Exploring Subset Analyses in the EMPOWER-Lung 1 Trial for NSCLC

Martin F. Dietrich, MD, PhD

Targeted Oncology^TM: What did the EMPOWER-Lung 1 trial [NCT03088540] demonstrate about the efficacy of cemiplimab [Libtayo] in key subgroups of the population of patients with non–small cell lung cancer?

DIETRICH: The hazard ratio [HR] for death for patients with brain metastases at baseline is 0.17 [95% CI, 0.04-0.76], an 83% risk reduction in this subset.¹ This is clear: immunotherapy has virtually no boundary through the blood-brain barrier, while the chemotherapy arm clearly underperforms. It’s a sicker population, so it’s not meant to compare one versus the other population, but only those 2 regimens. There is an improved benefit for older patients; I think the relative benefit was made in the underperformance of chemotherapy; it may not be as well tolerated.

Cemiplimab had the same benefit for locally advanced and metastatic disease. There was an [HR of 0.48] for squamous cell carcinomas, and that may not be as easy to treat in this setting, as well

How did patients with a greater than 50% PD-L1 expression respond to cemiplimab versus chemotherapy?

Looking at the disease response, there is about a 40% objective response rate for cemiplimab.¹ This is basically going back to the question, do we need chemotherapy? If you think about the 60% response rate that we would expect, here you have 40% for immunotherapy versus 20% for chemotherapy; I think they are additive in their respective value.

The median duration of response in this PD-L1–high subset is 6.0 months for chemotherapy—and this is the ECOG 1594 baseline—versus almost 17 months for the cemiplimab group. The time to response is really the time to first scan; there’s not a big difference [between the arms], but we know from our clinical assessments that chemotherapy works much faster between the 2 treatments.

What does this trial suggest about cemiplimab’s efficacy versus chemotherapy for different subsets of patients with high PD-L1 expression?

This brings us back to the case for PD-L1 levels. [Looking at the] depth and time of response there’s a clear correlation between the level of PD-L1 and the depth and length of response between the 2 responder populations. So, the closer we get to PD-L1 of 100%, the more impact we can expect, and the less likely we need chemotherapy. The way I do it in my clinic—aside from clinical assessment—is how favorable is the biomarker setup? If the patient has a PD-L1 level of less than 50%, I always add chemotherapy. If a patient has a PD-L1 level of greater than 50%, then it depends on the clinical assessment and how close the patient comes to a PD-L1 level of 100% between the 2 parts.

There are 3 subgroups of PD-L1 greater than 50%: 60% to 90%, 50% to 60%, and greater than 90%. There’s almost an exponential benefit increase as we move closer to 90% and greater. The response rate goes from 32% [for the 50% to 60% PD-L1 group] to 46% [for the greater than 90% PD-L1 group]. Median PFS [progression-free survival] goes from 4.3 months to 6.2 months to 15.3 months [in ascending order of PD-L1 expression]. And then, median OS [overall survival] goes from [21.9 months], to 22.1 months, to unreached, respectively. So, I think those would be a very interesting patient population for single agent PD-1 [inhibitor].

There is a risk of adverse events [AEs], and [concern for] how well the single agent PD-1 inhibitors are tolerated. We have these discussions with patients, what to expect, and I don’t want to undervalue the severity of AEs that can occur. But especially if you look at the higher grades, AEs are not common. There’s a high rate of continuation, and a high rate of tolerability. When we look at chemotherapy, the comparison is not quite fair; they’re such different [types] of AEs.

But 94% of patients were able to continue on cemiplimab. There was a surprisingly high number of AE-related discontinuations for chemotherapy, with 96%. The duration of therapy was much different.

What data support the use of cemiplimab plus a platinum doublet chemotherapy?

I do want to share [an important] point about the width of data that the EMPOWER-Lung 3 trial [NCT03409614] has reported. The approval for first-line cemiplimab and platinum doublet is currently in FDA review, and we have seen some of the data. This trial was for patients who are treatment-naive, with any PD-L1 expression. There was stratification by histology and PD-L1 levels for low, negative, and high—greater than 50%; with the primary end point being OS.

The median OS was 22 months with cemiplimab and chemotherapy versus 13 months for chemotherapy alone.² This seems very similar to what we’ve seen for the PD-L1 high groups, including those that are not immunotherapy responsive. Median PFS was 8.2 months versus 5.0 months, respectively. And the objective response rate was 43% for the cemiplimab group versus 22% for chemotherapy alone; the median DOR was also in a very similar level. So, it’s clearly [supporting] a notion that the chemotherapy will cover the PD-L1 less than 50% population quite well. With that width of data is coming, the FDA is reviewing the data, and hopefully by this summer there will be a decision on that, as well.

_References:

_{1. Sezer A, Kilickap S, Gümüş M, et al. Cemiplimab monotherapy for first-line treatment of advanced non-small-cell lung cancer with PD-L1 of at least 50%: a multicentre, open-label, global, phase 3, randomised, controlled trial. Lancet. 2021;397(10274):592-604. doi:10.1016/S0140-6736(21)00228-2}

_{2. Gogishvili M, Melkadze T, Makharadze T, et al. LBA51 - EMPOWER-Lung 3: Cemiplimab in combination with platinum doublet chemotherapy for first-line (1L) treatment of advanced non-small cell lung cancer (NSCLC). Ann Oncol. 2021;32(suppl_5):S1283-S1346. doi:10.1016/annonc/annonc741}