Completing our tour of clinical factors that we might use for predicting benefit with the EGFR monoclonal antibody Erbitux (cetuximab) in the FLEX trial is the development of a rash. We're actually discussing rash separately from the other clinical factors from the FLEX trial that we discussed in the prior post for a couple of reasons. First, rash on the treatment isn't something you can know about until you've committed yourself to at least starting the treatment (you can't have an Erbitux-induced rash before giving Erbitux). Second, it's not really clear whether this rash is predictive of doing particularly well with Erbitux or prognostic of doing well in general (on Erbitux or anything else). For years, the EGFR inhibitor side effect of rash has been associated with more favorable outcomes, as I've described in one of my earliest posts, on the potential significance of a rash on EGFR inhibitors. The investigators who developed the FLEX trial decided prospectively (pre-planned, not a "fishing expedition") to look at how patients did as a function of whether they developed a rash in the first 21-day cycle of treatment (all patients alive at day 21 were included). They found that 55% developed a rash to some degree, while 33% developed a grade 1 (mild) rash, 18% developed a grade 2 (moderate) rash, and 5% developed a grade 3 (severe) rash. It was very interesting to see that the patients who developed a rash on Erbitux did far better than those who did not. In fact, it was also interesting to see that the patients who didn't develop a rash on Erbitux had a worse median survival than the patients who were assigned to chemo alone:

FLEX Trial Survival by Rash

(click on figure to enlarge)

What's fascinating to me is that this reminds me incredibly of a similar analysis of survival by rash with Tarceva (erlotinib), an oral EGFR inhibitor, in the TRIBUTE trial of chemo with either Tarceva or a placebo. As in the FLEX trial, not only did patients who didn't develop a rash do relatively poorly, they did worse than the patients who received chemo and a placebo:

TRIBUTE Trial Survival by Rash

This gets to the heart of the issue. This work might lead some to presume that it would be a good idea to give everyone Erbitux, see who develops a rash in the first three weeks, then continue it only for those people. Perhaps, but I think there are two problems with this. First, I think it's extremely unlikely that the majority of oncologists and patients will actually stop the Ebritux they've started based on the absence of a rash. It just isn't going to happen (and I believe the people selling Erbitux suspect this and are happy about that). The second issue is that this work doesn't say that the people who develop a rash do better because they're getting Erbitux. In fact, we see that the people who don't develop a rash despite getting Erbitux (or Tarceva, for that matter) tend to do worse, not only compared to the people on an EGFR inhibitor who do develop a rash, but also compared to "average": the general population of people who haven't been tested on an EGFR inhibitor (the chemo alone or chemo/placebo group). Perhaps development of a rash is a measure of a person's general constitution or immunoreactivity. We might imagine that failure to develop a rash is like a new measure of poor performance status, so that this breakdown by rash isn't really a cause and effect relationship but a marker. In the meantime, perhaps some people will find the rash analysis makes the argument in favor of giving Erbitux more compelling than the rather modest benefit seen in the larger population. I think that only really works if people are inclined to actually stop the Erbitux if a patient doesn't have a rash, and I'm not convinced that the Erbitux is actually vaulting the survival as much as identifying people who are going to do better. Still, it's reasonable enough to consider Ebritux anyway -- the real question is whether this analysis is really more about clinical research or marketing.