COMBINATION IMMUNOTHERAPY FOR NSCLC-NOT YET FOR PRIME TIME

In just a few short years, the introduction of immune checkpoint inhibitors directed against PD-1 or PD-L1 has ushered in nothing less than a new era in the treatment of non–small cell lung cancer (NSCLC) as well as many other cancers. Starting with previously treated patients with advanced NSCLC, multiple agents soon demonstrated superior efficacy to second-line docetaxel. In short order, pembrolizumab monotherapy became a standard first-line therapy for the subset of patients with high PD-L1 expression, based on the positive results of the KEYNOTE-024 trial,^[1] followed earlier this year by an approval for the carboplatin/pemetrexed/pembrolizumab combination as first-line treatment for patients with advanced nonsquamous NSCLC regardless of PD-L1 status,^[2] based on the favorable results of the randomized phase 2 KEYNOTE-021g trial.^[3]

But even as our treatment landscape has been reshaped to accommodate PD-1 or PD-L1 inhibitors as monotherapy or a single immunotherapy in combination with chemotherapy, many have projected that this is just a transient stop on the way to combinations of immunotherapy agents.

Specifically, combinations of PD-1 or PD-L1 antibodies with inhibitors of CTLA-4 are leading the charge, with a few variants of a nivolumab/ipilimumab regimen generally demonstrating promising efficacy and feasible tolerability in advanced NSCLC,^[4] as well as small cell lung cancer.^[5] But before we call these combinations—or those with novel immunotherapy agents—a presumptive winner, we should be cautioned that preliminary results from phase 3 studies are negative. We should be appropriately circumspect and walk back some of the prognostication around these combinations as "irrationally exuberant," unless or until we actually see prospective randomized phase 3 data demonstrating an improvement in overall survival (OS) over our current best standards.

We should be appropriately circumspect and walk back some of the prognostication around these combinations as 'irrationally exuberant.

One of the most important trials testing the role of immunotherapy combinations is MYSTIC, a phase 3 study of approximately 1100 advanced NSCLC patients randomly assigned to the PD-L1 inhibitor durvalumab with or without the CTLA-4 inhibitor tremelimumab versus any of several standard chemotherapy combinations as first-line treatment.^[6] In July, we learned from a press release that this trial failed to demonstrate superior efficacy in the co-primary endpoint of progression-free survival (PFS) for the durvalumab/tremelimumab combination or durvalumab alone compared with first-line chemotherapy.^[7] The results for OS, the other co-primary endpoint, are expected in the first half of 2018, with no preliminary information offered.

What are the implications for the concept of immunotherapy combinations from this single press release with no actual data? First, it is possible—though I think quite unlikely—that the MYSTIC trial will demonstrate significant improvement in OS for the combination or durvalumab alone, despite neither one demonstrating a significant improvement in PFS. It is uncommon to see a difference in OS, which is a product of the combined benefit of typically several lines of therapy along with the underlying pace of the disease, that is far more pronounced than the impact of the treatment on PFS, which isolates the impact of that specific line of therapy. For instance, looking at the results of the KEYNOTE-021g trial of first-line chemotherapy with or without concurrent pembrolizumab,^[3] the dramatic and highly statistically significant difference in PFS eclipses the nonsignificant trend toward improved OS. I expect that with the vast majority of patients in MYSTIC who are recipients of first-line chemotherapy likely to receive immunotherapy upon progression, any differences in OS will be less pronounced than those for PFS.

Even if there is a statistically significant difference favoring an immunotherapy combination over chemotherapy, the threshold for this strategy to become a standard of care is higher than that. The toxicity of these combinations have presented a steep learning curve marked by high rates of grade 3 and higher toxicities, hospitalizations, and even treatment-related deaths, even among the well-selected patients treated at centers of excellence with the most experience using immunotherapies. We should be concerned about the potential toxicity of such a combination in a population that may be older, sicker, and less selected than in the early phase 1-2 experience, and who are being treated at centers with limited experience and support for patients.

The absolute difference in efficacy must justify the remarkable financial cost of a combination of two immunotherapy agents.

In addition, the absolute difference in efficacy must justify the remarkable financial cost of a combination of two immunotherapy agents, probably the most egregious challenge to the concept of appropriate value for a treatment that we have ever considered as a potential standard of care in lung cancer management. The oncology community rightly rejected the marginal efficacy of necitumumab combined with chemotherapy for advanced squamous NSCLC, despite the statistically significant improvement in OS,^[8] probably because it was recognized to represent a very poor value for the minimal benefit.^[9] To justify the cost premium of a combination of immunotherapy agents, we should expect this approach to not just cross a statistical threshold in a trial of more than 1100 patients, but also to demonstrate remarkably dramatic benefit. We should not expect this from a trial for which we have heard only disappointing results thus far.

We should expect this approach to not just cross a statistical threshold but also to demonstrate remarkably dramatic benefit.

Finally, we need to ask what this press release implies about other trials of different immunotherapy combinations, such as the CheckMate-227 trial, an over-2200-patient, multifaceted series of randomized comparisons of nivolumab with or without ipilimumab to standard chemotherapy in patients with or without tumor PD-L1 expression, as well as separate arms comparing chemotherapy with concurrent nivolumab versus chemotherapy alone.^[10] We received a noninformative press release on this trial in January,^[11] as we did with MYSTIC, just stating that the focus will be on OS results that are still maturing. While we have yet to learn of even preliminary results from this trial, I would contend that the general results of the immunotherapy trials have been far more similar than different. This makes me dubious that the results with nivolumab/ipilimumab will be remarkably different from those with durvalumab/tremelimumab. I suspect that the silence speaks volumes: No company sits on good news, and month after month passes by without a hint of commentary disabusing us of the presumption of negativity. And even if CheckMate-227 manages to demonstrate a statistically significant benefit for an immunotherapy combination, it will need to do so with an absolute benefit threshold sufficient to justify the added cost and toxicity burden of such a combination.

I suspect that the silence speaks volumes: No company sits on good news.

Just as it is premature to consider combinations of immunotherapy agents as the heir apparent and anticipated new standard of care in advanced NSCLC, it is premature for me to declare them a failure based on a press release from a single trial with no data presented. In the coming months, we should expect to see data not only from the MYSTIC trial but also from CheckMate-227 and potentially others to help us clarify the role for these combinations. In the meantime, however, these sobering results should at least remind us that more is not necessarily better.