NSCLC: Similar Immunotherapy Trials, Different Results. Early Signals From PD-1 Checkpoint Inhibitors in First-Line NSCLC

Similar Immunotherapy Trials, Different Results: Early Signals From PD-1 Checkpoint Inhibitors in First-Line NSCLC
H. Jack West, MD
Disclosures|August 25, 2016

The evolution of the role of immunotherapy for patients with non-small cell lung cancer (NSCLC), as well as many other cancers, has been rapid and dramatic. The PD-1 checkpoint inhibitor nivolumab was approved by the US Food and Drug Administration (FDA) as a second-line therapy on the basis of a demonstrated significant survival benefit compared with docetaxel in PD-L1 expression–unselected patients, first in those with squamous NSCLC[1] and later in patients with nonsquamous NSCLC histology.[2] Pembrolizumab, with the same mechanism of action, was also FDA approved in late 2015 following previous treatment for patients with advanced squamous or nonsquamous NSCLC. But the focus with pembrolizumab has been on patients with significant PD-L1 expression[3]; in contrast to the "PD-L1 agnostic" approval of nivolumab, approval of pembrolizumab was specifically for patients who were positive on testing with the 22C3 antibody as a companion diagnostic test,[4] which defines positivity using a 50% expression level as a cutoff. This high bar is present in less than 30% of patients, making it restrictive but enriching for those patients most likely to benefit from immunotherapy. Though improved survival with pembrolizumab compared with second-line docetaxel has also been demonstrated in a more liberally defined population with any degree of PD-L1 expression,[5] this has yet to translate into a change in indication for advanced NSCLC.

The fate of these two remarkably similar agents has been largely dictated by their differing indications.
The fate of these two remarkably similar agents has been largely dictated by their differing indications, which are the result of strategic decisions by the companies that developed them. Despite evidence that the efficacy of nivolumab as second-line therapy for at least nonsquamous NSCLC patients is correlated with PD-L1 expression (as measured by a different antibody and different cutoffs from those used for pembrolizumab),[2] the survival benefit of nivolumab over docetaxel in the broad range of unselected patients was sufficient to justify approval and use without prospective PD-L1 testing. This has led to treatment patterns in the United States that have overwhelmingly favored the use of second-line nivolumab without advance testing for PD-L1, in contrast with the requirement to test for PD-L1 using a relatively high threshold level of expression in order to have the opportunity to treat with pembrolizumab. Treating without a required threshold of PD-L1 expression led to a clear win in round 1 for nivolumab.

The promise of dramatic and prolonged responses to immune checkpoint inhibitors with little or no significant toxicity logically leads to extensive testing in the first-line setting. Accordingly, scores of trials with multiple PD-1 or PD-L1 checkpoint inhibitors integrated into the first-line setting have been conducted and are continuing enrollment.The array of studies varies across several critical variables:

Scores of trials with multiple PD-1 or PD-L1 checkpoint inhibitors integrated into the first-line setting have been conducted and are continuing enrollment.
Patient selection: All patients or limited by squamous or nonsquamous histology, and/or restricted by PD-L1 status?

Which PD-1 or PD-L1 checkpoint inhibitor is studied?

Is the PD-1 or PD-L1 inhibitor administered as monotherapy or as part of a combination (with standard first-line therapy or another immunotherapy, such as a CTLA-4 inhibitor), or both?

Is treatment with the immunotherapy for a limited or indefinite duration?

For patients assigned to the control arm, is crossover to immunotherapy permitted?

For the clear majority of these trials, progression-free survival (PFS) has been the primary endpoint, largely out of concern that subsequent therapies will eliminate the ability to discern the benefit of first-line treatment. However, response rate (RR) and PFS are potentially confounded by the long-recognized difficulty of interpreting response to immunotherapy, making it difficult or impossible to define an optimal primary endpoint for first-line treatment.

In this setting of so many uncertainties, the preliminary results of two critically important, seemingly similar first-line trials have become available, at least in the form of official press releases, with very discordant results. The KEYNOTE-024 trial randomly assigned 305 treatment-naive advanced NSCLC with PD-L1 expression of >50% (approximately 28% of all patients) to either any of several platinum-based doublets or pembrolizumab at a new, fixed dose of 200 mg IV every 3 weeks. The press release noted that recipients of first-line pembrolizumab demonstrated not only a significantly longer PFS but also a significantly improved overall survival (OS), despite the fact that patients assigned to initial chemotherapy could cross over to pembrolizumab upon progression.[6] Though we have yet to view the actual data, this report indicates that pembrolizumab is not only highly effective for this subset of patients, but also that it is most beneficial when administered as early as possible. At the same time, this subset is the minority of patients most likely to benefit from an immune checkpoint inhibitor, so it would be a mistake to presume that these results can be generalized to a less stringently selected patient population.

Is there reason to suspect that there are clinically significant differences among the many checkpoint inhibitors being tested?
This last point was underscored by the sobering press release describing the negative results of the CheckMate-026 trial,[7] which randomly assigned 541 advanced NSCLC patients with a less strict PD-L1 cutoff point of 5% with a different antibody (corresponding to about 40%-45% of patients being included) to one of multiple platinum-based doublet chemotherapy regimens or nivolumab at 2 mg/kg IV every 2 weeks. The study failed to demonstrate a significant improvement in PFS, and though no information was provided about OS, we might infer that a positive result was probably not withheld; companies rarely sit on good news.

We await the presentation of the data from these trials, but in the meantime, we should ask ourselves how to make sense of these similar trials producing such different outcomes. Is pembrolizumab, seemingly so similar in its efficacy and tolerability in the second line, a significantly better agent than nivolumab? Is there reason to suspect that there are clinically significant differences among the many checkpoint inhibitors being tested?

My explanation is that these agents are still remarkably similar in efficacy and that the marked difference in outcome is attributable to the critical importance of patient selection in the first-line setting, far more so than among previously treated patients. Even with only a minority of patients demonstrating a response to second-line nivolumab—approximately 20% in the studies of squamous[1] and nonsquamous NSCLC[2]—and many other patients diluting the very good outcome with nivolumab in that minority, nivolumab bested a weak opponent in second-line docetaxel. But considerably more active first-line doublet chemotherapy offers less opportunity to dilute the benefit seen in a minority. Whether the proportion of patients who benefit significantly more from PD-1 checkpoint inhibitor therapy is 20% or closer to the approximately 28% of patients "cherry-picked" for KEYNOTE-024, the selected but broader population eligible for CheckMate-026 diluted this subset of more immune-sensitive patients with too many patients who were not actually beneficiaries of immunotherapy. This approach undermined the benefits of immunotherapy that would likely also have been conferred by nivolumab in a more select subset of patients who met the eligibility criteria for the pembrolizumab trial.

Immunotherapy will transform the first-line landscape, but not by overturning standard therapy for most or all patients.
Of course, this remains only a hypothesis, especially without the data to review from these studies. As we continue to receive data on an ever-growing collection of trials featuring first-line immunotherapy, our understanding will continue to be refined, and we will need to integrate data on findings of combinations of PD-1 or PD-L1 checkpoint inhibitors with other immunotherapies, conventional chemotherapy, and targeted therapies. But even this first volley of press releases already confirms that immunotherapy will transform the first-line landscape, but not by overturning standard therapy for most or all patients.