Vassiliki Papadimitrakopoulou, MD: There was a clear-cut benefit in all subgroups of patients who were treated with the combination of chemotherapy and pembrolizumab, but the benefit was most pronounced in the high—PD-L1 subset of patients with a hazard ratio of 0.42. The benefit extended to patients with a TPS of 1% to 49% and less than 1%, demonstrating that patients in all settings could benefit from this therapy.

Corey J. Langer, MD: Toxicity is a concern, particularly with immunotherapy. These agents introduce a whole constellation of unique toxicities that are not typically observed with chemotherapy: thyroid disease, potentially hypothyroidism in 10% to 20% of our patients, arthritis, arthralgias, rash, and sometimes pruritus in the absence of rash. These are common, but there are even several less common immune-mediated side effects centering on the liver, kidney, and lungs. We do see pneumonitis and sometimes nephritis or hepatitis, although those are fairly rare. All these immune-mediated toxicities were higher in the pembrolizumab arm but at manageable levels. By and large, the vast majority were grade 1 or 2, very rarely grade 3 or 4, toxicities. Beyond that, there really wasn’t much difference in toxicity. Most of the rest of the toxicities were typical of chemotherapy, those that we would usually see with combination pemetrexed/carboplatin.

The crossover lends particular credibility to the survival difference. If crossover had not been included, we could say that survival was better because the control arm never had exposure to pembrolizumab or to other immunotherapy agents. But in the original KEYNOTE-021g trial, well over 70% of patients crossed over to either pembrolizumab or some other similar agent. In the randomized phase III trial, the crossover rate exceeds 50%, and as time goes on, that crossover percentage will go up further as more patients have disease progression and become eligible for crossover.

Vassiliki Papadimitrakopoulou, MD: The trial included maintenance pemetrexed, as per standard of care, in combination with pembrolizumab for the investigational arm versus pemetrexed in the standard-of-care arm. The actual use of maintenance therapy was great. It was in the range of 69% to 75% in both arms. Patients received the full benefit of maintenance therapy in both standard-of-care and investigational arms.

Corey J. Langer, MD: KEYNOTE-189 really defines the current maintenance approach for the nonsquamous population. Here pemetrexed was continued in the absence of disease progression or untoward toxicity. It’s given at 3-week intervals. Pembrolizumab is continued along with the pemetrexed for up to 2 years. I have to confess, for some of my patients who are treated off study, at that 1-year mark, I may start to lengthen the interval to 4 weeks. I certainly will accommodate vacations and family events. But otherwise, I pretty much stick to that schedule. There’s a major debate regarding just how long we need to continue immunotherapy.

There are data from one of the CheckMate trials with nivolumab looking at folks with stable disease or response at 1 year who are randomized to continuing nivolumab or stopping, and there’s clearly a PFS benefit for those who continued the nivolumab. Over time, we’re starting to see the emergence of a survival difference. One year is insufficient. Should we do the same study at 2 years? I don’t really know, although I would be open to that notion. In the absence of disease progression in a patient who’s really tolerating pembrolizumab well, it’s actually my inclination to keep going beyond even the 2-year mark. But in truth, we have no prospective data to validate one approach or another.

Transcript Edited for Clarity