This week’s New England Journal of Medicine includes not one but two seminal publications on EGFR mutation status, response to EGFR tyrosine kinase inhibitors (TKIs) and sequencing of therapy in advanced NSCLC.

First, Dr. Tony Mok and colleagues reported detailed results of the IPASS trial, which have been outlined previously by Dr. West. The investigators selected for inclusion patients from clinical subgroups associated with high rates of EGFR mutations and response to EGFR TKIs: women, Asians (the study was conducted entirely in Asia), patients with adenocarcinomas and light- or never-smokers. However, the presence of an EGFR mutation was NOT part of the study entry criteria. The trial randomly assigned over 1200 patients with these clinical characteristics to either front-line chemotherapy with a maximum of 6 cycles of carbo/Taxol or to Iressa. The investigators were able to determine the EGFR mutation status for 36% of patients.

The primary endpoint of the trial was progression-free survival (PFS). Patients in both the Iressa and the chemotherapy groups had a median age of 57, were overwhelmingly female (80%) and most were never-smokers (94%). For the population as a whole, the primary endpoint of PFS was met, with the group on Iressa showing a statistically significant 26% decrease in the risk of progression (also known as a hazard ratio of 0.74), as shown in panel A. In the first six months, the survival curve favors chemotherapy but thereafter the curves cross and PFS is more favorable with Iressa. Why? Patients with EGFR-mutant tumors had high response rates with chemotherapy (47.3%) also so both EGFR-mutation-positive and EGFR-mutation-negative patients could have responded to chemotherapy, which lasted a maximum of six cycles, thus favoring chemotherapy early-on. After six months, the high response rates and prolonged PFS in EGFR-mutation-positive patients would have driven the superiority of Iressa.

For EGFR-mutation-positive patients, there was a more dramatic 52% reduction in the risk of progression (panel B) if they were assigned to Iressa. While the survival curves are pretty close together at the beginning, they subsequently separate in favor of Iressa in this patient population. What happens to EGFR-mutation-negative patients (panel C) is more dramatic – right from the beginning, these patients were much more likely to progress if they were assigned to Iressa compared to chemotherapy.

These PFS curves can help us understand what happens with overall survival in this trial.

Though the overall survival was not statistically different between those who received Iressa initially (18.6 months) and those who received chemotherapy (17.3 months), the results broken down by subgroup are interesting. For EGFR-mutation-negative patients, the curves separate early with a strong trend in favor of chemotherapy (though this did not reach statistical significance). Although patients who progressed on Iressa could then receive carbo/Taxol, only 39% of them did. Another 10 percent received a different anti-cancer therapy (not specified). That leaves 50% of patients not getting anything; I’d be willing to bet at least some progressed too rapidly after initial treatment with Iressa to be eligible for further therapy.

For the EGFR-mutation-positive patients, the curves are very close together initially, reflecting the fact that EGFR-mutation-positive patients are likely to do well initially with either Iressa or chemotherapy (though they did have higher response rates with Iressa). Over the long-term; however, there is a trend towards improved survival for those patients who were treated with Iressa initially. Again, the results were not statistically significant but the number of patients with mutation status available was relatively small and this does represent an EARLY analysis of overall survival since many patients were still living at the time of the analysis.

This study included quality of life (QoL) data, assessed using the well-validated FACT-L questionnaire. Results overall favored Iressa but were more impressive when presented by mutation status. Seventy percent of mutation-positive patients on Iressa reported sustained improvement in QoL (versus 45% of mutation-positive patients on chemo). In contrast, mutation negative patients were more than twice as likely to report improved QoL when they got chemo compared to Iressa (36% versus 15%). In terms of toxicity, patients on Iressa were half as likely to experience a severe side effect and were also half as likely to stop therapy because of side effects.

So what does this study tell us? In my opinion, it argues strongly for knowing EGFR mutation status upfront and making one’s initial therapeutic decision based on that information. Some might quibble over the lack of statistical significance for an overall survival benefit and would argue that EGFR-mutation-positive patients will do well as long as they receive an EGFR TKI at some point. However, the QoL data and safety data cinches it for me. A highly effective therapy with better QoL and fewer side effects: why NOT give it first in this population?

Perhaps more important is the data with respect to EGFR mutation negative patients. I think knowing EGFR status is more critical for this group – it showed me that I really could do a disservice to a patient by making a decision to go with a front-line EGFR inhibitor because the patient had a high likelihood of having a mutation based on clinical characteristics. If one extrapolates from this study, even in a highly enriched population, at least 40% will not have a mutation. Of course, sometimes tissue is not available and we still have to treat the patient using our best judgment. My fear about this trial is that the results highlighting the benefit of Iressa for EGFR-mutation-positive patients and the benefit of chemotherapy for mutation-negative patients will be extrapolated beyond the front-line setting. Increasingly, I am seeing patients who are never offered Tarceva at any point in treatment because they are known to be mutation-negative. This is in spite of evidence that patients without mutations do benefit from this therapy in the second- or third-line setting.

In my opinion, there is yet another important reason to try to obtain EGFR status on patients. The novel EML4-ALK fusion gene also occurs more frequently in light or never-smokers but does not co-exist with EGFR. A recent publication highlighting the clinical characteristics of patients with the ALK rearrangement noted that this mutation occurred in 33% of their light or never-smokers whose tumors were negative for the EGFR mutation (more to come later on this paper). With the advent of a novel treatment for these patients, it will be important to identify them.

Stay tuned some interesting findings from the Spanish experience with EGFR mutation testing and Tarceva treatment.