The improvement in median survival of 1.2 months with the monoclonal antibody to EGFR erbitux (cetuximab) in the FLEX trial that I've previously described was statistically significant, but there's plenty of room to debate whether it's really clinically significant (see prior post). What If we could add some way to refine our predictions of who will benefit from the addition of erbitux?

At the Chicago Lung Cancer meeting in November 2008, Dr. Ulrich Gatzemeier presented results of a planned subset analysis of the FLEX trial, in which the results compared the outcomes of patients randomized to receive cetuximab by whether they developed an acne-like rash within three weeks of starting this treatment. This was based on a growing and rather consistent experience that patients who receive EGFR inhibitors, whether oral tyrosine kinase inhibitors or IV antibodies, generally show a a strong trend or significantly better survival than patients who develop no rash (see prior post). The investigators found that there was a very significant difference in efficacy with cetuximab among the patients who developed a rash (of any severity) after two weeks (56%) compared with those who didn’t (44%). The difference in median survival was a near doubling: 15.0 months for the patients with any rash, compared with 8.8 months for those who didn’t:

First, when you compare the median overall survival of 15 months for the patients who developed an early rash on erbitux to the median survival of 10 months for patients who received chemo alone, this suggests to me that some patients are getting a pretty meaningful benefit from erbitux, but it may be only about half of the people who receive it. If these results are real, and I believe they are, the question is whether we can use the early development of a rash to distinguish between the real beneficiaries of this treatment and those for whom it will just add side effects and cost. The other possibility is that patients who don't develop a rash with erbitux would do worse with any treatment that they happen to receive -- that the association of rash with better survival is prognostic of better outcomes, but not necessarily predictive of doing better specifically with erbitux.

I haven’t been in the practice of pulling people off of a treatment after just a few weeks on the basis of a lack of symptoms, and I think that many patients may be inclined to still continue erbitux, whether they develop a rash or not. But I think that at least we might be able to feel more comfortable that patients who receive erbitux with chemo and develop a rash are likely to get a survival benefit that we’d consider to be genuinely clinically significant.