Immunotherapy in advanced non-small cell lung cancer: Outcomes from EMPOWER-Lung 3: Written recap

In this Managed Healthcare Executive Between the Lines video series, Chul Kim, M.D., and Brian Henick, M.D., discussed the five-year outcomes of the EMPOWER-Lung 3 trial that compared treatment with cemiplimab plus chemotherapy with chemotherapy alone in patients with advanced non-small cell lung cancer (NSCLC). Kim is director of thoracic oncology and an associate professor at MedStar Georgetown University Hospital in Washington, D.C. Henick is the director of the Herbert Irving Comprehensive Cancer Center Phase I program at Columbia University in New York and an associate professor at the Columbia University Medical Center.

Background

Kim and Henick began with a review of what treatment for advanced NSCLC was like before the advent of immunotherapy.

Henick traced the preimmunotherapy era back approximately a decade. The dominant clinical question at the time was simply which platinum partner to use, he said. A major randomized study that had evaluated different chemotherapy combinations in first-line metastatic NSCLC found broad equipoise across regimens, though subgroup analyses revealed histology-based differences: Patients’ nonsquamous tumors tended to benefit more from pemetrexed-based combinations, while those with squamous tumors showed better results with gemcitabine and taxanes. Those findings set the stage for histology-stratified thinking, but the outlook remained poor — median overall survival with platinum doublet chemotherapy rarely exceeded 12 months.

In parallel, molecular profiling began identifying actionable driver mutations (EGFR, ALK, ROS1), which shifted adenocarcinoma management toward tyrosine kinase-based therapy. “Adenocarcinoma has undergone a true sea change since that time with the recognition or identification of molecular drivers that are susceptible to targeted drugs,”Henick said. Per current National Comprehensive Cancer Network guidelines, molecular testing is now the obligatory first step for adenocarcinoma, he noted. Patients with cancers with actionable drivers go to targeted therapy; immunotherapy is considered only for those who do not.

Henick noted that early data raised questions about whether patients with EGFR and ALK alterations could benefit from checkpoint inhibitor combinations — those questions had largely been answered in the negative, and in some populations, such combinations had shown safety signals that remain incompletely understood.

The immunotherapy era

Henick gave a brief overview of the history of PD-1 and PD-L1 inhibitor development in NSCLC, including landmark trials. He noted that early signals emerged from phase 1 trials — CHECKMATE studies and KEYNOTE-001 with pembrolizumab, a phase 1 trial that enrolled approximately 1,000 patients as an all-comers study and ultimately generated data that led to the first FDA approval of immunotherapy for NSCLC in the later-line setting. From there, the field moved toward first-line approvals across multiple agents.

Key landmark trials included KEYNOTE-024,KEYNOTE-042, KEYNOTE-189, KEYNOTE-407 and the CHECKMATE series, which led to pembrolizumab approvals in first-line NSCLC for PD-L1 greater than or equal to 1% and greater than or equal to 50%, both as monotherapy and in chemotherapy combinations. Atezolizumab, cemiplimab, nivolumab plus ipilimumab, and durvalumab plus tremelimumab also received first-line approvals, with nivolumab/ipilimumab and durvalumab/tremelimumab combinations applying across all PD-L1 levels. Henick described the resulting treatment framework as a “checkerboard” based on PD-L1 expression: monotherapy with pembrolizumab, cemiplimab or atezolizumab at high levels; chemoimmunotherapy combinations at intermediate levels guided by histology; and CTLA-4 inhibitor-containing regimens at low or undetectable PD-L1. Kim observed that the multiple agents reflect the progress made in the treatment of NSCLC.

EMPOWER-Lung 3 study design

Kim then turned the conversation to EMPOWER-Lung 3,a randomized, double-blind, placebo-controlled phase 3 study comparing cemiplimab plus chemotherapy with chemotherapy alone as first-line treatment for advanced NSCLC. Key eligibility criteria included treatment-naive advanced disease (Stage IIIB/C or IV) and no EGFR, ALK or ROS1 alterations. The researchers did not select patients by PD-L1 expression level, and patients with squamous or nonsquamous histology were included. A total of 466 patients were enrolled, and the researchers randomly assigned patients in a 2-1 ratio to receive cemiplimab plus chemotherapy or chemotherapy alone, Kim explained. Cemiplimab was administered for up to 108 weeks or until progression. Median follow-up for this five-year analysis was 60.9 months. The primary end point was overall survival, with secondary end points of progression-free survival, overall response rate, complete response rate and duration of response.

Study population and generalizability

Both physicians commented on how well the study population mapped to their clinical practice. Henick noted, though, that the study was predominantly conducted outside the U.S. in Asia and Europe. More than 40% of patients were aged 65 or older, though patients in their 70s and 80s — common in academic oncology clinics — were underrepresented. A large majority (85%) of patients were male; approximately 40% had squamous histology. PD-L1 distribution was roughly one-third each across less than 1%, 1%-49%, and 50% or greater, which is consistent with the published literature. The majority of patients had an Eastern Cooperative Oncology Group (ECOG) performance status of 1; rates of baseline brain metastases were below 10%, which was on the low end of what is typically observed in practice. Never-smokers comprised 13% to 15% of the cohort, which both physicians considered reasonably reflective.

Henick said patients treated at his clinic tended to have a higher ECOG performance status, indicating a lower ability to perform daily tasks. Kim agreed, adding that treatment selection by clinicians needs to account for patient-level factors alongside tumor biology, summarizing the principle as “tailor the treatment regimen based on the patient factors, the biology of the tumor, and clinical factors.”

Histology and PD-L1 as factors in treatment choice

Kim raised the question of whether squamous histology should drive different treatment decisions. Henick confirmed it should — not only because of differences in optimal chemotherapy backbone (pemetrexed-based maintenance is appropriate for nonsquamous but not squamous), but because the different histologies probably represent biologically distinct diseases. “It is important to separate these groups because they probably are, in fact, different diseases on some level,” he said. Henick cautioned that study results lump all NSCLC together without teasing apart interpretations of response rates and survival.

As far as PD-L1 levels are concerned, Kim and Henick agreed it remained an imperfect but essential biomarker. Kim added a nuance: Within the high PD-L1 group (50% or greater), there appeared to be a gradient, with patients at 90%-100% tending to have substantially better outcomes than those at the lower end of the high category. He treated PD-L1 as a “dynamic biomarker” and noted that even in PD-L1-high patients, high tumor burden or rapid disease progression could tip the decision toward chemoimmunotherapy. Henick acknowledged the field hadn’t gone far enough in tailoring therapy, particularly for patients with nonsquamous disease without actionable drivers.

Key results

Kim shared highlights from the data from the five years of follow-up of EMPOWER-Lung 3. Median overall survival was 21.1 months with cemiplimab plus chemotherapy compared with 12.9 months with chemotherapy alone, a 34% reduction in risk of death. Median progression-free survival also favored the combination. Kim noted the five-year survival probability was 19.4% in the cemiplimab plus chemotherapy group versus 8.8% in the chemotherapy group. “There was more than about a 10% difference in the five-year overall survivor rate, which is not insignificant,” he commented. The overall response rate was 43.6% with the combination versus 22.1% with chemotherapy alone, added Kim. Moreover, the overall survival improved with the combination approach in patients with squamous or nonsquamous NSCLC. On safety, grade 3 or greater treatment-emergent adverse events occurred in 49.4% of the combination arm versus 32.7% in the chemotherapy arm.

Henick contextualized the results against the five-year data from pembrolizumab plus chemotherapy, noting that the outcomes looked comparable and that the consistency of benefit across PD-L1 subgroups reinforced established principles about when to combine with chemotherapy.

Kim spotlighted the percentage of patients who were alive after five years. “One number that I keep coming back to is the five-year overall survival rate, which is 19.4% in this study,” he said. “What this really means is nearly 1 in 5 patients receiving this combination of treatments are alive at five years. And this is, I think, meaningful progress.”

Henick sounded a mild note of caution, pointing to a discontinuation rate that was comparable to the discontinuation rate in the trial assessing the pembrolizumab-chemotherapy combination. “Despite the benefit that we do see for about a fifth of patients, there certainly are a number of patients, especially those who are benefiting from treatment, who need to discontinue because of toxicity at some point. That’s something that we all have to navigate with our patients who are going on to this and similar regimens,” he said.

Monotherapy versus a combination approach

Both cancer specialists described similar approaches. Kim’s default for patients with NSCLC PD-L1 less than 50% was an immunotherapy in combination with chemotherapy. For PD-L1 50% or greater, monotherapy was a reasonable option, they agreed, although high tumor burden, symptomatic disease or rapid progression could shift the decision toward combination therapy even in that group.

Henick added that patient fitness is a primary factor in whether to treat with an immunotherapy-chemotherapy combination. “It is really about [patients’] ability to tolerate platinum-doublet chemotherapy and weighing that against the pace and extent of their cancer,” he said.

Immune-related adverse events

Kim noted that the adverse events from immunotherapy can affect virtually any organ system. Common presentations included thyroid dysfunction, skin rash, hepatitis and diarrhea from colitis. Less common but more serious toxicities included myositis, myocarditis and neurologic complications. He said the overall toxicity profile was manageable with vigilant monitoring but noting “any organ system in the body can be affected.”

Henick focused on the challenges of counseling patients. The unpredictability of which patients would experience which side effects — and how severe — can make the conversation difficult. You are offering hope to patients, he said, “but at the same time you’ve got essentially a grab bag of side effects, and we don’t really have great tools to predict who’s going to experience what and how bad it is going to be.”

His approach involved establishing a clinical baseline before treatment began, then prompting patients to report changes after starting therapy. He emphasized eliciting autoimmune history carefully: asking about prior steroid use, rheumatology or endocrinology visits, or relevant gastrointestinal symptoms might be more productive than asking broadly about “autoimmune disease,” a term that some patients don’t recognize. Henick described a case in which a patient’s Crohn’s disease history wasn’t identified until after immune-based colitis had developed, despite a thorough patient history. The lesson, he said, was that “you have to ask questions in the way that’s going to really elicit the true history from the patient and kind of meet them at
their level.”

Gaps in understanding and future directions

Henick identified several gaps in the data, including more granular outcomes by the type of adverse event, survival in patients who discontinued therapy, retreatment outcomes, and results stratified by site of metastasis, particularly liver and brain.

On the future of the field more broadly, Henick saw the landscape as “really wide open,” with the possibility of combining KRAS inhibitors and other targeted agents with checkpoint inhibitors and the potential for next-generation sequencing to identify patients with differential benefit. Kim mentioned the possibility of entirely new combinations that would use antibody-drug conjugates, T-cell engagers, and cell therapies with existing PD-1/PD-L1 and
CTLA-4 backbones.

Concerning extended dosing intervals, Henick noted progress with other approved agents in reducing infusion frequency, viewing this as a meaningful quality of life gain that “can make an impact on patients’ lives.”

On affordability, Henick flagged that ensuring that patients have access to cemiplimab at a reasonable cost would be an important consideration as it enters a competitive landscape. Kim added that from a payer standpoint, durable long-term responses could translate to reduced overall resource utilization that could be seen as “a good investment” over the longer term.