How to select a patient for EGFR-TKI treatment

It was a difficult picture to conjure up a decade ago—by even the most optimistic and seasoned oncologist caring for patients with lung cancer—that a pill taken orally once a day could dramatically relieve interminable cough almost overnight, resolve numerous lesions radiographically within days and, at times, even return a morbidly sick individual with an enormous tumor burden to a productive work life. All accomplished with fairly acceptable and manageable toxicities. For sure, only a small subset of patients with advanced non–small-cell lung cancer (NSCLC) treated with epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) experienced these dramatic benefits, and even when this happened, the respite lasted generally only for months. Nevertheless, the successful introduction of two agents from this class, gefitinib and erlotinib, across the globe ensured that lung cancer would be at the front and center of drug development in the era of molecularly targeted therapies.

Astute clinicians quickly observed that striking benefits with EGFR TKIs were seen predominantly in certain subsets of patients with NSCLC (life-long never smokers, Asian people, women, and those with adenocarcinoma histology). It soon became clear that the molecular mechanism underpinning dramatic responses in these patients was the presence of activating mutations in the EGFR TK domain.1,2 Now, 5 years after the publication of these two seminal articles, an entire new field of research devoted to EGFR mutation has added a significant body of knowledge to the literature. Apart from EGFR mutation analysis, other markers, such as EGFR expression analysis by immunohistochemistry (IHC), EGFR copy number by fluorescent in situ hybridization (FISH), and K-ras mutations have also been evaluated to identify patients likely to respond to EGFR TKIs. Not surprisingly, a conflicting and confusing body of information now confronts practicing physicians. Among these four molecular markers, which one would most reliably identify patients who would respond dramatically to EGFR TKIs? Has the time arrived to incorporate molecular analyses in our routine clinical practice? Finally, what have we learned from this experience that could guide us to develop better strategies in the future?

The Iressa NSCLC Trial Evaluating Response and Survival versus Taxotere (INTEREST) represent one such attempt to address the first two issues confronting practicing physicians.3 INTEREST is a large multi-institutional international phase III study randomly assigning patients with advanced NSCLC who had received one prior cycle of chemotherapy to docetaxel or gefitinib with the primary goal of establishing noninferiority of gefitinib over docetaxel in producing improved overall survival. INTEREST clearly met its (modestly set) primary goal. No single biomarker (EGFR protein expression by IHC, EGFR copy number by FISH, EGFR TK mutation or K-ras mutation) tested in the INTEREST trial were predictive for differential survival between gefitinib and docetaxel. However, presence of EGFR TK mutation identified patients likely to have better response and longer progression-free survival (PFS) with gefitinib compared with docetaxel. A high EGFR copy number by FISH identified those who would have a greater response to gefitinib compared with docetaxel, but unlike those with EGFR mutation, it was not accompanied by a corresponding increase in PFS. Gefitinib was significantly superior to docetaxel in improving PFS in patients with EGFR TK mutations as demonstrated by an impressive hazard ratio of 0.16 (95% CI, 0.05 to 0.49; P = .001) although this was not accompanied by a similar impact on overall survival. The discordance between PFS and overall survival is likely related to poststudy therapy, with patients randomly assigned to docetaxel receiving gefitinib after developing progressive disease.

The biomarker analyses from INTEREST are similar to those from two large studies reported recently. The first one, the Iressa Pan-Asia Study (IPASS) study randomly assigned patients (never smokers, former light smokers from East Asia) with previously untreated advanced NSCLC to gefitinib or the doublet therapy of paclitaxel and carboplatin.4 The study not only met its primary objective of demonstrating noninferiority of gefitinib over chemotherapy in terms of PFS but also demonstrated its superiority (HR = 0.74; 95% CI, 0.65 to 0.85; P < .001). While there were no differences in overall survival in the entire study population, those with EGFR TK mutations had significantly improved response rates and PFS when treated with gefitinib compared with chemotherapy (9.5 v 6.3 months; HR = 0.48; P < .001) The striking biologic effect of gefitinib in EGFR TK mutation tumors was evident from the remarkable differences in PFS with or without mutation when treated with gefitinib (9.5 v 1.5 months). In patients with no EGFR TK mutations, PFS was significantly superior in the group treated with chemotherapy compared with gefitinib (HR = 2.85; 95% CI, 2.05 to 3.98; P < .001). High EGFR copy numbers were EGFR FISH was predictive for efficacy only when accompanied by the presence of concomitant EGFR TK mutation. Similar to the INTEREST study, no differences in overall survival were seen in patients with EGFR TK mutation regardless of whether they were randomly assigned to receive gefitinib or chemotherapy, once again likely from cross-over effects.

The third study of relevance here is the Sequential Tarceva in Unresectable NSCLC (SATURN) trial, which randomly assigned patients with advanced NSCLC who had stable disease after four cycles of platinum-based doublet therapy to erlotinib or placebo.5 The SATURN trial met its primary end point of improving PFS (median PFS from 11.1 weeks with placebo to 12.3 weeks with erlotinib; HR = 0.71; 95% CI, 0.62 to 0.82; P < .0001) with erlotinib. Although the study even met its coprimary end point of improving PFS among those who had EGFR expression by IHC and in all subgroups tested; the most dramatic benefit was seen in those with EGFR mutation (HR = 0.10; 95% CI, 0.04 to 0.25; P < .0001) when treated with erlotinib compared with placebo. Sequential administration of gefitinib following three cycles of platinum-based therapy did not improve overall survival (the primary endpoint) in patients with advanced NSCLC when compared with six cycles of platinum based doublet therapy in a phase III study reported by the Japanese investigators in this issue of JCO.6 Similar to the SATURN trial, early use of EGFR TKI was associated with a modest improvement in median PFS (HR = 0.68; 95% CI, 0.57 to 0.80). The design of the study, relatively smaller sample size, and lack of any biomarker data seriously limit the interpretation and applicability of the findings in routine clinical practice.

Taken together, the three recent representative studies (INTEREST, IPASS, and SATURN) underscore the fact that presence of EGFR TK mutation best identifies those who would derive the most benefit, as measured by PFS. As more and more effective subsequent lines of therapy become available, at least for a significant proportion of patients after discontinuation of study therapy, we may have to learn to live with PFS as a preferred primary end point. The cross-over effect of patients receiving "the other arm" therapy on progression clouds the interpretation of overall survival even more. Other tests, most notoriously EGFR protein expression by IHC and to a lesser extent EGFR copy number analysis by FISH and K-ras mutations have not been consistently shown to be useful in identifying those who would have a significant prolongation in PFS when treated with EGFR TK inhibitors. It is too simplistic to view that all known EGFR mutations would have similar biologic effects. The two most common mutations in the TK domain, exon 19 deletions and L858R have been associated with best responses.

The lack of overall survival benefit, even in patients with EGFR mutations, in all the three studies raises an interesting problem. Does it really matter whether we give EGFR TKIs earlier (front line/maintenance) in the course of the disease or later (in the second- or third-line setting)? In essence, does sequencing of therapy matter in patients with EGFR mutation NSCLC? Would they be better off with upfront EGFR TKI compared with chemotherapy in the front-line setting? No study has been specifically designed to address this issue in patients with EGFR mutations. On a practical note, in this specific group of patients with NSCLC, it is hard to make a case for delaying a therapy that is dramatically effective, convenient, and well tolerated compared with a more toxic intravenous cytotoxic chemotherapy, be it in the front line or subsequent lines of therapy. More importantly, it would be unwise to use EGFR TKI in the front-line setting in patients with lung cancer who harbor wild-type EGFR. One could make a case then for EGFR mutation testing, certainly when EGFR TKIs are being considered rather than chemotherapy in the front-line setting or even in the maintenance setting.

What lessons have we learned so far and how best can we move forward? It is inconceivable that molecularly targeted therapies would be effective in unselected groups of patients with NSCLC, a notoriously diverse disease at the molecular level. The era of using the molecularly targeted therapies in unselected population is drawing to a close. Designing combination therapy regimens based on poorly characterized preclinical models has repeatedly led us down wrong paths. A bewildering array of novel therapies has suddenly become available for testing in an increasingly harsh and competitive climate for clinical trials. As we unravel the molecular complexities of lung cancer, we should get used to the notion of rethinking NSCLC not as one disease, but as several subcategories with unique aberrations, uniting at best 5% of them under a distinct molecular genotype. The challenge is to get the right drug to the right population. Once toxicities and appropriate doses are defined, our first task should be to identify those who would benefit the most. The early phase nonrandomized studies should focus primarily on identifying predictive biomarkers. They should include reasonable numbers of patients (larger than the traditional phase II studies) with enough power to identify a small number who would respond well. These studies should incorporate mandatory collection of blood and tissue specimens. The optional tissue collection approach has been quite ineffective. Once a biomarker has been identified, randomized studies should focus on this susceptible population to assess the impact of new therapy and address well–thought out questions on how best to further improve efficacy (eg, sequencing, novel combinations). Finally, and more importantly, we should apply these lessons in earlier stage resected NSCLC to continually improve the cure rates in molecularly defined populations. Concurrent with these efforts, we should continue to encourage technological developments to assay these changes in the peripheral blood, as is being done with EGFR mutations, and to use functional imaging to take a snapshot of the constantly evolving tumor. The latter two approaches are crucial to get real-time, noninvasive assessment of ongoing molecular changes to guide therapy. Archived tissue specimens, embedded in paraffin and frozen in time, are unlikely to be representative of the clones surviving and thriving despite the assault of previous therapies.

An organized thoughtful approach that focuses on early biomarker identification could dramatically change the lives of our patients with cancer by improving cure rates. An empiric shot gun approach will not. The progress made in breast cancer using the former approach has led to spectacular improvements in cure rates after the use of trastuzumab in the adjuvant setting in patients who have overexpression of human epidermal growth factor receptor 2/neu in the tumor cells. The fact that we have not begun designing one randomized clinical trial in the adjuvant setting comparing EGFR TKIs with platinum-based doublet chemotherapy in patients with resected NSCLC who have EGFR TK mutations, 10 years after the first of them became available, is telling. It is time to change our approach and put the lessons learned to practice


1. Lynch TJ, Bell DW, Sordella R, et al: Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350:2129–2139, 2004.

2. Paez JG, Janne PA, Lee JC, et al: EGFR mutations in lung cancer: Correlation with clinical response to gefitinib therapy. Science 304:1497–1500, 2004.

3. Douillard JY, Shepherd F, Hirsch V, et al: Molecular predictors of outcome with gefitinib and docetaxel in previously treated non–small-cell lung cancer: Data from the randomized phase III INTEREST trial. J Clin Oncol 28:744–752, 2010.

4. Fukuoka M, Wu Y, Thongprasert S, et al: Biomarker analyses from a phase III randomized, open-label, first-line study of gefitinib (G) versus carboplatin/paclitaxel (C/P) in clinically selected patients (pts) with advanced non–small-cell lung cancer (NSCLC) in Asia (IPASS). J Clin Oncol 27:408s; 2009 (suppl) abstr 8006.

5. Cappuzzo F, Ciuleanu T, Stelmakh L, et al: SATURN: A double-blind, randomized, phase III study of maintenance erlotinib versus placebo following nonprogression with first-line platinum-based chemotherapy in patients with advanced NSCLC. J Clin Oncol 27:407s; 2009 (suppl) abstr 8001.

6. Takeda K, Hida T, Sato T, et al: Randomized phase III trial of platinum-doublet chemotherapy followed by gefitinib compared with continued platinum-doublet chemotherapy in Japanese patients with advanced non–small-cell lung cancer: Results of a West Japan Thoracic Oncology Group trial (WJTOG0203). J Clin Oncol 28:753–760, 2010.

Add a New Comment
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License