There are a few interesting new stories coming out of the 2013 European Society for Medical Oncology (ESMO) conference in Amsterdam that's just wrapping up today.  One has been on the  anti-PDL1 immunotherapy MPDL3280A, an agent from Roche that we saw some initial promising data on at ASCO 2013 earlier this year.  Dr. Jean-Charles Soria from the Institute Gustave Roussy in Villejuif, France presented data on this agent in a larger cohort now of 85 patients for safety data and 53 patients for efficacy.

While again confirming the overall efficacy of this agent, with a response rate of 23%, which tended to be long-lasting at greater than 24 weeks, the angle discussed most enthusiastically was the subgroup analysis of smokers vs. non-smokers. This showed that the response rate was 26% (11 of 43) among smokers vs. 10% (1 of 10) among never-smokers, a finding that is especially striking in the face of so many treatment options that are more frequently directed toward never-smokers. The hypothesis is that smokers are exposed to far more antigens and may have a more immunoreactive form of lung cancer.

Another interesting finding in the trial is that those patients whose tumors expressed higher protein levels of the PD-L1 target had a higher response rate, which was 83% in the minority with 3+ PD-L1 expression and 46% in those with 2+ or 3+ protein expression.

To me, these results again confirm the provocative findings we've seen with "immune checkpoint inhibitors" like nivolumab, an immunotherapy directed against PD-1. It may well be that this tendency toward greater responsiveness in smokers is shared among most or all of these immunotherapies.  But it's also worth bearing in mind that, while these results were called a potential "game changer" in Amsterdam, they are still based on small numbers, and we should be cautious when we know that the definitions of response can be especially challenging with immunotherapy, where new nodules that turn out to not be cancer may appear and you sometimes see what appears to be progression before a response (and these cases are not categorized as a response).  If a single additional never-smoker had demonstrated an objective response, the response rate would have been essentially the same between smokers and never-smokers (26% vs. 20%).   

We need to see this and other immunotherapy agents tested in larger numbers, as is starting to be done on a larger scale.  The early findings are very promising, and I'm very hopeful that MPDL3280A and other agents in this class are active, but it's also worth bearing in mind that Taxotere (docetaxel), an agent we consider pretty mundane now, had a response rate of 25% in phase II testing but turned out to be more like 5-10% in larger studies.