NSCLC: A Tale of Two Checkpoint Inhibitors

Where Are We Now With PD-1 Inhibitors and the Utility of PD-L1 Screening for Advanced Non-Small Cell Lung Cancer?

H. Jack West, MD


October 21, 2015

In the wake of two new approvals for programmed cell death 1 (PD-1) inhibitors in previously treated advanced non-small cell lung cancer (NSCLC)—one dependent on screening for programmed death-ligand 1 (PD-L1) expression, and another without—oncologists are left with the questions of whether to test for PD-L1 and whether to use pembrolizumab or nivolumab. As we sift through the data, I'll offer my perspective on why I am inclined to check for PD-L1 expression and consider it clinically relevant, with a caveat that there is no remote consensus on this issue.

Nivolumab was approved by the US Food and Drug Administration for previously treated patients with advanced squamous NSCLC in early March 2015, on the basis of the 3.2-month improvement in median overall survival (OS) (9.2 vs 6.0 months; hazard ratio [HR], 0.59; P < .001) reported in the CheckMate 017 trial[1] comparing second-line nivolumab with standard docetaxel. However, we had to wait until earlier this month for the approval of nivolumab for advanced nonsquamous NSCLC, after the similarly positive CheckMate 057 trial[2] of the same design, presented at the 2015 meeting of the American Society of Clinical Oncology, also showed a significant OS benefit for nivolumab (12.2 vs 9.4 months; HR, 0.73; P = .0015).

One key difference between the two studies was that although nivolumab was superior in all subgroups of patients with squamous NSCLC irrespective of PD-L1 expression, the OS benefit was confined to the patients who had nonsquamous disease with PD-L1 expression; there was no hint of better efficacy of nivolumab among the patients without PD-L1 expression. Nevertheless, the approval for nonsquamous NSCLC is also irrespective of PD-L1 expression.

In contrast, pembrolizumab is now approved for previously treated patients with advanced NSCLC (either squamous or nonsquamous) who test positive for PD-L1 by the specific companion diagnostic test PD-L1 IHC 22C3 PharmDx (Quest Diagnostics; Madison, New Jersey). Although this specific test defines "positive" as greater than 50% staining—a group in which pembrolizumab was shown to have a response rate of 45% and median OS was not reached[3]—it is not clear how rigorous different payers or centers will be in using this specific test and cutoff to determine eligibility for pembrolizumab.

So with two different PD-1 inhibitors approved in the same second-line setting, one of them specific to patients with PD-L1 expression, is there value in pursing PD-L1 testing, or should clinicians follow a default option of second-line nivolumab without testing for PD-L1?

Although nivolumab is a strong choice for all patients with squamous NSCLC, for those with nonsquamous NSCLC, I think the clearest answer is that it remains a judgment call, but one for which I personally favor PD-L1 testing. Specifically, I am most inclined to pursue this testing initially—at the time of fact-gathering after initial diagnosis, when we send for driver mutations, such as EGFR and ALK—rather than delay second-line therapy.

Why? Among the PD-L1–negative patients in CheckMate 057, the progression-free survival trended in the wrong direction for nivolumab, with a hazard ratio of 1.19. Although this is not statistically significantly more favorable with docetaxel, it raises a concern that old, unexciting docetaxel is modestly more effective for patients with PD-L1–negative cancer (who comprised nearly one half of the patients in the trial) while costing a tiny fraction of the approximately $150,000 per year for nivolumab. Nivolumab is better tolerated, but I'm dubious of the overall value of a treatment that trends toward inferior efficacy and costs far more than 10 times that of the alternative.

Those who favor nivolumab as the default second-line therapy point to the fact that there are patients who are negative for PD-L1 but respond to nivolumab (approximately 10%), and these responses are often dramatic and long-lasting. Although that is true, I would prioritize having PD-L1–negative patients receive nivolumab as a third-line therapy in order to ensure that all patients have the opportunity to benefit from an immune checkpoint inhibitor. With a safety profile markedly superior to that of docetaxel, nivolumab is likely to remain a very feasible treatment option for just about all patients after progression on an alternative second-line therapy. We cannot be nearly as confident that many patients will remain candidates for survival-prolonging docetaxel in the third-line setting.

Future developments are likely to make PD-L1 testing even more compelling as immune checkpoint inhibitors are tested in the first-line setting. If proven to provide a benefit in addition to or instead of current standard first-line therapies, the vast majority of this work is being directed toward patients with PD-L1 expression, making it very likely that first-line use will be limited to PD-L1–positive patients. Similarly, immunotherapy combinations in lung cancer may follow the same pattern observed in metastatic melanoma,[4] in which a combination is incrementally beneficial for PD-L1–negative patients, whereas single-agent therapy may be sufficient for patients with PD-L1 expression.

A final question is whether PD-L1–positive patients with advanced NSCLC who are eligible for either nivolumab or pembrolizumab should preferentially receive one over the other. Although the general patterns of efficacy and tolerability of these agents should lead us to strongly suspect that they produce remarkably comparable results in the same population of patients, the greater convenience of an every-3-week schedule leads me to slightly favor pembrolizumab, all other factors being equal. We might expect that payers will develop agreements that lead to a "preferred checkpoint inhibitor" for patients in the future, leading us to be directed to one or another remarkably comparable agent, like the way in which we so often administer one proton pump inhibitor or fluoroquinolone antibiotic over another.

As debatable as these questions are, we can certainly expect more data and a changing landscape of immunotherapy treatment options in the next few years. In the meantime, the availability of two more effective treatment options for advanced nonsquamous NSCLC leads us to the enviable dilemma of how best to integrate them into our current management strategies.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.