Answer: We are in the midst of a transition to a much more biomarker-driven world of treating non–small-cell lung cancer. Whenever possible, we should be making treatment decisions based on molecular, rather than demographic, features like smoking status, patient sex, or tumor histology. With regard to whether to initiate therapy with an EGFR tyrosine kinase inhibitor (EGFR-TKI) or to start with chemotherapy, the EGFR-TKI–based therapy would really be reserved for patients with an identified activating EGFR mutation. At ASCO 2011, Rafael Rosell and colleagues presented interim data from the EURTAC trial of standard platinum-based chemotherapy versus single-agent erlotinib in patients with a prospectively identified EGFR mutation.¹ As we’ve seen in several other trials presented in the past few years using gefitinib or erlotinib, patients with an EGFR mutation had an overwhelming improvement in progression-free survival with an EGFR-TKI relative to chemotherapy in this study (Table 1). The most common adverse effects in the erlotinib arm were rash (80%) and diarrhea (57%); patients in the chemotherapy arm experienced more anemia (46%) and neutropenia (36%). Dose modification or interruption due to treatment-related adverse events was higher for the chemotherapy arm (47% vs 23%), as were discontinuations due to treatment-related adverse events (14% vs 5%). These results certainly make a compelling argument to test prospectively in anyone with more than a negligible chance of having an EGFR mutation.

In terms of mutations that may predict response to chemotherapy, we have evidence suggesting that the excision repair cross-complementing gene 1 (ERCC-1) mutation is associated with sensitivity or resistance to platinum-based chemotherapy.² Ribonucleotide reductase M subunit 1 (RRM-1) expression is associated with responsiveness to gemcitabine.³ Thymidylate synthase may well be correlated with reduced clinical benefit with pemetrexed.⁴ While all of these are promising, they haven't been validated. I would, therefore, be wary about obviating what is otherwise a proven standard of care for something that is still best categorized as provocative but investigational. So, I am not inclined to order predictive tests for chemosensitivity—unless and until there is validated prospective research that shows a significant improvement in a definitive patient outcome such as survival. We don't yet have that for any of these chemosensitivity assays.

There is a growing consensus toward conducting molecular marker testing, at least for patients with recurrent nonsquamous non–small-cell lung cancer. That recommendation is now integrated into the National Comprehensive Cancer Network (NCCN) guidelines.⁵ The guidelines recommend EGFR testing tor patients with nonsquamous-cell carcinoma; that's the only mutation that is currently included. There is also compelling reason to strongly consider testing for ALK translocations and perhaps KRAS as a guide for any molecular targeted therapies.⁶ ALK translocations have been clearly shown to be associated with a high probability of response, typically of a duration in the range of a year or longer, with the ALK inhibitor crizotinib.⁷ Crizotinib was approved by the FDA in August 2011 for the treatment of locally advanced or metastatic NSCLC that is ALK-positive as detected by the ALK Break Apart FISH Probe Kit. And, while KRAS mutations have not yet been proven to serve as a clinically relevant decision-making variable, patients with KRAS mutations have consistently been shown to have a low probability of major response to EGFR-TKIs (though not a clearly worse survival benefit with these agents than is seen in other patients with EGFR wild type) and to have a nearly mutually exclusive relationship with EGFR mutations and ALK rearrangements.⁸

Moreover, there are novel agents targeting KRAS mutations in development that we hope could make KRAS mutations more clearly a biologically relevant target. For now, however, we cannot yet say that KRAS mutations have a clearly evidence-based role in clinical decision-making for advanced NSCLC.

Back to top

Answer: Clear pathways for squamous versus nonsquamous histology have emerged in terms of an algorithm for NSCLC. In patients with a nonsquamous histology, there is now a recommendation from the NCCN to test for EGFR mutation right at the time of diagnosis—and to use this result to direct patients to EGFR-TKI–based therapy or chemotherapy, depending on the outcome.⁵

In patients with a squamous histology—who comprise about 20% or 30% of the US non–small-cell lung cancer population—most of us would use a platinum/taxane or platinum/gemcitabine combination. I, personally, tend to favor cisplatin or carboplatin with gemcitabine, depending on the fitness of the patient.

With regard to the chemotherapy of choice, bevacizumab is contraindicated in patients with a predominantly squamous non–small-cell lung cancer. Pemetrexed, which is one of our more active agents in NSCLC overall, is not effective in patients with squamous NSCLC histology, and the package insert for this agent specifically notes that it is not indicated in these patients.⁹

For patients with a nonsquamous subtype, who are the majority of patients with NSCLC in the United States, pemetrexed is very effective for first-line,¹⁰ maintenance,^11,12 or second-line therapy. There have been a couple of other studies that have looked retrospectively at other types of chemotherapy, either single-agents or combinations. But, there really hasn't been any compelling evidence of a histology-specific effect with the other, more commonly used, chemotherapy agents in non–small-cell lung cancer right now. That lack of histology-specific effect also applies to erlotinib, which is certainly associated with some profound responses in NSCLC, generally in patients with an adenocarcinoma subtype. But, the survival benefit with it is not limited to those with an adenocarcinoma or nonsquamous histology. It is approved and has activity in all of the non–small-cell subhistologies.

We do not have any really strong evidence that one regimen is clearly, significantly superior to several other comparable options. There is an FDA-approved choice of carboplatin, paclitaxel, and bevacizumab for nonsquamous non–small-cell lung cancer in those patients who are otherwise eligible for bevacizumab (no history of hemoptysis, untreated brain metastases, significant cardiac history). A platinum and pemetrexed combination, sometimes with bevacizumab, is another very strong consideration as well. I am one of many oncologists who have often used carboplatin with pemetrexed and bevacizumab as a very good intersection of efficacy and tolerability in the first-line setting. Although it has not been FDA-approved for that specific setting, this combination has looked extremely promising in some clinical trials¹³ and is the subject of a phase 3 prospective randomized trial comparing carboplatin, paclitaxel, and bevacizumab with carboplatin, pemetrexed, and bevacizumab.¹⁴ I would certainly consider that to be a fine alternative.

Finally, cetuximab was also shown to be associated with a statistically significant improvement in overall survival when added to cisplatin/vinorelbine in the FLEX trial,¹⁵ though I would say that the majority of people in the oncology community were disappointed in the absolute difference in outcomes. It appears as an option in the NCCN guidelines, but I would consider it an option that has yet to really find its place in advanced NSCLC. In an interesting trial presented at the World Conference on Lung Cancer, investigators performed a post hoc analysis of results from the FLEX trial and noted a far more substantial benefit with the addition of cetuximab in the subset of patients (approximately 30%) with a high “H-score,” the product of the proportion of tumor cells positive by immunohistochemical staining for EGFR protein on tumor cells by the intensity of the staining.¹⁶ If this strategy is validated in a prospective fashion, it could make cetuximab a far more compelling addition to standard chemotherapy in the subset of patients with advanced NSCLC most likely to benefit from it.

But, any of the commonly used platinum-based doublets—in some cases with bevacizumab added—are reasonable choices, with cetuximab as an option that may emerge as more clinically relevant if we can reliably identify the patients more likely to benefit substantially from its addition.

Back to top

Answer: I would say that there is now a consensus that you would want to start EGFR-TKI therapy at the earliest opportunity, as soon as you know that a patient has an activating EGFR mutation. Although erlotinib has not been approved by the FDA for first-line therapy at this time, the NCCN guidelines list it as a strong recommendation. The evidence is extremely consistent, including the recent presentation of the EURTAC trial at ASCO this year,^1,17 that there is a very significantly longer progression-free survival with first-line erlotinib (or another EGFR inhibitor in other trials) compared with standard chemotherapy in the first-line setting for patients with a prospectively identified EGFR mutation. Although we haven’t seen significant improvements in overall survival, we do see more favorable trends for survival in patients with an EGFR mutation getting first-line EGFR-TKI therapy. You may be able to a get a comparable benefit by giving an EGFR-TKI as second-line therapy or later, or maintenance, but there is always a chance that the patient will have some complication that precludes him/her from getting subsequent therapy. You would really not want to take the chance of missing out on such a potentially beneficial intervention.

Perhaps the bigger question, which is unanswered, is what to do if you find out after the first cycle of chemotherapy that a patient has an EGFR mutation. Would you discontinue the chemotherapy and immediately switch them over to erlotinib, or would you add erlotinib to their chemotherapy? There is no clear answer to that question, but I think many of us would be inclined to continue the first-line therapy rather than discard a potentially effective therapy prematurely. In that situation, we would have a very low threshold to transition to erlotinib at the earliest evidence of any progression, and potentially to switch to erlotinib as maintenance immediately after 4 or 6 cycles of chemotherapy, even in the absence of progression. We want to ensure that patients who have such a high probability of a dramatic and prolonged response to these agents will get their opportunity.

I would be very wary about presuming that the results we’ve seen in the metastatic setting will definitely translate to earlier stage disease. The evidence we have is with stage IV disease, and we don't at this point have evidence that supports giving erlotinib or another EGFR inhibitor as a maintenance therapy, or an adjuvant therapy, for earlier stage disease. In fact, the BR19 trial, which looked at adjuvant gefitinib versus placebo in patients with resected Stage IB-IIIA NSCLC, showed a troubling trend toward a detrimental effect of gefitinib as an adjuvant or maintenance therapy compared with placebo.¹⁸ In a multivariate analysis, there was actually a strong but nonsignificant trend toward shorter survival duration in the gefitinib arm among patients with an EGFR mutation (P = .097). We can't really explain why that happened, but we just don't have good evidence that giving EGFR inhibitor therapy in early stage NSCLC is beneficial in the same way that has been demonstrated in advanced disease.

We should get some useful insight on this question from an ongoing phase 3 trial called RADIANT, being led by Dr. Karen Kelly at the University of California at Davis.¹⁹ In this study, patients with Stage IB-IIIA NSCLC who are positive for EGFR by immunohistochemistry or FISH testing will receive postoperative erlotinib. Patients may or may not have received adjuvant chemotherapy after resection. The study has completed enrollment, but it will be several years before we get an idea of the results.

Back to top