An acne-like rash or dry skin is a very common side effect of the drugs that target the epidermal growth factor receptor, with approximately 3/4 of patients who receive the EGFR tyrosine kinase inhbitor tarceva/erlotinib experiencing skin toxicity. Similar skin toxicities are also seen, but a bit less commonly, with the very similar drug iressa/gefitinib, and also frequently with erbitux/cetuximab, a monoclonal antibody that is less well studied in lung cancer. But since the earliest clinical trials of these agents, there have been questions of whether the rash is something more than just a potentially problematic side effect, but rather a marker of someone likely to do well with these agents.

One initial point we can make is that there is a dose dependency associated with these agents. As shown in the table below for iressa, the more drug you give patients, the more commonly you see a rash.

(click to enlarge)

However, even at doses far beyond what was considered the standard or target dose for iressa (250 mg in general, 500 mg in early studies), not all patients develop a rash. I believe that is important, because it may suggest that some people cannot generate a skin response to even high doses of EGFR inhibitors.

The idea that there may be an association between rash development and how well patients with NSCLC do on EGFR inhibitors traces back to the first reported clinical trial of tarceva in NSCLC, by Dr. Roman Perez-Soler in New York. Among 57 previously treated patients with advanced NSCLC, he reported that survival was best in patients with a moderate or severe rash, worst in the patients with no rash, and in between for the group of patients with a mild rash, as shown here.

(click to enlarge)

In fact, the degree of rash that a patient developed was more predictive of a patient doing well on that trial than their activity level going into the trial. These results were interesting, but it was a single small trial. However, similar trends were seen in other tumor types, as shown in the table below:

(click to enlarge)

And similar results were being seen with erbitux, particularly in colon cancer, and to a lesser extent with iressa. My own study with iressa in BAC showed a correlation as well, although it wasn't as clear that more severe rash was better than a milder rash. Instead, it did support the idea that patients who did not develop a rash did not do well.

(click to enlarge)

In fact, in both my trial and one with tarceva in BAC, there were no patients who had a favorable response of tumor shrinkage who failed to develop a rash.

The TRIBUTE trial of first-line treatment for advanced NSCLC, in which patients received carboplatin/paclitaxel alone or with tarceva, showed something very interesting. Like other trials, it showed that patients with a rash did better than those who did not, and the worse the rash, the better patients did. But another VERY important point was that the patients who received tarceva but didn't develop any rash actually did worse than the folks who received chemo and placebo. Here's the breakdown:

(click to enlarge)

So these results suggest that there's something particularly unfavorable about not developing a rash, if they do worse than not even getting tarceva at all.

In the BR.21 trial of previously treated patients with advanced NSCLC, rash again appeared to be an important predictive factor. In that trial, the 75% of patients on tarceva who developed a rash had a much better survival than the 25% who didn't (9.5 vs. 2.2 months, quite a difference!). I have never seen a breakdown by severity of rash. Also interesting was the fact that 17% of patients on the placebo arm developed some kind of rash/skin toxicity, and they did better than the patients on the placebo arm that didn't develop any skin toxicity!

What can we conclude? Well, we certainly don't have all of the answers, but there does seem to be a clear and consistent association of rash with better outcomes. It's not completely clear that a worse rash is associated with better results than a mild rash, but no rash or skin toxicity is associated with worse outcomes. And these results have been seen with multiple agents, in combination with chemo or as single-agents, and in many types of cancers.

But is it how much drug is given or how sensitive a person is? I notice that some patients don't develop a rash even at very high doses, that the folks who don't develop a rash actually do worse than the folks getting a placebo, and finally that the folks on the placebo arm of BR.17 who developed a rash actually do better than others, and I conclude that it has to do with immunocompetence/immune sensitivity. However, I'll admit that nobody understands this especially well, including me at this point.

Importantly, we have not yet seen ANY good evidence that you can convert a non-responder to a responder on EGFR inhibitors by increasing the dose (so please don't start doubling your dose!!). There are a few studies of dose escalation to get to developing a rash, being looked at for years, but I've never heard any results from them, which makes me believe that there's nothing promising there. Nor is there any evidence that patients who develop a major rash do any worse after reducing dose to manage the rash better. I think that someone's tendency to develop a rash is more intrinsic, and that you can't change that.

Finally, this is a correlation, and it's definitely NOT perfect. I've had patients do very well with no skin reaction, and I've had patients with very bad rashes who had their lung cancer progress through them

I'll talk more about managing the rash, and perhaps how dose and rash and smoking status may interact, in future posts. In the meantime, I welcome your comments, questions, or objections.