Paul K. Paik, MD:There are a few things that we have to consider when thinking about giving the carboplatin/pemetrexed/pembrolizumab regimen to our patients with newly diagnosed lung adenocarcinoma. The first is, I think, their PD-L1 score. If their PD-L1 expression is high, ≥50%, I think we really need to consider giving pembrolizumab as a single agent by itself. These are data that are backed up by a randomized phase III trial, and this is a regimen that is the standard of care now for this subgroup of patients. We have no data that show that the triplet combination is better than single-agent pembrolizumab in this setting. So, I think we have to be cautious about giving the triplet combination in these high-expressing patients. Here we may, unless there are other mitigating circumstances, favor pembrolizumab as really the standard-of-care option.
I think in patients who have PD-L1 expression levels <50%, it then becomes a discussion with the patient about some possible options that are there. And I think that discussion is led by how the patient is feeling, how symptomatic they are, what their performance status is, what their past medical history is, whether or not they have autoimmune diseases, whether or not they have a very large burden of disease, and whether or not they’re fit enough to receive a platinum doublet chemotherapy as a backbone. These are all considerations clinically that I think we’re all used to making when it comes to recommending certain regimens over the other.
Again, I think we have to be cautious that the subgroup analyses are very small. There are N = 20 patients in each arm. So, for us to make firm conclusions and to say in a sort of de facto manner that the triplet regimen is the standard of care is not something that we can make. And so, this really does require some careful consideration for who to give it in.
I think one of the things that KEYNOTE-021 cohort G does not addressand in fact, KEYNOTE-024 does not address with the first-line pembrolizumab regimen—is, what about bevacizumab? Arguably, the standard of care, based on the data, is actually triplet chemotherapy carboplatin/pemetrexed/bevacizumab followed by maintenance pemetrexed and bevacizumab. It’s important to mention this because we do know that giving these agents in the maintenance setting, adding bevacizumab to pemetrexed for example, and even just maintenance pemetrexed by itself, extends PFS by a significant amount. And we’ve shown this in past chemotherapy trials.
So, we don’t have any data as to whether or not any of these regimens is superior than carboplatin/pemetrexed/bevacizumab. The data suggest, in terms of cross-trial comparisons, that yes, probably the pembrolizumab regimens will emerge as better. But we don’t know for sure, and so I would not discount carboplatin/pemetrexed/bevacizumab in our patients in the first-line setting. It is a reasonable triplet chemotherapy regimen to give. It is a generally well-tolerated chemotherapy regimen to give. There is some thought in patients who have effusions from their diseasespericardial effusions or pleural effusions—that bevacizumab is an angiogenesis inhibitor and may have some added efficacy in controlling these things. So, it is something, again, that we need to take into consideration as an option that was not actually formally investigated as a control in any of the randomized phase III trials or phase II trials for pembrolizumab so far in the first-line setting.
The one subgroup of patients, of course, in whom bevacizumab would not be an indication would be in patients who have very high thromboembolic risksso very recently diagnosed pulmonary emboli, for example—patients who have sites of disease that have an increased propensity for bleeding, patients who have active hemoptysis, and patients who have CNS metastases that have a component of hemorrhage. These are all patients in whom bevacizumab as an angiogenesis inhibitor should not be considered and a subgroup of patients where if a third agent wanted to be added by the patient or the oncologist, pembrolizumab may have a role as a result.
Transcript edited for clarity.