A central question since the introduction of the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) like Tarceva (erlotinib) and Iressa (gefitinib) has been how best to use them. Specifically, one standard way that we integrate new agents in cancer care is to combine them with the treatment that is our current standard of care. However, four large trials of first line doublet chemo with or without Iressa (the INTACT-1 and INTACT-2 trials) or Tarceva (TALENT and TRIBUTE trials) were both definitively negative. One exception was that the subgroup of never-smokers who received carboplatin/taxol (paclitaxel) with concurrent Tarceva on the TRIBUTE trial actually did significantly better than never-smokers who received chemo alone. (click on image to enlarge) However, one important issue with the curves above is that the group receiving Tarceva doesn't really seem to do better (as indicated by the separation of the curves) until about 6 months into the trial. And because the chemotherapy portion was limited to six cycles given every three weeks, this means that the recipients of Tarceva didn't really seem to do better until they had completed chemotherapy and were continuing on Tarceva alone. This, along with some other suggestive bits of inconclusive evidence, led me to be quite disinclined to give an EGFR concurrent with standard chemotherapy, out of concern that they may even be antagonistic with each other. I think a very relevant test would be chemo/Tarceva vs. Tarceva alone in never-smokers, or better yet, people with an EGFR mutation, but the TRIBUTE trial didn't test was chemo/Tarceva vs. Tarceva alone. But a trial by the Cancer and Leukemia Group B (CALGB) (Cancer and Leukemia Group A is long gone), known as CALGB 30406, has tried to address this question directly.

My friend and medicine residency colleague (together at Boston's Brigham and Women's Hospital), Dr. Pasi Janne, led this trial that is being reported in an oral session at ASCO in just a few days. But the abstract tells us some valuable information already. Back when the trial was designed, the idea was to try to focus primarily on patients with an EGFR mutation, or at least to try to enrich the trial with as many people with an EGFR mutation as possible, through mutation testing was not readily available. Instead, the trial compromised by including patients with a previously untreated advanced lung adenocarcinoma and with a history of either never-smoking or having just a minimal smoking history ( a term now referred to as being an "oligo-smoker", with oligo meaning "few" in Greek). They enrolled a total of 188 patients, of whom 182 were actually randomized. It's unclear to me why more were randomized to chemo and Tarceva (100 patients) than to Tarceva alone (82). They provided results based on the overall trial population and also broken down by whether the patients had an EGFR mutation or not (available for 172 patients). They report on several important findings. First, the mutation rate was 39%, which is in keeping with our consistent observation that the EGFR mutation are is higher in never- or oligo-smokers with an adenocarcinoma than in the general North American population (about 10%), but it's definitely not a proxy for having an EGFR mutation, and it's less than we see for Asian never-smokers with a lung adenocarcinoma (about 60%). And as with the IPASS trial, we see substantial differences depending on whether patients have an EGFR mutation or not. One initial conclusion is that there doesn't seem to be a detrimental effect to giving EGFR TKIs concurrent with chemotherapy. For the overall population, the response rate was higher in the chemo/Tarceva group (47% vs. 34%), no difference in progression free survival (PFS) (6.0 vs. 6.7 months for chemo/Tarceva vs. Tarceva alone, respectively), and a non-significant trend toward better median overall survival (OS) in patients receiving Tarceva alone (24.0 vs. 19.6 months). Obviously, OS is also affected by subsequent therapies after the first line setting. When we look at results for patients with no EGFR mutation, they are clearly better served by receiving chemotherapy in addition to Tarceva in the first line setting. Response rate (31% vs. 8%) and PFS (4.8 vs. 2.8) were convincingly superior with chemo added, but there wasn't any difference in OS (15.4 vs. 13.7 months, actually favoring the Tarceva alone group). The results for the patients with an EGFR mutation were substantially different. First, they did significantly better in all efficacy measures than those who don't have an EGFR mutation, as shown in the numbers that follow. There weren't any differences between chemo/Tarceva and Tarceva alone in terms of response rate (69% vs. 66%) or PFS (17.2 vs. 16.4 months). In terms of OS, the numbers look superior for chemo/Tarceva (39.0 vs. 27.6 months), but this is based on only a relatively small number of patients, since only 53% of the patients in the overall trial population had died (that's good), and the patients with an EGFR mutation were doing clearly better. Because medians are especially unreliable in very small groups, I suspect that this difference in OS for patients with an EGFR mutation will become far less pronounced with longer follow-up. Plus, it's biologically implausible for a treatment to improve OS but not improve PFS or response rate, which are the measures of how the treatment is actually doing while a patient is on that treatment. We'll get more information from the full presentation of the data at ASCO, not to mention the subsequent publication with longer follow-up, but we can definitely glean good information from the CALGB 30406 trial. 1) About 40% of never-smokers or oligo-smokers in North America have an EGFR mutation. 2) Those with an EGFR mutation do especially well overall. 3) For those without an EGFR mutation, Tarceva alone as a single agent is clearly inferior compared to chemo/Tarceva in this trial, but also compared to chemo alone in the IPASS trial. I'd say that chemo alone remains the standard here: OS may well be far better with sequential chemo followed by an EGFR inhibitor (but this trial didn't test that). 4) For people with an EGFR mutation, there doesn't seem to be a clear difference between Tarceva alone and chemo/Tarceva concurrently. There certainly isn't evidence for a detrimental effect of an EGFR inhibitor and concurrent chemo (for either patients with an EGFR mutation or EGFR wild type (no mutation)). While there is a numerically longer median survival OS here in patients who received chemo/Tarceva, I suspect this is just a product of small numbers and very variable outcomes for individual patients. We'll learn more about this trial over time, but these are already some very important conclusions.