Another lung cancer trial that received a good deal of attention at the recent European Society for Medical Oncology (ESMO) conference in Stockholm this past week was conducted by the Spanish Lung Cancer Group and led by Dr. Rafael Rosell, who is chief of medical oncology at Catalan Institute of Oncology in Barcelona and one of the true international greats in the field who has made important contributions for a couple of decades now. His leading interest these days is in refining treatment plans based on molecular characteristics of lung tumors, and this particular study focuses on how patients who have been selected based on EGFR activating mutations do when they receive EGFR inhibitors. We have a pretty good hint of this from several smaller studies that have already been conducted around the world: (click to enlarge) So while these are smallish trials, the largest being 127 patients who actually had a mutation, they all show response rates in the 55-82% range, and the median time to progression is 9-12 months (with some going far beyond).

The Spanish trial is really the same design, but writ larger. They screened 2,312 patients to identify 307 who had an EGFR mutation predictive of favorable response, then treated them all with tarceva (erlotinib). Of these, they had clinical data available from 193 patients, 54% of whom receiving tarceva first line, and 46% receiving it as second line therapy. The group was very disproportionately never-smokers (68%), female (71%), and with adenocarcinomas (75%), so we again see that there is a major overlap between molecular features and clinical variables. Not surprisingly, the results looked similar to the smaller studies done elsewhere: the response rate was 71%, with a particularly high likelihood of response in patients with an "exon 19 deletion" (the other important mutation to be eligible is on exon 21), which refers to a particular part of the EGFR gene where the mutation is found. Survival was also very favorable, with a mean of 22 months, but notably better for women than for men (28 vs. 17 months). Mean time to progression was 12 months, also better for women than for men. Dr. Rosell stated, "There is no excuse not to test for these mutations...Testing is the first step toward a dramatic improvement in subgroups of patients." Well, actually, with all due respect, I beg to disagree...and I'm not alone here. For starters, tarceva is already FDA approved and very widely used throughout the world. This study corroborated the finding that never-smokers were remarkably more likely to carry EGFR mutations than former or current smokers (38.4% vs. 9.2% vs. 6.1%, p < 0.0001), and they were also far more common in women than men, and in patients with adenocarcinoma or bronchioloalveolar carcinoma (BAC) than other histologies. I've already explained that if I have a never-smoking woman, or a perhaps a remote former smoker with BAC, tarceva's likely to be high on the list and very possibly my first treatment anyway. And if I have a never-smoking Asian woman with adenocarcinoma (not rare in my practice, and extraordinarily common throughout Asia), there's no way I'm going to decide against giving tarceva to her, no matter what her EGFR mutation testing shows. Importantly, even with mutation testing, the response rate is 71%, not 100%, and there are also responders who don't have an EGFR mutation. So they are neither necessary nor sufficient to benefit from oral EGFR inhibitors. Another key point is that there's no evidence at all that you need to give tarceva first line to get the benefit. In fact, when they compared the survival of patients who received tarceva as first line treatment vs. second line treatment, both the progression-free survival and overall survival were 16% better for patients who received tarceva later. So let's just take a step back and say that if you don't test for EGFR mutations, and you give tarceva anyway as a later therapy, there's not a scintilla of evidence that it's worse than giving it immediately (and in fact, the evidence is that it's a little better later). I'll just clarify that I'm not militantly against molecular testing, but rather trying to provide a reality check. The mutation zealots have a tendency to become wild-eyed and suspend thinking critically about the issue, so I think it's important to take a step back. Mutations definitely do a good, not excellent, job of predicting who will benefit from EGFR inhibitors, but we've also seen that... 1) they don't guarantee a response, nor does absence translate to a lack of benefit 2) they are also correlated with very unusually favorable response and survival with chemo alone (see prior post) 3) they overlap consistently and significantly with clinical factors that don't require scant tumor tissue or cost extra to assess 4) if you plan to give an EGFR inhibitor anyway as second or third line therapy, there is no evidence yet that giving it earlier is better We're certainly going to continue to see more studies that incorporate molecular testing, but mutation testing will need to counter these issues before there is a compelling reason to make them a part of routine clinical practice.