Bunn Discusses Frontline Nivolumab NSCLC Results, Issues With PFS Endpoint

Paul A. Bunn Jr, MD

The CheckMate-026 results were a surprising twist for the oncology community, as the frontline use of nivolumab (Opdivo) for the treatment of patients with PD-L1—positive non–small cell lung cancer (NSCLC) did not improve progression-free survival (PFS) when compared with physician’s choice of chemotherapy.

A total of 541 patients with nonsquamous and squamous NSCLC were enrolled in CheckMate-026 and randomized to receive nivolumab at 3 mg/kg every 2 weeks or standard combination chemotherapy. All patients enrolled in the study had PD-L1 expression on ≥5% of tumor cells.

As the lung cancer community awaits the full study data, which will be presented at an upcoming medical meeting, Paul A. Bunn Jr, MD, offered his expert insight in an interview with OncLive.

Bunn, Distinguished Professor, Division of Medical Oncology/University of Colorado, James Dudley Chair in Lung Cancer Research, University of Colorado Denver, discussed the meaning of the CheckMate-026 results and issues with the trial’s primary endpoint of PFS, as well as his views on combining immunotherapy with chemotherapy in patients with NSCLC.

OncLive: Let’s first discuss nivolumab as a single agent in frontline NSCLC. What happened in CheckMate-026?

Bunn: Not surprisingly, the initial immunotherapy studies were in second-line after people already progressed on chemotherapy. Two phase III trials and one phase II trial randomized with nivolumab, pembrolizumab (Keytruda), and atezolizumab (Tecentriq) were all positive. Then, obviously, they moved these into the first-line to investigate.

In first-line, people with no PD-L1 expression were excluded from the trials. Bristol-Meyers Squibb (BMS) did 1 trial where people with any positivity were included. Merck did 2 trials, 1 included people with low PD-L1—positivity—1% to 49% PD-L1 expression on cells—and one with high PD-L1 expression of more than 49% positivity. Merck announced that the trial of patients with high PD-L1 levels was positive, so people sort of thought that any PD-L1–positive trial like this will be positive, as well.

It was announced that the BMS “any PD-L1—positivity trial” was not positive. Like anything, when there is a press release, you don’t know all of the data. Originally, most of these trials were designed with PFS as the primary endpoint. However, it turns out that for immunotherapy, PFS isn’t a very good endpoint.

It is possible that this trial could be negative for PFS, but be positive for OS. It is also possible that the high level of PD-L1—positivity cohort will have an advantage but the low PD-L1–positivity cohort will not. It is, in a way, disappointing, but we don’t really have the details to know what to do. Undoubtedly, there is another trial called CheckMate-227 that is ongoing. There will probably have to be some modifications of that study.

You said PFS is not an appropriate endpoint when studying these agents. What are some better ones to use?

RECIST criteria response rate is a better criterion than PFS because the people who respond tend to respond for a long time. It is not as good as OS. A number of the trials have changed their endpoint. Many of the trials used PFS and then they changed to have 2 coprimary endpoints—PFS and OS—for this very problem.

What modifications need to be made for the ongoing CheckMate-227 trial?

The CheckMate-227 trial had 6 arms. It had 3 arms for people who were PD-L1—positive. Those arms consisted of nivolumab, nivolumab plus ipilimumab (Yervoy), and chemotherapy. Something will have to happen with the nivolumab-alone arm. In the arms with PD-L1–negative patients, no one received nivolumab alone. They received nivolumab plus ipilimumab, chemotherapy plus nivolumab, or chemotherapy alone. Five of the arms probably don’t need to be changed, depending on what the results were. In the arm with nivolumab alone for PD-L1–positive patients, there will have to be some kind of change there.

Given with what we know with these agents already, what clinical impact could these combinations have?

We don’t have the data yet, and there are all kinds of combinations. The combinations that we’ll know the answer to first will be a study of chemotherapy plus immunotherapy versus chemotherapy alone. There is nivolumab plus ipilimumab, and there is also durvalumab plus tremelimumab.

The pulmonary data for the chemotherapy arms are mixed. It is very hard to predict whether that will be positive or not. It does look like the CTLA-4 combinations seem a little bit better than the single-agent one. They have more toxicity, but I don’t think tremelimumab adds as much toxicity as chemotherapy. Therefore, the additional toxicity is acceptable if the survival is better.

It is good; there are a lot of trials. Of course, essentially, all of the antibodies have combinations with chemotherapy; however, only durvalumab and nivolumab have them with ipilimumab or tremelimumab.

The main one—and always the most important one—is OS. For a lot of patients, these drugs do not work. The first half—up to the median—oftentimes, you don’t see much of a swing. However, the people who respond have a really good response and it will last for a long time. It is the tail of the curve where you sometimes see bigger differences.