Treating Patients With NSCLC and High PD-L1 Expression in the First Line

Immunotherapies Offer Alternatives to Chemotherapy for PD-L1–High Lung Cancer

During a Targeted Oncology live virtual event, Martin F. Dietrich, MD, PhD, discussed data from 3 trials of immunotherapy agents for patients with non–small cell lung cancer with high PD-L1 expression.


  • A 72-year-old White man presents with chest discomfort, cough, and dyspnea ​
  • Medical history: mild chronic obstructive pulmonary disease​
  • Former smoker, 10 pack-years; quit 25 years ago​
  • CT of the chest and abdomen found a 9.0-cm spiculated mass in the left lower lobe, loculated pleural effusion in the left hemithorax​, right adrenal metastasis​
  • PET/CT: 18F-fluorodeoxyglucose uptake in the left lung mass, pleura and right adrenal gland​
  • Brain MRI: solitary brain lesion (1.5 cm)
  • Note: no central nervous system symptoms​
  • Physical exam: notable for decreased breath sounds in the left lung base; ECOG performance score of 1​
  • Laboratory work unremarkable with normal organ function and blood counts​
  • Image-guided biopsy of the lung mass shows a poorly differentiated adenocarcinoma of the lung​
  • Immunohistochemistry: PD-L1 tumor proportion score of 90% (22C3 pharmDx assay) ​
  • Molecular testing:​
  • Tissue NGS (next-generation sequencing): negative for molecular aberrations in EGFRROS1BRAFALK, RET, MET, ERBB2, NTRK; KRAS
  • Microsatellite stable​

Targeted OncologyTM: What regimens are recommended for a patient with non–small cell lung cancer (NSCLC) with PD-L1 expression above 50%, who is negative for actionable biomarkers?

DIETRICH: Here we get very little guidance. The NCCN [National Comprehensive Cancer Network] guidelines have the full slew of opportunities between squamous and non-squamous histology…and I think the difference between single agent and combination is a mix of biomarker, PD-L1 assessment, maybe even some secondary markers like KEAP1, p53, and STK11 that may signify either a positive or negative response to immunotherapy, as well as the clinical assessment.

We looked at [my practice] and, overwhelmingly, we’re seeing the use of chemotherapy up front. So our patients, certainly during the pandemic, are coming in more symptomatic and requiring, in clinical judgement, the addition of chemotherapy. Otherwise, we don’t have any side-by-side comparison. So, those are all category 1 recommendation, based on phase 1 data, and so basically the selection is dependent on the clinician.1

Which immune checkpoint inhibitor monotherapy are you most likely to recommend for this patient?

What data support the use of pembrolizumab (Keytruda) for patients with stage IV lung adenocarcinoma PD-L1 greater than 50%?

This is the KEYNOTE-024 phase 3 trial [NCT02142738]. [Patients enrolled on this trial] had untreated stage IV NSCLC, independent of histology. They had a total percentage score greater than 50%, no EGFR/ALK mutations, no untreated brain metastases, and no active autoimmune disease requiring systemic therapy.

They were randomized [between] chemotherapy and pembrolizumab. There was a second course of pembrolizumab [for 1 year] versus crossover to pembrolizumab upon progression with chemotherapy. They allowed basically every conceivable platinum doublet for an international study. There was a steep drop-off of patients that didn’t receive subsequent second-line therapy, so choosing the right [first-line therapy] is important. The data are here very strong. The crossover has relatively small numbers.

The 5-year data were presented in 2020.2 The long-term follow-up [for overall survival (OS)] is important. The tail is at the beginning of the [Kaplan-Meier] curve. There’s [not much] crossover. So, the curves start to split within 3 to 6 months in favor of immunotherapy, and we maintain a relatively stable absolute OS benefit between the years to over 30% at 5 years for pembrolizumab in this subset. The hazard ratio [HR] is 0.62 [95% CI, 0.48-0.81], and the median OS is 26.3 months [95% CI, 18.3-40.4] for pembrolizumab versus 13.4 months [95% CI, 9.4-18.3] for chemotherapy. It’s a pretty good intrinsic marker, that is pretty consistent across the phase 3 trials to gauge how sick the individual patients are that were selected to participate.

Which trial demonstrated the efficacy of atezolizumab (Tecentriq) for this subset of patients with lung cancer?

So, looking at the IMpower110 study [NCT02409342], it has very similar trial design [to KEYNOTE 024]. It is a prospective evaluation for squamous or non-squamous stage IV NSCLC. This was a little bit less aggressive with PD-L1 levels in a hierarchical manner. This has only been approved for the first hierarchy of PD-L1 greater than 50%. Here, [patients received] atezolizumab versus chemotherapy with cisplatin or carboplatin plus pemetrexed versus cisplatin/carboplatin/gemcitabine for squamous cell carcinoma. Then there is a maintenance option for non-squamous disease, and a best supportive care for the squamous cell population. Patients received treatment until progression of disease.

In this study, there is crossing of the Kaplan-Meier curves for OS [due to more OS events in the first 4 months for patients receiving atezolizumab].3,4 I am not quite sure why that is; the numbers look relatively alike. The median OS for chemotherapy was 13.1 months [95% CI, 7.4-16.5], so the comparator arm is certainly looking very similar [to that of the KEYNOTE 024 trial]. And then, the median OS for atezolizumab was 20.2 months [95% CI, 16.5-not estimable (NE)] in this subset.

We have a crossing of the curves [between 4 to 6 months]. So, this would be a trial where the chemotherapy upfront may have been a more reasonable option, based on the visuals. The HR [with 17 additional months of follow-up] is only 0.76 [95% CI, 0.54-1.09], which is not quite as impressive as we’ve seen with KEYNOTE-024. This is also a different [PD-L1 measurement] because it’s not the 22C3 Dako but the SP142 assay for assessment, in this trial.

Please describe the design of the cemiplimab (Libtayo) trial in this patient population.

The third registration trial in this setting is EMPOWER-Lung 1 [NCT03088540]. This patient population was all-treatment-naive NSCLC, with PD-L1 greater than 50%; EGFR, ALK, and ROS1 mutations were excluded. They allowed clinically stable central nervous system metastases that were treated, and then patients with hepatitis B or C and HIV were allowed. There is not a good reason not to include them, but they’ve never done it before in this larger setting.

The 1:1 randomization was to cemiplimab or chemotherapy of investigator’s choice; this was an international study, so there was more flexibility. Then, cemiplimab was given for 2 years, and there was a little bit of a different concept: not a second course of treatment, but an addition of chemotherapy. So basically, immunotherapy to chemoimmunotherapy augmentation upon progression versus crossover [from chemotherapy to cemiplimab monotherapy]. About three-quarters of patients took that crossover option, and there weren’t many patients that were able to continue treatment who didn’t choose that option. So this was very interesting, like a sequencing approach to those that were clinically able to.

When you look at the baseline characteristics, you see in both the intent-to-treat [ITT] population—there was a little bit of a difficulty confirming all the PL-L1 levels centrally for tissue availability, so that we have 2 sets of data: PD-L1 greater than 50% confirmed, and then the full ITT population. About 1 in 8 patients had brain metastases; and, as a particular area of interestwe see a little bit of an over-representation of [current] smokers [versus past smokers]. We don’t have any never-smokers in this trial, I think this was felt to be not necessarily helpful for non-smokers as much as it is for smokers. There is also a slightly over-representation of patients with squamous cell disease; obviously, when you exclude non-smokers or light smokers, you see that increase in patients with squamous histology. And, it’s also [related to] the way [cigarettes] are smoked internationally, filter-free, [which can lead to]a little bit of a different precipitation of carcinogens in the airway.

What was the response to therapy on the EMPOWER-Lung 1 study?

We see a picture that’s very similar to KEYNOTE-024. The Kaplan-Meier curves run in parallel for the first 3 to 6 months or so, and then start parting.5,6 In the ITT population, there was median OS of 22.1 months [95% CI, 17.7-NE] for cemiplimab, and 14.3 months [95% CI, 11.7-19.2] for chemotherapy [HR, 0.68; 95% CI, 0.53-0.87; P = .0022].5 The median progression-free survival showed slight increase in the ITT population: 6.2 months for cemiplimab versus 5.5 months for chemotherapy. There is a much larger tail of the curve on an ongoing basis.

So, looking at OS in the confirmed PD-L1 greater than 50% population, it looks better.6 There was certainly the suspicion that there were patients slipped in that didn’t have central confirmation of PD-L1 levels greater than 50%. The HR [of 0.57] is in the same ballpark, maybe even trending a little bit better than with KEYNOTE-024, with an unreached benefit here for OS [95% CI, 17.9-NE] versus 14.2 months [95% CI, 11.2-17.5] of median OS for the chemotherapy arm, and that’s within the same range that we’ve seen for IMpower110, and for KEYNOTE-024.2,4,6 There is almost a 3-month benefit in median PFS [8.2 months for cemiplimab versus 5.7 months for chemotherapy]. But the largest part of benefit here is not extending to the median number of population, but for the strong responder subset.

This was also a crossover study, so there was not necessarily a catching up [from the comparator arm]; I think that’s important for those trials in evaluation.


1. NCCN. Clinical Practice Guidelines for Oncology. Non–small cell lung cancer, version 3.2022. Accessed June 27, 2022.

2. Brahmer JR, Rodriguez-Abreu D, Robinson AG, et al. LBA51 - KEYNOTE-024 5-year OS update: First-line (1L) pembrolizumab (pembro) vs platinum-based chemotherapy (chemo) in patients (pts) with metastatic NSCLC and PD-L1 tumour proportion score (TPS) ≥50%. Ann Oncol. 2020;31(suppl_4):S1142-S1215. doi:10.1016/annonc/annonc325

3. Herbst RS, Giaccone G, de Marinis F, et al. Atezolizumab for first-Line treatment of PD-L1-selected patients with NSCLC. N Engl J Med. 2020;383(14):1328-1339. doi:10.1056/NEJMoa1917346

4. Jassem J, de Marinis F, Giaccone G, et al. Updated overall survival analysis from IMpower110: Atezolizumab versus platinum-based chemotherapy in treatment-naive programmed death-ligand 1-selected NSCLC. J Thorac Oncol. 2021;16(11):1872-1882. doi:10.1016/j.jtho.2021.06.019

5. 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

6. Sezer A, Kilickap S, Gümüş M, et al. LBA52 - EMPOWER-Lung 1: Phase III first-line (1L) cemiplimab monotherapy vs platinum-doublet chemotherapy (chemo) in advanced non-small cell lung cancer (NSCLC) with programmed cell death-ligand 1 (PD-L1) ≥50%. Ann Oncol. 2020;31(suppl_4):S1142-S1215. doi:10.1016/annonc/annonc325