Pooled analysis of the prospective trials of gefitinib monotherapy for EGFR-mutant non-small cell lung cancers

Abstract

Purpose

Epidermal growth factor receptor (EGFR) mutations have been found in the majority of gefitinib-responsive non-small cell lung cancer (NSCLC) patients from retrospective studies. We sought to compile the available phase II and prospective trials of this EGFR tyrosine kinase inhibitor (TKI) to better understand the efficacy and safety of selecting patients to receive gefitinib based on their genotype.

Design

We searched published trials involving EGFR-mutant patients and gefitinib. Five reports were identified (published between June 2006 and April 2007) in which gefitinib was given in a prospective manner to EGFR mutation positive patients at a dose of 250 mg a day. Responses were determined by RECIST and toxicities by NCI-CTC.

Results

A total of 101 patients were pooled from these studies. 59 received gefitinib as their first line of therapy and 42 after having received chemotherapy. The combined rate of complete and partial response (CR+PR) in the 99 measured patients was 80.8% (80/99) and only 7.1% (7/99) had progressive disease as best response. The response rate (CR+PR) for exon 19 deletion and L858R patients were 80.3% (53/66) and 81.8% (27/33), respectively. The median progression-free survival ranged from 7.7 to 12.9 months. Overall survival had not been reached in 4/5 reports and was 15.4 months in one of them. Gefitinib administration was safe (<50% of patients developed grade 1-2 skin rash or diarrhea) and interstitial lung disease was only reported in 2 patients (2%), without deaths.

Conclusions

Gefitinib monotherapy leads to objective responses in most patients with EGFR mutations. Both L858R and deletion 19 mutations derived similar clinical benefits. Small molecule TKIs are the new treatment paradigm for EGFR-mutant NSCLC.

INTRODUCTION

Lung cancer is the leading cause of malignancy-related death in the United States [1] and the world. Patients diagnosed with metastatic and advanced non-small cell lung cancer (NSCLC) have a dismal prognosis that seldom reaches over 1 to 2 years. The use of combination platinum-based chemotherapeutics has been the cornerstone of palliative treatment of advanced NSCLC [2]. However, even the addition of the vascular endothelial growth factor monoclonal antibody bevacizumab to a platinum-based doublet was only able to achieve response rates (RR) of around 30% in the phase II [3] and III trials [4] and the progression-free survival (PFS) did not exceed 7.5 months. Median survival was barely above 12 months [4]. It is increasingly necessary, in this heterogeneous disease, to identify sub-sets of NSCLC patients who can receive tailored therapies to improve outcomes.

In 2004, three groups of investigators identified somatic mutations in the tyrosine kinase domain of the epidermal growth factor receptor (EGFR) in patients with NSCLC [5–7]. The most exciting translational finding of these initial studies was the almost ubiquitous presence of these mutations in patients who had radiographic and clinical responses the specific EGFR tyrosine kinase inhibitor (TKI), gefitinib [5,6]. Subsequent population based efforts to sequence EGFR in NSCLC have consistently identified EGFR mutations in an enriched cohort of women, non-smokers, adenocarcinomas and East Asians [8,9]. The most prevalent EGFR mutations consist of small inframe deletions around the conserved LREA motif of exon 19 (residues 747–750), followed by a single point mutation in exon 21 – L858R. Both of these mutations compromise around 90% of known EGFR mutations [5,8,9].

Both cell line and mouse models of EGFR mutations demonstrate that tumor cells that harbor such mutations are exquisitely sensitive both to suppression of the EGFR driving signal or EGFR TKIs [6,10,11]. As data grows, it seems clear that EGFR-mutant NSCLCs compromise a distinct molecular class of lung cancers that display the phenomenon of “oncogene addiction” [12] and therefore have an “Achilles’ heel” in the inhibition of EGFR by small molecule tyrosine kinase inhibitors. The success of such an oncogene-targeted approach for cancer therapy has been fulfilled in chronic myeloid leukemia (CML) and gastrointestinal stromal tumor (GIST) [13,14].

Recent reviews on EGFR mutations and lung cancer have been published, both exploring the biology of this disease [15] and also the retrospective patient data gathered over the last years [16–18]. Over 268 EGFR-mutant patients have been identified in retrospective case series in which gefitinib or erlotinib (another anilinoquinazoline EGFR inhibitor) were given to patients with metastatic NSCLCs, mostly as 2^nd or 3^rd line therapies [6–8,16,19–37]. The overwhelming observation was that the objective response rate of these EGFR mutant tumors reached 78%. In the same studies, the EGFR mutation negative tumors only showed a response rate of 10% [16]. In contrast to the astonishing responses seen in these molecularly classified tumors, the phase III trials of the anilinoquinazoline EGFR inhibitors in unselected NSCLCs who had failed chemotherapy demonstrated that the benefits of TKI therapy in lung cancer are not applicable to all patients. Erlotinib, in the 2^nd or 3^rd line setting, derived a small, statistically significant, improvement in survival compared to placebo [38], but RR was only 8.9%, median PFS 2.2 months and overall survival 6.7 months in the treatment arm. In the other hand, gefitinib, versus placebo in a similar group of patients, only achieved a statistically significant survival improvement in a planned sub-group analysis of non-smoking and Asians patients and not in the complete population studied [39]. The later results led to the Food and Drug Administration (FDA)’s limitation in the use of gefitinib in unselected patients with lung cancer in the United States. Patients in the United States can only receive gefitinib now as part of investigational protocols.

Our goal in this report is to compile the current phase II and prospective trials of gefitinib monotherapy for EGFR-mutant patients.

MATERIAL AND METHODS

Literature search

We performed a systematic search of the MEDLINE and PUBMED databases to identify all recent clinical trials and prospective patient cohorts that contained EGFR mutant NSCLC patients that were treated with gefitinib. The search strategy included articles from 2004 through April of 2007 indexed under the subject headings epidermal growth factor receptor mutation, EGFR, gefitinib, Iressa, tyrosine kinase inhibitor, prospective, phase II and trial. We chose to start with the year 2004 because this marked the identification of EGFR mutations. The search did not restrict the type of publication or periodical. We did not include preliminary sets published as abstracts or meeting’s proceedings. We selected all prospective trials that included patient selection based on EGFR genotyping and that clearly described that gefitinib was offered to patients based on this genotype identification. The search was also restricted to published manuscripts in the English language.

Patient selection and EGFR genotype in the identified trials

All the identified trials had to include an institutional approved protocol for human studies and genomic analysis of stored/collected tumor tissue. The inclusion criteria for administration of gefitinib had to include the identification of EGFR mutations in exons 18 to 21 of EGFR. However, we did not exclude any of the trials based on the method of DNA isolation (if from fresh tissue or paraffin-embedded tissue) or the strategy used to enhance tumor-derived DNA (either micro-dissection or use of more sensitive polymerase chain reaction (PCR) amplification techniques). The trials identified included only adult patients, most with advanced stage IIIB or IV NSCLC. Two trials only included chemotherapy naive patients who were treated in the first line setting, while the others contained a mixed population of patients who had received prior chemotherapy for their NSCLC or were receiving gefitinib as first line therapy.

Treatment schedules, response, survival assessment and statistical analysis in the identified prospective studies

All the identified trials had the same treatment schedule for gefitinib. This medication was given orally at a dose of 250 mg a day. The primary end point of most of these studies was tumor response rate and feasibility. Progression-free survival (PFS) and overall survival (OS) were also measured, but they were not primary end points in any of the studies identified. Toxicities, as defined by the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTC, either version 2 or 3), were monitored in five trials identified.

The objective tumor response was determined by RECIST criteria (Response Evaluation Criteria in Solid Tumors) [40]. A complete response (CR) was defined as complete disappearance of all clinically detectable tumors for at least 4 weeks, a partial response (PR) as a 30% or more decrease in the sum of the longest diameter of the target lesions for a minimum of 4 weeks with no new areas of tumors, and progressive disease (PD) defined as at least an increase in 20% in the sum of the longest diameter of the target lesions or a new tumor lesion. Stable disease (SD) was defined as any measurement in between a PR and PD for at least 8 weeks. Response rates (RR) were defined as CR+PR. The trials that monitored for PFS and OS used the date of enrollment as the starting point of measurement. Median PFS and OS were calculated using the Kaplan-Meier method [41] and differences based on type of mutation compared by the log-rank test [42]. Fisher’s exact test was used to compare response rates in the different types of mutations and the two-tailed probability reported.

We used direct data as extracted from the author’s publications for response rate, PFS and OS. However, for 1-year survival rates, when not available, we extrapolated the results based on the available Kaplan-Meier survival curves provided.

RESULTS

Identification of five prospective trials of patients with EGFR mutations treated with gefitinib

Based on our search criteria, we identified six reports that prospectively evaluated the effects of gefitinib monotherapy for advanced NSCLC based on the presence of an EGFR tyrosine kinase domain mutation in the patient’s tumor specimens [43–48]. All trials were published between 2006 and 2007. Five of these stemmed from Japanese and one from European groups. Since the European trial [48] only had 3 patients in which gefitinib was given prospectively and no data on survival or toxicity, we decided to exclude this trial from further analyses. Table 1 summarizes the 5 identified Japanese-based studies. The characteristics of these EGFR-mutant patients in these publications [43–47] were similar among the different studies. Most patients were women, nonsmokers and had adenocarcinomas (Table 2).

Table 1.

Characteristics of the prospective trials of gefitinib monotherapy for advanced EGFR mutation positive non-small cell lung cancers (NSCLC)

Publication source and year (reference)	Gefitinib 250 mg a day treatment	Number of patients (1st line/2nd line)	Country of origin	EGFR mutation analysis
Inoue 2006 (43)	1st line therapy (chemotherapy naive patients)	16 (16/0)	Japan	PCR amplification and sequence of exons 18–23
Asahina 2006 (44)		16 (16/0)	Japan	PCR amplification and sequence of exons 18–21
Sutani 2006 (45)	Mixed 1st and subsequent lines of therapy (both patients who had received or not prior chemotherapy)	27 (4/23)	Japan	Peptide nucleic acid-locked nucleic acid (PNA-LNA) clamp of exons 18–21
Yoshida 2007 (46)		21 (12/9)	Japan	Common fragment analysis of PCR amplification for exon 19 deletions and Cycleave real-time PCR for the L858R mutation
Sunaga 2007 (47)		21 (11/10)	Japan	PCR amplification and sequence of exons 19–21
Total **	-	101 (59/42)	Japan	-

EGFR, epidermal growth factor receptor; PCR, polymerase chain reaction

^*one of the patients received both gefitinib and erlotinib

^**Another prospective study by van Zandwijk et al. [48] was exluded from this analyses due to the small number of patients that received gefitinib (n=3) and lack of survival data.

Table 2.

Clinical characteristics of the EGFR-mutant patients (also includes EGFR genotyped patients that never received gefitinib as part of the prospective treatment)

Source (reference)	Gender (female/male)	Histology (adenocarcinoma/other)	Smoking status (non-smoker/smoker)*
Inoue (43)	21/4	24/1	22/3
Asahina (44)	15/5	19/1	15/5
Sutani (45)	23/15	33/5	25/13
Yoshida (46)	17/10	27/0	17/10
Sunaga (47)	17/4	21/0	19/2
Total (n=131)	93/38	124/7	98/33

EGFR, epidermal growth factor receptor

^*not all trials defined either lifetime non-smokers vs former smokers vs smokers, however most detailed non-smokers vs eversmokers. This separation was used for this table.

Overall 101 patients with EGFR mutations received gefitinib. Two of the reports were phase II trials of first line gefitinib therapy for chemotherapy naïve EGFR-mutant stage IIIB or IV NSCLCs [43,44], while the remaining four contained a mixed population of patients, which were either receiving gefitinib as first line therapy or after tumor exposure to chemotherapy (second and subsequent lines of therapy). Overall, 59 patients received gefitinib as first line therapy and 42 after having received chemotherapy (Table 1).

The method of identification of EGFR genotypes is listed in Table 1. Most of the trials isolated tumor DNA from either diagnostic paraffin-embedded or fresh tissue (pleural fluid). Most of the reports identified the need to use micro-dissection techniques; however the details of the percentage of tumors where this was necessary are not disclosed. Four of the trials used direct polymerase chain reaction (PCR) amplification of DNA using EGFR primers that spanned between exons 19 and 21. One of the studies [45], used a more sensitive method for mutation identification, called peptide nucleic acidlocked nucleic acid (PNA-LNA) PCR clamp that has been reported to identify EGFR mutations in the presence of 100-fold background levels of wild-type EGFR from normal cells [45]. And the remaining trial only analyzed for the presence of exon 19 deletions or L858R [46].

Almost all of the patients available for analysis contained either exon 19 in frame deletions or the L858R mutation. In none of the patients was the TKI-resistant T790M mutation [49] identified or reported during the initial sequencing of the tumors.

Response rates in the total population and by different EGFR mutations

Of the 99 patients that were evaluated for a response in the five identified trials, 80 (80.8%) achieved either a complete or partial response (CR+PR). Another 12 (12.1%) had SD and only 7 (7.1%) displayed PD as best response (Table 3). In the phase II trials that evaluated only chemotherapy naïve patients, the combined CR+PR rates were 75% [43,44].

Table 3.

Overall responses in the 99 EGFR-mutant evaluated patients

Patient population	Source (reference)	CR+PR (%)	SD (%)	PD (%)	Total
Chemotherapy naïve (1st line therapy)	Inoue (43)	12 (75%)	2 (12.5%)	2 (12.5%)	16
	Asahina (44)	12 (75%)	1 (6%)	3 (19%)	16
Mixed 1st and subsequent lines of therapy	Sutani (45)	21 (78%)	5 (18.5%)	1 (3.5%)	27
	Yoshida (46)	19 (90.5%)	1 (4.8%)	1 (4.8%)	21
	Sunaga (47)	16 (84%)	3 (15%)	0 (0%)	19*
Total		80 (80.8%)	12 (12.1%)	7 (7.1%)	99

CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease

^*in the prospective trial by Sunaga et al. (47) 21 patients were entered, however 2 discontinued treatment and only 19 were available for analysis of overall response to gefitinib therapy

All of the trials identified objective response rates in the different types of EGFR mutations. As indicated in Table 4, the total number of patients with exon 19 deletions was 66 (around 67% of the patients evaluated for a response) and the other 33 patients had the L858R mutation (around 33% of total). This distribution of EGFR mutations had been expected by the retrospective studies of EGFR sequencing in non-treated patients [8,9].

Table 4.

Responses rates (combined CR+PR) by type of EGFR mutation

	Source (reference)	EGFR Mutation		p-value*
		deletion 19	L858R
CR+PR/total (%) in 1st line therapy	Inoue (43)	6/9 (67%)	6/7 (86%)	0.585
	Asahina (44)	10/13 (77%)	2/3 (67%)	1
CR+PR/total (%) in mixed 1st subsequent lines of therapy	Sutani (45)	15/20 (75%)	6/7 (86%)	1
	Yoshida (46)	7/8 (87.5%)	12/13 (92.3%)	1
	Sunaga (47)	15/16 (94%)	1/3 (33%)	0.050
CR+PR/total (%) in the combined group	Total (99 patients)	53/66 (80.3%)	27/33 (81.8%)	1

EGFR, epidermal growth factor receptor; CR, complete response; PR, partial response; SD, stable disease; PD, progressive

^*two-tailed probability by Fisher’s exact test

The percentage of patients achieving either a CR or PR for exon 19 deletions and L858R mutations were 80.3% and 81.8%, respectively (Table 4). In only one of the trials did the exon 19 deletion cohort have a significantly higher response rate than the L858R patients, however even in that trial the investigators noted that the L858R patients with SD by RECIST had measurable tumor responses that did not qualify as PR [47]. Overall, in these prospectively collected patients the L858R patients were as likely as the exon 19 deletion ones to derived responses from gefitinib monotherapy.

Progression-free and overall survival in EGFR-mutant patients treated with gefitinib

The follow-up of these patients was short. In the trials that reported median follow-up, it ranged from 7.6 to 12.7 months (Table 5). In only one of the reports did the authors obtain enough number of deaths to calculate median overall survival [45]. In all the other trials, median overall survival had not been reached at time of publication.

Table 5.

Progression-free survival (PFS) and overall survival (OS)

	Source (reference)	Median PFS in months (95% CI)	1-year survival (%)	OS in months	Median follow-up (months)
1st line therapy	Inoue (43)	9.7 (7.4–9.9)	-	not reached	7.6
	Asahina (44)	8.9 (6.7–11.1)	83%	not reached	12.7
Mixed 1st and subsequent lines of therapy	Sutani (45)	9.4 (-)	-	15.4	-
	Yoshida (46)	7.7 (-)	80%	not reached	-
	Sunaga (47)	12.9 (-)	85%	not reached	12.6

PFS, progression-free survival; OS, overall survival; -, not reported in the published material

The five publications reported PFS (Table 5). In the phase II first line therapy trials, the PFS were 9.7 and 8.9 months [43,44]. In the prospective cohorts with a mixed treatment population, the PFS ranged from 7.7 to 12.9 months [45–47]. The 1-year survival, as derived from the investigators or extrapolated by the Kaplan-Meier curves, ranged from 80 to 85% (Table 5).

It was not possible to extract PFS and OS for the individual types of EGFR mutations from 3 out of the 5 trials. In the phase II trial by Asahina and colleagues [44], the median PFS for exon 19 deletion patients was 8.3 months [44,50] and not reached for the L858R patients (of the three L858R patients, 2 were alive and progression-free after more than 11.7 months of therapy). The other only trial that reported PFS by mutational status was that of Yoshida and colleagues [46]. They noted a median PFS of 7.8 months (95% CI, 7.6-not reached) for exon 19 deletions (n=8) and 6.0 months (95% CI, 2.6 to 7.7 months) for the L858R group (n=13) (log rank test, p=0.04). No other results can be extracted from the other trials regarding differential PFS by mutation type.

Gefitinib-related adverse events reported in these prospective trials

Adverse events were reported based on the NCI-CTC guidelines. A total of 101 patients had their common grades 1 to 4 toxicities reported (Table 6). The most common toxicities were those related to the skin. The trials reported them as rash, skin toxicity, skin pruritus or dry skin. 46.5 % of the patients developed grade 1 or 2 skin side-effects, however only 3% were reported as having serious grade 3 or 4 skin toxicities. The second most common adverse event was diarrhea, with 23% of patients reporting grade 1-2 and 2% grade 3-4, respectively. Other less common effects reported, but not displayed in Table 6, included elevations in aspartate and alanine aminotransferases, fatigue and anorexia [43–48].

Table 6.

Gefitinib-related adverse events

	Skin		Diarrhea		ILD
Source (reference)	Grades 1-2/no. pts (%)	Grades 3-4/no. pts (%)	Grades 1-2/no. pts (%)	Grades 3-4/no. pts (%)	Grades 1-4/no. pts (%)
Inoue (43)	12/16 (75%)	0/16 (0%)	3/16 (19%)	0/16 (0%)	0/16 (0%)
Asahina (44)	7/16 (44%)	1/16 (6%)	7/16 (44%)	0/16 (6%)	1/16 (6%)
Sutani (45)	9/27 (78%)	1/27 (18.5%)	5/27 (3.5%)	0/27	0/27 (0%)
Yoshida (46)	6/21 (28.5%)	0/21 (0%)	6/21 (28.5%)	2/21 (9.5%)	0/21 (0%)
Sunaga (47)	13/21 (62%)	1/21 (4.7%)	2/21 (9.5%)	0/21 (0%)	1/21 (4.7%)
Total	47/101 (46.5%)	3/101 (3%)	23/101 (23%)	2/101 (2%)	2/101 (2%)

no. pts, number of patients; ILD, interstitial lung disease

The most dreaded adverse effect associated with TKI exposure in NSCLC, interstitial lung disease (ILD), was noted in only 2% of the treated patients. In 1 patient this was classified as grade 3 toxicity and the patient recovered after steroids [47]. The other ILD patient (grade 1) recovered without need of steroids [44]. No deaths were attributed to ILD or gefitinib-treatment in these 101 patients.

DISCUSSION

The discovery of EGFR mutations has sparked an input of renewed interest in lung cancer translational research. Our current report pooled from five prospective cohorts, of Japanese patients, who were genotyped for EGFR mutations and selected to receive gefitinib based on the presence of an activating somatic mutation. We believe that our pooled analysis, combining 101 patients (99 of which had their tumor’s radiographic response measured by RECIST), is able to strengthen the individual observations of each of these small prospective and phase II studies alone. The primary outcome measured in all trials was objective tumor response rate and therefore they have not been powered or followed long enough to generate detailed information on patient survival.

A great majority, over 80%, of the 99 patients displayed either a CR or PR as best response to 250 mg a day of gefitinib. These unheard of results in NSCLC confirm the retrospective RR of 78% previously reported in 268 EGFR-mutant patients [16]. Gefitinib monotherapy in EGFR-mutated patients has been described as a “weapon of mass destruction” [51] and our compilation reaffirms this concept.

Due to the large number of homogenous patients in the group described here we were also able to evaluate the predictive role of the two most common EGFR mutations. Two well conducted retrospective reports from the Dana-Farber/Harvard Cancer Center (DF/HCC) [52] and the Memorial Sloan-Kettering Cancer Center [53] combining 32 (22 exon 19 deletions, and 10 L858R) and 34 patients (23 exon 19 deletions, and 11 L858R), respectively, indicated that the response rates, progression-free and overall survival of gefitinib or erlotinib treated individuals was superior in exon 19 deletions than the L858R mutation patients [52,53]. Two retrospective East Asian cohorts did not show significant differences in survival between the gefitinib-treated L858R and exon 19 deletions [25,28,50], and a recent Japanese retrospective report actually described an improved survival for L858R-bearing patients after gefitinib therapy [54].

From our data, it is clear that, in the 99 patients reviewed here, the response rates were almost identical for patients with either exon 19 deletions or L858R mutations – 80.3% and 81.8%, respectively (p=1). Of note, all of the patients described in our report were Japanese. The Massachusetts General Hospital and DF/HCC Thoracic Oncology groups have completed a phase II trial of gefitinib as first line therapy for EGFR-mutant tumors (iTARGET). The results of these patients have not been reported yet, however we believe that this North American trial will help both confirm the remarkable response rates of gefitinib in a western population and also determined if L858R and exon 19 deletion mutations confer different prognostic degrees in this non-Asian ethnical groups. Despite the current controversy regarding possible clinical differences between the exon 19 and 21 mutations, we view our report as reassurance that both of these classes of mutations are highly sensitive to TKIs initially in a predominantly Asian population. The mature follow-up of the 99 patients compiled here will, in the future, permit a better understanding of patterns of survival in the different mutations.

The time to progression in the 5 trials reported here are also impressive. In the two prospective phase II studies of gefitinib for chemotherapy naïve patients the median PFS exceeded 8.9 months [43,44]. And even more striking was the result of the three other trials with a mixed population of previously treated and untreated advanced NSCLC. In two of those trials, the median PFS exceeded 9.4 months [45,47]. The overall survivals have not been reached yet, however we speculate that longer follow-up will confirm the improved survival seen in retrospective studies of patients treated with TKIs in which EGFR-mutations had been identified [19,23,52,53].

Another observation made possible by our pooled analysis was that of the safety of gefitinib administration to EGFR-mutant NSCLC patients. The most common toxicities were those of skin rash and diarrhea, which had been expected from the large phase III trial of gefitinib [39]. Almost all were reported as mild toxicities and few patients required a dose reduction or discontinuation. Most notably, the rate of ILD was only 2% and no deaths occurred in the patients in which this developed. Previous case series had identified this as a feared complication of TKI therapy. In over 1900 gefitinib-treated patients, around 3.5% developed ILD and over 1.5% died [55]; with male sex, smoking history and interstitial pneumonia identified as risk factors. Of interest, the only patient that developed a grade 3 ILD from our pooled analysis was a male smoker with an EGFR mutation [47].

Despite the incredible response rates and efficacy of gefitinib monotherapy for EGFR-mutant NSCLC, our results confirm that acquired resistance to EGFR TKI therapy is seen in most of the patients. As described above, in almost all trials the PFS did not exceed 12 months. This is a reminder that, different than CML where most patients retain a response even after 5 years of imatinib [56], EGFR-mutant tumors have a high propensity to loose their sensitivity to continuous TKI therapy. Two cohorts of patients with TKI-resistant NSCLCs, in which a second biopsy was obtained after progression, identified the secondary resistant T790M mutation [49] in around 50% of the samples [57,58]. The acquired amplification of the MET oncogene was recently reported as occurring in around 20% of TKI-resistant patients [59]. The mechanisms of resistance in the remaining tumors have not been completely clarified and very few other secondary mutations identified [57]. T790M, by introducing a bulkier amino acid side chain at this position, may affect the binding of gefitinib/erlotinib to the EGFR tyrosine kinase domain [49] and results in ineffective TKI function and sustained down-stream signaling from the mutant EGFR [60]. In vitro, T790M causes significant resistance to the effects of gefitinib but this can be partially overcome by irreversible EGFR inhibitors [60]. These findings have prompted ongoing trials of second-generation irreversible EGFR inhibitors in this patient population, such as an ongoing trial of the HKI-272 compound [61]. The clinical efficacy of this approach is still unknown. The re-administration of gefitinib in previously responsive patients that show radiographic progression has also been reported to improve symptoms and the clinical course of patients [62], suggesting a role for continued TKI use to control the non-TKI resistant clones of these “oncogene addicted” cancers. The optimal management of EGFR-mutant tumors that progress on gefitinib or erlotinib has not been established.

Our compilation of the first 101 published patients selected by their EGFR-mutant genotype to receive gefitinib monotherapy for advanced NSCLC confirms the high response rate (over 80%) of this agent in this molecularly classified population. Longer follow-up and phase III trials comparing gefitinib to conventional platinum-based chemotherapy are necessary to define the extent of improvement in survival for these patients. We believe that gefitinib will be a most valuable “weapon” in the fight against EGFR-mutant NSCLC.