Oncologists often have their own ideas about how effective or ineffective a targeted therapy like an EGFR tyrosine kinase inhibitor (TKI) (Tarceva (erlotinib) or Iressa (gefitinib)) is compared with standard chemotherapy options. We've seen a lot of new information emerge in the field over the last few years, and we now have evidence that patients with an EGFR mutation generally have excellent outcomes with an EGFR TKI, but many physicians believe that patients who don't have an EGFR mutation don't get any benefit from this class of agents, or certainly far less than they get with a standard chemo agent. The INTEREST trial directly compared Iressa with the standard chemo agent Taxotere (docetaxel), which was the first agent proven to improve survival for previously treated patients with advanced NSCLC. As I described in a post when the INTEREST trial was initially reported, this was a worldwide trial of 1466 previously treated patients with advanced NSCLC who were randomized to one treatment or the other, and I previously reported the results showing that Iressa and Taxotere led to an identical survival: (click on image to enlarge)

That result is somewhat "interesting" (ugh..the horrible puns write themselves), in that 1) many people felt (and still d0) feel that chemo is simply more effective than a targeted therapy pill, and 2) Iressa actually didn't show a significant improvement in survival compared to placebo in a trial of previously treated patients with advanced NSCLC, unlike Tarceva. So I might have expected equivalence of Tarceva with Taxotere, but if anything these results were better than I might have expected for Iressa in a population of patients who weren't selected for having an EGFR mutation. This brings us to the next point. The study also looked at the outcomes of Iressa vs. Taxotere in about a quarter of the patients who had tissue available for molecular marker testing. Many of us might have expected that patients with an EGFR mutation, and perhaps with EGFR gene amplification (by fluorescence in situ hybridization (FISH)) would do far better with Iressa than chemo, while some people have claimed that people with a KRAS mutation would get no benefit from Iressa and should do much better with chemotherapy. The direct comparison of Iressa vs. Taxotere in different molecular marker subgroups is shown below. The right side of the figure shows progression-free survival, and the right side shows overall survival. The dots falling on one side of the vertical lines or another show a group doing better with Iressa (if the dot falls to the left) or with Taxotere (if the dot falls to the right), and the horizontal bars represent the uncertainty of where the real result could actually be (the smaller the group, the greater the uncertainty). What these results show is that there are almost no differences in any groups, in that the dots pretty much fall close to the vertical line of equivalence, with the exception that progression-free survival was far better with Iressa in patients with an EGFR mutation. However, even in that group of patients most likely to do exceptionally well with Iressa, overall survival wasn't different, probably because people with an EGFR mutation could end up doing just as well if they received an EGFR inhibitor later (and I suspect most did). But there wasn't evidence that other EGFR markers, such as the protein as measured by immunohistochemistry (IHC) or gene amplification as measured by FISH could discriminate which patients would do better with one approach vs. another, nor was KRAS mutation a useful tool. The best answer may well be to try to give previously treated patients both chemo and an EGFR inhibitor over time, but there isn't evidence to favor the order of one vs. another as a second line therapy. And for those of you asking why we shouldn't combine chemo and an EGFR inhibitor together, there's no evidence that this is a good idea, and there is even some suggestive evidence that these two approaches can interfere with each other. In the meantime, this work shows that it's still dealer's choice.