PACIFIC Trial Implications Concerning NSCLC Mutations

Mark A. Socinski, MD:Right. There were about 6% of patients on the PACIFIC trial who were known to haveEGFRmutations. Now, patients withEGFRmutations, if you look at the information we have about checkpoint inhibition in stage IV disease, this is a group of patients that does not appear to derive significant benefit from checkpoint inhibition as monotherapy. If you look at all the second-line trials, whether you look at pembrolizumab, nivolumab, atezolizumab, and specifically look atEGFRmutation—positive patients, they, in checkpoint inhibition, were no better than docetaxel.

So I think there are patients, and I think one as a group,EGFRmutation—positive patients, these are not patients who derive great benefit from checkpoint inhibition as monotherapy. It may be a different situation if you’re using checkpoint inhibitors in combination with various chemotherapy regimens. We don’t have time to discuss that; that’s mostly a stage IV issue today.

In stage III disease, I’m not surprised with that—given the relatively small numbers, in the data that we have with regard to the use of checkpoint inhibition in all of the oncogenic subset drivers, includingEGFRmutation—positive patients—that we saw this. I would caution clinicians today to put too much weight on that, as it’s a is very small numbers of patients. I don’t think that we can exclude a benefit that could be significant to individual patients. For instance, in my practice, I would not necessarily exclude a patient from the use of durvalumab in this setting based on theEGFRmutation status, because the numbers are really small and it may be misleading and we clearly need more data in this setting.

Across the board in all the trials of checkpoint inhibition, PD-L1 has been an important marker. And I think what we can conclude from the analysis of PD-L1 status is that this really is pretty consistent across all the trials as well as the use of different antibodies to measure PD-L1, as well as the use of different PD-1 or PD-L1 inhibitors.

The more strongly you stain for PD-L1, the greater the benefit you derive from immunotherapy. The question is, what’s the cut point? Is it 1%, is it 25%, is it 50%? You know in stage IV disease if you have greater than 50% PD-L1 positivity, we use immunotherapy as a standard of care. Those are patients who are very strongly staining, and they get great benefit from immunotherapy compared to chemotherapy.

Now specifically in the PACIFIC trial, PD-L1 was measured. Historically, durvalumab has used the cutoff of 25%, greater or less than 25%. There was, as part of the regulatory analysis of this trial, by the European authorities, there was a request to look a bit more granular at the PD-L1 status. This was an unplanned retrospective analysis. It was not either stratified for and therefore could not be protected by the power of randomization. When they looked specifically at the subset of patients that had less than 1%, or so-called negative patients, there didn’t seem to be a survival advantage for those patients.

Now, again, in my opinion, given the fact that this was not designed into the trial, that it was not powered for, it was unplanned, retrospective in nature, I think we have to be very careful about interpreting that. I think it is hypothesis generating, and I think it would be very reasonable to do a randomized trial in PD-L1—negative patients of durvalumab, yes or no, with placebo, pretty much a PACIFIC trial design in a PD-L1–negative population.

In my practice in stage III disease, I am agnostic to the PD-L1 status. I don’t make clinical decisions based on PD-L1 because I want to potentially provide the benefit of checkpoint inhibition to all the patients, and I do think that this does raise the question, but it certainly doesn’t answer the question definitively, about the PD-L1—positive versus –negative population. And again, I think we have to be very careful about unplanned retrospective analysis of very small subsets and subsets that weren’t necessarily stratified for.

In my opinion, PD-L1 is a useful biomarker, but it’s not a perfect biomarker. In my opinion, the greatest usefulness of PD-L1 is in stage IV disease. And it’s really in a subset of patients in which you find them to be greater than 50% positive and you can offer those patients mono-immunotherapy rather than chemotherapy and have a greater benefit. So I think that’s the main utility.

Beyond identifying those very strong staining patients for PD-L1 in stage IV disease, I think PD-L1 is kind of a continuum. It makes me feel either more confident or maybe less confident about the benefit that the patient may get from immunotherapy in this setting. But it doesn’t necessarily cause me to exclude patients or include patients for use of immunotherapy based on PD-L1 status. It’s very clear, for instance, in stage IV disease that patients who get chemotherapy plus immunotherapy who are PD-L1—negative still get benefit from the combination of this versus chemotherapy alone. So, again, a PD-L1 is a useful marker, a good biomarker, not a perfect biomarker. It can be informative in certain clinical situations, but I do think we need to go beyond PD-L1 as a biomarker and look for novel biomarkers in the future. And there’s lots of work around that at this particular time, but none has risen to the top yet.

Transcript edited for clarity.