Immunotherapy: novel anti-PD-L1 antibodies & various combination regimens

OAK subgroup analyses

As compared to anti-PD-1 antibodies, the advantage of antibodies directed against PD-L1 is that they can inhibit PD-1/ PD-L1 interactions while leaving the PD-1/ PD-L2 pathway intact, thus potentially preserving peripheral immune homoeostasis. OAK was the first randomised phase III trial to assess an anti-PD-L1 agent in advanced NSCLC. Patients with locally advanced or metastatic NSCLC received either atezolizumab 1,200 mg every 3 weeks or docetaxel. Prior to the trial, they had already been treated with one or two lines of chemotherapy, including at least one platinum-based regimen. The population was enrolled irrespective of PD-L1 status, and stratified according to PD-L1 expression. OAK had two primary endpoints: OS in the ITT population, and OS in patients with PD-L1 expression on ≥ 1 % of tumour cells or infiltrating immune cells. Cross-over was not permitted, which is of relevance for the interpretation of the OS data.

The primary analysis was presented at the ESMO Congress 2016. Here, OAK met both of the primary endpoints [1]. In the ITT population, atezolizumab treatment resulted in a relative reduction in mortality compared to docetaxel of 27 % (median OS, 13.8 vs. 9.6 months; HR, 0.73; p = 0.0003). Also, atezolizumab improved survival at all levels of PD-L1 expression, with the greatest benefit for the patients with the highest PD-L1 expression. However, atezolizumab also improved survival in patients whose tumour did not express PD-L1.

Subgroup analyses conducted in the OAK trial to evaluate the efficacy of atezolizumab in several clinically relevant subgroups revealed broad efficacy of this treatment [2]. OS benefits were observed regardless of PD-L1 expression levels, as measured by immunohistochemistry (IHC) or gene expression, and of histology (non-squamous vs. squamous) across PD-L1 expression levels, and for all age groups. OS improvement also occurred in never-smokers and in patients with brain metastases at baseline. On the other hand, docetaxel was more effective than atezolizumab in the subgroup of patients with EGFR mutation, while wild-type patients fared better with atezolizumab. This lack of improved efficacy of atezolizumab relative to docetaxel in the EGFR-mutant population has already been noted for other in-pathway agents [3].

BIRCH: promising first-line efficacy of atezolizumab

The single-arm phase II BIRCH study evaluated atezolizumab monotherapy in PD-L1–selected patients with locally advanced or metastatic NSCLC. This trial had three arms, to investigate atezolizumab at a dose of 1,200 mg every 3 weeks as first line, second line, and third/ later lines. PD-L1 expression on tumour cells (TC2 or TC3) and tumour-infiltrating immune cells (IC2 or IC3) was examined by IHC. The primary efficacy endpoint, which related the ORRs to historical controls, has already been met. Garassino et al. presented the data of an exploratory analysis that assessed the first-line portion of the trial [4]. This cohort comprised 138 patients. In this group, 47 % showed the highest PD-L1 tumour expression (TC3 or IC3). Fiftythree percent of the patients had TC2 and IC2.

First-line atezolizumab showed promising monotherapy efficacy. The overall population obtained objective responses and stable disease (SD) in 25 % and 42 %, respectively (Figure 1). In the TC3 or IC3 cohort, ORRs and SD rates were each 34 %. For those in the TC2 and IC2 cohort, these were 18 % and 49 %, respectively. Responses lasted for 16.5 months in the overall population, with the median duration of response of 12.3 months in the TC2 and IC2 population; this has not been established for the TC3 or IC3 cohort yet. The ORR benefit of atezolizumab extended to patients with both mutant and wild-type status for EGFR and KRAS, although the respective patient numbers are small. These results indicate that atezolizumab monotherapy has durable efficacy in the first-line setting.

Figure 1: Response rates with first-line atezolizumab in the BIRCH trial, according to PD-L1 expression status

Figure 1: Response rates with first-line atezolizumab in the BIRCH trial, according to PD-L1 expression status

Median PFS was 7.3 months in the overall population, with similar results across the different levels of PD-L1 expression. After a median follow-up of 22.5 months, median OS was 23.5 months. Again, OS trends were comparable across the PD-L1 expression subgroups, although the median OS estimates are not mature yet. The proportion of patients still alive at 1 year in the overall population was 66.4 %. EGFR and KRAS mutation status did not affect these results. The safety profile was similar to other atezolizumab NSCLC studies, and atezolizumab was well tolerated. Ongoing phase III trials, such as IMpower110, are evaluating atezolizumab compared to chemotherapy in the first-line setting in PD-L1–selected patients.

Durvalumab activity beyond second line in the ATLANTIC trial

Like atezolizumab, durvalumab falls into the category of anti-PD-L1 antibodies. Durvalumab was tested in the open-label, single-arm, phase II ATLANTIC trial at a dose of 10 mg/kg 2-weekly, for up to 12 months [5]. The patients who participated in the trial had at least two prior systemic treatment regimens, including one platinum-based chemotherapy. Initially, the protocol was designed for all comers, but after an amendment, the patient selection was restricted to those with highly PD-L1–expressing tumours. The population consists of three cohorts. Cohort 1 (n = 111) includes patients with EGFRmutation/ ALK aberration and high PDL1 expression (≥ 25 % of tumour cells). Patients in Cohorts 2 and 3 have EGFRALK wild-type. In Cohort 2 (n = 265), PD-L1 expression levels of ≥ 25 % on tumour cells and low/ negative PD-L1 expression (< 25 %) prevails. Cohort 3 (n = 68) includes patients with PD-L1 expression levels ≥ 90 %. The cohorts were independent, and Cohorts 2 and 3 were enrolled sequentially.

In this heavily pre-treated metastatic NSCLC population, durvalumab treatment showed activity and gave rise to durable responses. Stronger PD-L1 expression appeared to be associated with higher response rates. In Cohort 2, the ORRs for patients with low/ negative and high PD-L1 expression were 7.5 % and 16.4 %, respectively (Table). In Cohort 3, the ORR increased to 30.9 %. Disease control rates at ≥ 6 months were 20.4 %, 28.8 % and 38.2 %, respectively. Median duration of response had not been reached yet in Cohort 2 patients with low/ negative expression or in Cohort 3, and was 12.3 months in Cohort 2 patients with high expression. The ORR benefit became apparent across the subgroups; of note, it was independent of the line of treatment and the presence of CNS metastasis.

ATLANTIC: anti-tumour activity of durvalumab in Cohorts 2 and 3

The groups with low/ negative and high PD-L1 expression in Cohort 2 experienced median OS of 9.3 and 10.9 months, respectively. These results corresponded to 1-year OS rates of 34.5 % and 47.7 %, respectively. For Cohort 3, OS had not been reached yet, and 50.8 % of patients were alive at 1 year. Most AEs were classified as low grade, and immune-mediated AEs proved manageable. The authors concluded that these results are consistent with those obtained with other anti-PD-1/ PD-L1 therapies in metastatic NSCLC. Ongoing phase III trials will clarify the role of durvalumab alone or in combination with the CTLA-4 antibody tremelimumab.

Quadruple approach: chemotherapy plus combined immunotherapy

Durvalumab in combination with the CTLA-4 antibody tremelimumab was investigated in the IND.226 dose-escalation trial that focussed on quadruple therapy, thus combining chemotherapy with two immuno-oncological agents. This study is attempting to amplify the benefits of chemotherapy plus immunotherapy by adding not only a PD-L1 inhibitor, but also a CTLA-4 inhibitor. IND.226 includes patients with solid tumours and uses multiple chemotherapy backbones. Twenty-seven patients of the total cohort have been diagnosed with non-squamous NSCLC. They are PD-L1–unselected. Durvalumab 15 mg/kg 3-weekly and tremelilumab 1 mg/kg (multiple doses, 6-weekly) or 3 mg/kg (3 doses, 6-weekly) are being administered together with pemetrexed and cisplatin.

For safety, which is the primary endpoint of this trial, no significant additional toxicity was observed beyond what can be expected from chemotherapy and checkpoint inhibitor therapy with a CTLA-4 and a PD-L1 antibody [6]. As this is a phase I safety study, not all patients were required to have measurable disease. To date, 16 of 26 patients (61.5 %) have experienced partial responses. Stable disease has occurred in seven cases. Treatment is ongoing in many of these patients.

Overall, it was shown that durvalumab and tremelimumab can be safely combined with full doses of pemetrexed/ cisplatin chemotherapy. Future PD-L1 subset analyses will be performed. A phase II randomised follow-up study will compare platinum-based doublet chemotherapy plus durvalumab/ tremelimumab with durvalumab/ tremelimumab alone, in the first-line setting.

JAVELIN: avelumab in a range of solid tumours

Avelumab is another anti-PD-L1 antibody, and it is being tested in the international, phase I, multi-cohort, dose-escalation and dose-expansion JAVELIN Solid Tumor trial. This study enrolled patients with a range of malignancies, which include thoracic cancers and tumours of the skin, head and neck, genitourinary tract, and gastrointestinal tract. Across all of the cohorts, more than 1,700 patients are receiving avelumab 10 mg/kg 2-weekly in the dose expansion phase. Two cohorts with stage IV or recurrent NSCLC have been included; here, patients are treated with avelumab either in the first-line (n = 156) or second-line (n = 184) setting.

At the WCLC, the findings on safety and clinical activity of avelumab in the first-line cohort were reported [7]. These patients are unselected for PD-L1 expression and do not have activating EGFRmutations or ALK translocations. PD-L1 expression is positive in 56.4 % and negative in 14.7 %. Avelumab was shown to be well tolerated. Ten percent of the patients experienced potentially immune-related AEs, but only one patient developed a grade 3 event. No grade 3/4 pneumonitis occurred; grade 1/2 pneumonitis was observed in only four patients (2.6 %).

This early analysis has revealed durable anti-tumour activity of avelumab monotherapy. Complete and partial responses occurred in 22.5 %. Forty-three percent of patients experienced stable disease, which added up to a disease control rate of 65.4 %. The majority of patients showed tumour shrinkage (Figure 2). At data cut-off, 68.6 % of responses were ongoing. For PFS, the analysis yielded a median of 17.6 weeks, with a 24-week PFS rate of 37.2 %.

Figure 2: Waterfall plot from the JAVELIN study depicting the tumour shrinkage obtained with avelumab

Figure 2: Waterfall plot from the JAVELIN study depicting the tumour shrinkage obtained with avelumab

Additional follow-up will further characterise the clinical benefits of avelumab therapy. The analysis of PD-L1 expression as a predictive biomarker for avelumab is ongoing. Currently, a phase III trial is comparing avelumab monotherapy with a platinum-based doublet chemotherapy in untreated, PD-L1–selected NSCLC patients.

Long-term outcomes from CheckMate 012

The CheckMate 012 trial evaluated the anti-PD-1 antibody nivolumab alone versus two schedules for the combination of nivolumab and the CTLA-4 immune checkpoint inhibitor ipilimumab. This treatment was administered as a first-line strategy in patients with stage IIIB/IV NSCLC of any histology. The nivolumab-only arm (n = 52) received nivolumab 3 mg/kg every 2 weeks. In the two combination arms, nivolumab and ipilimumab were administered at doses of 3 mg/kg 2-weekly and 1 mg/kg, respectively, with one arm receiving ipilimumab every 12 weeks (n = 38), and the other, every 6 weeks (n = 39). PD-L1 expression status was assessed. Approximately 70 % of the patients in each arm had PD-L1 expression ≥ 1 %. The primary endpoint of CheckMate 012 was safety and tolerability. Gettinger et al. presented the long-term outcomes of CheckMate 012 at the WCLC [8].

After an additional follow-up of 6 months in the combination cohorts, the rates of treatment-related AEs and the safety profile remained similar to results reported previously. Treatment-related deaths did not occur. Both nivolumab monotherapy and the combinations demonstrated activity, with the combined administration resulting in higher ORRs, longer PFS, and numerically higher 1-year OS rates. At 2 years, ORR was 23 % with nivolumab and 43 % with nivolumab plus ipilimumab. Increasing PD-L1 expression enhanced the efficacy of both the monotherapy and combined treatments. In patients with ≥ 50 % PDL1 expression, ORR was 50 % with nivolumab alone and 92 % with the combination regimens. Likewise, PFS and OS where highest in the ≥ 50 % PDL1 expression groups, although even patients without PD-L1 expression (< 1 %) derived benefit from the treatments. In those with ≥ 1 % PD-L1 expression, median PFS was 3.5 months for nivolumab monotherapy, and 10.4 months and 13.2 months for nivolumab plus ipilimumab 12-weekly and 6-weekly, respectively. At 1 year, 69 %, 91 % and 83 % of these patients were alive, respectively.

Nivolumab 3 mg/kg every 2 weeks plus ipilimumab 1 mg/kg every 6 weeks is being evaluated in further studies. These include the phase III CheckMate 227 trial, which is comparing this regimen to nivolumab monotherapy, chemotherapy, and a regimen of nivolumab plus chemotherapy. The type of the comparison here depends on the PD-L1 expression levels, for which two groups have been defined (≥ 1 % and < 1 %).

Pembrolizumab plus chemotherapy: KEYNOTE-021 G

The combined use of the anti-PD-1 antibody pembrolizumab and chemotherapy as a first-line strategy for the treatment of stage IIIB/IV non-squamous NSCLC was tested in the open-label, randomised, phase II KEYNOTE-021 G study. In the experimental arm, pembrolizumab was administered at a dose of 200 mg every 3 weeks for 2 years, together with carboplatin and pemetrexed. Patients in the control arm received carboplatin and pemetrexed alone over four cycles. Pemetrexed was permitted as maintenance therapy. The primary endpoint was ORR. Approximately 60 patients were treated in each arm, while 20 patients from the control arm crossed over to the pembrolizumab arm when progression set in, and 12 received anti-PD-(L)1 treatment outside of the cross-over.

The confirmed ORR was nearly double with the addition of pembrolizumab (55 % vs. 29 %; p = 0.0016) [9]. According to the PFS analysis, the combination almost halved the risk of progression or death, with median PFS exceeding 1 year (13.0 vs. 8.9 months; HR, 0.53; p = 0.0102; Figure 3). OS was similar between the two arms (92 % at 6 months with both treatments; 75 % and 72 % at 1 year). Pembrolizumab plus chemotherapy showed high tolerability and a manageable safety profile. As the investigators noted, pembrolizumab in combination with carboplatin and pemetrexed could be an effective treatment option for chemotherapy-naïve patients with advanced non-squamous NSCLC.

Figure 3: PFS benefit due to the addition of pembrolizumab to chemotherapy

Figure 3: PFS benefit due to the addition of pembrolizumab to chemotherapy

Harmonisation study on PD-L1 IHC testing in France

PD-L1 expression as assessed by IHC is the main currently available predictive biomarker for the benefit of anti-PD-1/ PD-L1 antibodies. Assays used on the Dako (22C3, 28-8) and Ventana (SP142, SP263) platforms have been used as diagnostic tests in clinical trials. In France, harmonisation of assays and the development of laboratory-developed tests are urgently needed for several reasons. The Dako and Ventana platforms are not available in all pathology laboratories, and assays remain expensive, while PD-L1 testing reimbursement is insufficient to date in France. At the same time, PD-L1 testing has to be rapidly available for patients in the first-line setting, and multiple tests with different assays will not be feasible on small NSCLC samples. A multi-centric French study therefore evaluated the analytical performance of the Dako 28-8 and 22C3, and the Ventana SP263 PD-L1 assays across various centres, with the aim of determining whether laboratory-developed tests can achieve an analytical performance close to PD-L1 assays in a set of NSCLC cases [10]. It was confirmed that the 28-8, 22C3 and SP263 assays performed in several centres showed high agreement. Among 27 laboratory-developed tests developed in seven centres on the Dako, Ventana and Leica platforms, 14 (51.8 %) were in agreement as compared to the reference assays for tumour-cell staining. Low agreement was observed for immune-cell staining when using a four-category scale with 1 %, 5 % and 10 % thresholds. Clone SP263 achieved the highest concordance rate across all of the platforms.

This study also highlights that caution is required for validation and further use of laboratory-developed tests. Selected laboratory-developed tests will be validated on larger cohorts and using external quality assessment programmes in France. These results will provide the basis for national recommendations on PD-L1 testing in NSCLC.


  1. Barlesi F et al., Primary analysis from OAK, a randomized phase III study comparing atezolizumab with docetaxel in 2L/3L NSCLC. ESMO
    2016, abstract LBA44_PR
  2. Gadgeel SM et al., OAK, a randomized ph III study of atezolizumab vs docetaxel in patients with advanced NSCLC: results from subgroup
    analyses. WCLC 2016, PL04a.02
  3. Borghaei H & Brahmer J, Nivolumab in nonsquamous non-small-cell lung cancer. N Engl J Med 2016; 374(5): 493-494
  4. Garassino MC et al., Atezolizumab as 1L therapy for advanced NSCLC in PD-L1–selected patients: updated ORR, PFS and OS data from the
    BIRCH study. WCLC 2016, OA03.02
  5. Garassino MC et al., Durvalumab in ≥3rd-line locally advanced or metastatic, EGFR/ALK wildtype NSCLC: results from the phase 2 ATLANTIC
    study. WCLC 2016, PL04a.03
  6. Juergens RA et al., Pemetrexed/cisplatin + durvalumab ± tremelimumab in patients with advanced non-squamous NSCLC: A CCTG phase Ib study – IND.226 Cohort 1. WCLC 2016, MA09.03
  7. Verschraegen C et al., JAVELIN Solid Tumor: safety and clinical activity of avelumab (anti-PDL1) as first-line treatment in patients with advanced NSCLC. WCLC 2016, OA03.03
  8. Gettinger SN et al., First-line nivolumab monotherapy and nivolumab plus ipilimumab in patients with advanced NSCLC: long-term outcomes from CheckMate 012. WCLC 2016, OA03.01
  9. Langer CJ et al., Pembrolizumab + carboplatin and pemetrexed as 1st-line therapy for advanced non-small cell lung cancer: KEYNOTE-021 Cohort G. WCLC 2016, MA09.02
  10. Adam J et al., Multicentric French harmonization study of PD-L1 IHC testing in non-small cell lung cancer. WCLC 2016, PL04a.04