall survival compared to patients with tumors containing non-amplified EGFR. Interestingly, amplified EGFR was not associated with acinar or BAC patterns, the latter of which has been reported to predict responsiveness to TKIs. These findings support the concept that EGFR amplification may occur after a mutation and result in a higher-grade, more aggressive tumor.