Immune checkpoint inhibition: the picture is slowly completing itself

KEYNOTE-024: first-line, PD-L1–enriched population

The anti–PD-1 antibody pembrolizumab has been approved for treatment of patients with PD-L1–expressing, advanced NSCLC. The KEYNOTE-024 study focused on the first-line comparison of pembrolizumab with platinum-doublet chemotherapy [1]. Chemotherapy regimens comprised five options, two of which (pemetrexed plus carboplatin; pemetrexed plus cisplatin) were used with non-squamous non-small–cell lung cancer (NSCLC) only. In all, 305 patients were randomised across 142 sites in 16 countries. The population was enriched for PD-L1 expression, as a key eligibility criterion was PD-L1 tumour proportion score (TPS) ≥ 50 % (i. e., PDL1 expression on at least 50 % of tumour cells). Approximately 20 % of patients had tumours with squamous histology.
The progression-free survival (PFS) obtained with pembrolizumab was significantly greater than that with the platinum-doublet chemotherapy, which translated into a risk reduction of 50 % (10.3 vs. 6.0 months; HR; 0.50; p < 0.001; Figure 1). PFS at 12 months was 48 % versus 15 % for patients with pembrolizumab and chemotherapy, respectively.
Patients treated with pembrolizumab also experienced significant OS benefit despite 50 % total crossover from chemotherapy (HR, 0.60; p = 0.005). Median OS had not been reached in either arm. At 12 months, 70 % versus 54 % of the patients, respectively, were alive. Likewise, the confirmed objective response rate (ORR) differed by 17 % in favour of pembrolizumab (45 % vs. 28 %; p = 0.0011).

Figure 1: Progression-free survival in KEYNOTE-024: benefit obtained with pembrolizumab compared to chemotherapy

Figure 1: Progression-free survival in KEYNOTE-024: benefit obtained with pembrolizumab compared to chemotherapy

Six complete responses were observed with the anti-PD-1 antibody. Despite longer exposure to pembrolizumab (7.0 vs. 3.5 months), adverse event (AE) rates of all grades were lower in the experimental arm. The Data Monitoring Committee recommended stopping the trial because of this superior efficacy with pembrolizumab.
According to the authors, a PD-L1 TPS ≥ 50 % is detected in approximately one third of patients with advanced NSCLC, and this identifies those most likely to benefit from anti-PD-1 therapy. Pembrolizumab should become a new standard of care as first-line therapy for advanced NSCLC with high levels of PD-L1 expression.

Pembrolizumab plus chemotherapy: KEYNOTE-021

There is a rationale for combining chemotherapy and immunotherapy, as chemotherapy itself has several immunological effects and can induce PD-L1 expression on tumour cells. Clinical data in this area were provided by the multi-cohort phase I/II KEYNOTE-021 trial that evaluated pembrolizumabbased combination therapy for patients with advanced NSCLC [2]. The patients in Cohort G of this study, who had untreated stage IIIB or IV non-squamous NSCLC, were randomised to either pembrolizumab 200 mg every 3 weeks for 2 years plus carboplatin and pemetrexed (n = 60), or carboplatin plus pemetrexed only (n = 63). Pemetrexed maintenance therapy was permitted.
The objective response rate (ORR) according to blinded independent central review, which was defined as the primary endpoint, was almost doubled with the addition of pembrolizumab to chemotherapy (55 % vs. 29 %; p = 0.0016; Figure 2). In the responding population, the time to response was shorter in the experimental arm than in the control arm (1.5 vs. 2.7 months), and a higher percentage of patients showed ongoing responses in the experimental arm (88 % vs. 78 %). Of note, only two pembrolizumab-treated patients experienced primary progression of disease at the initial assessment of 6 weeks (i. e., 3 % vs. 17 % in the control arm). The ORRs were similar for PD-L1 expression of < 1 % and ≥ 1 % in the pembrolizumab arm.
Progression-free survival favoured the pembrolizumab combination, and here the risk of progression or death was nearly halved, with PFS for pembrolizumab plus chemotherapy exceeding 1 year (13.0 vs. 8.9 months; HR, 053; p = 0.0102). OS did not differ between the two arms. At 6 months, 92 % of patients were alive with both treatment regimens. Grade 3/4 AEs were more frequent with the pembrolizumab combination, but this did not translate into higher discontinuation rates. Overall, pembrolizumab in combination with carboplatin and pemetrexed appears to be an effective treatment option for patients with chemotherapy-naïve, advanced non-squamous NSCLC.

Figure 2: Confirmed objective response rates in KEYNOTE-021

Figure 2: Confirmed objective response rates in KEYNOTE-021

OS improvement of 4 months with atezolizumab in OAK

The anti–PD-L1 antibody atezolizumab showed superiority with regard to OS over docetaxel in patients with advanced NSCLC in the phase II POPLAR study [3, 4]. In the randomised phase III setting, the OAK trial compared atezolizumab 1,200 mg every 3 weeks with docetaxel 75 mg/m2 every 3 weeks, in patients with locally advanced or metastatic NSCLC who had received one or two prior lines of chemotherapy, including at least one platinum-based regimen [5]. Patients were recruited regardless of their PD-L1 expression status. Crossover was not allowed. The co-primary endpoint consisted of OS in the ITT population and OS in patients with PD-L1 expression on ≥ 1 % of their tumour cells (TCs) or immune cells (ICs). The OAK data are the first phase III results obtained for a PD-L1–directed antibody, with a total of 1,225 patients projected to be recruited into the study.
The analysis of the first 850 patients showed that OAK met its co-primary endpoint. In the ITT population, atezolizumab treatment was associated with significant and clinically meaningful OS benefit (13.8 vs. 9.6 months; HR, 0.73; p = 0.0003). The survival curves separated early on, at 3 months, and remained separated over time. At 18 months, almost twice as many patients were alive in the atezolizumab arm as in the docetaxel arm (40 % vs. 27 %). A comparable OS benefit was observed for the 55 % of the patient population with PD-L1 expression on ≥ 1 % of their TCs or ICs (TC1/2/3 or IC1/2/3) (15.7 vs. 10.3 months; HR, 0.74, p = 0.0102). However, the subgroups of patients with no or minimal PD-L1 expression (< 1 %; TC0 and IC0) also benefited, with a similar HR of 0.75 (12.6 vs. 8.9 months; p = 0.0205). The greatest OS improvement occurred in the group with the highest PD-L1 expression (on ≥ 50 % of TCs or ≥ 10 % of ICs; TC3 or IC3), which made up 16 % of the total population. Here, the OS benefit achieved with atezolizumab treatment translated into 59 % reduction in mortality risk (median OS, 20.5 vs. 8.9 months; HR, 0.41; p < 0.0001).
As the forest plot for OS by PD-L1 expression shows (Figure 3), the HRs were comparable across all of the subgroups, except for those with the highest expression, where the patients experienced even greater benefit. The OS benefit conferred by atezolizumab is further supported by the 17 % of patients who were randomised into the chemotherapy arm who subsequently received immunotherapy.

Figure 3: OAK trial: OS according to PD-L1 expression for atezolizumab and docetaxel

Figure 3: OAK trial: OS according to PD-L1 expression for atezolizumab and docetaxel

Additional analyses

As patients with both non-squamous and squamous histological subtypes were included in the OAK trial, the investigators also assessed the OS effects of the treatments in these subgroups. In both cohorts, the HRs were 0.73 in favour of atezolizumab. A similar OS advantage was seen across most subgroups irrespective of gender, age, ECOG performance status, number of prior treatment lines, smoking history, and baseline CNS metastasis. The only exception to this were patients with EGFR-activating mutations, who did not benefit from this treatment with the anti–PD-L1 antibody. This phenomenon has already been observed with other PD-L1 inhibitors.
As in previous trials assessing immunotherapy, significant PFS benefit was only seen for the group of patients with the highest PD-L1 expression. Accordingly, response rates only showed benefit for atezolizumab in this high-expression subgroup. Generally, responses lasted considerably longer in the atezolizumab arm than in the docetaxel arm (median duration of response for ITT population, 16.3 vs. 6.2 months). This effect was seen across all of the PD-L1 subgroups.
In spite of the prolonged duration of therapy, atezolizumab was well tolerated. Grade 3/4 AEs were less frequent in the experimental arm, which also applied to AEs that led to withdrawal, dose modification, delays or interruptions. Only musculoskeletal pain and pruritus occurred more frequently with atezolizumab than with docetaxel; for all of the other AEs, the reverse was the case. Immune-mediated AEs were reported with low incidence rates, at below 1 %.

News from the pivotal pembrolizumab trial

The phase III KEYNOTE-010 study demonstrated the efficacy and safety of pembrolizumab in comparison with docetaxel in 1,034 previously treated patients with PD-L1–expressing, advanced NSCLC [6]. This thus provided the basis for European approval of pembrolizumab for this indication. An updated analysis after six additional months of follow-up showed that OS continued to be superior with pembrolizumab 2 mg/ kg and 10 mg/kg compared to docetaxel in the TPS ≥ 50 % and ≥ 1 % populations [7]. PFS was similar to that previously observed, and responses continued to be durable. Overall, these findings confirm pembrolizumab as a standard of care in patients with pre-treated, PD-L1–expressing, advanced NSCLC.
Barlesi et al. assessed the effects of pembrolizumab and docetaxel on health-related quality of life in KEYNOTE-010 using the EORTC QLQ-C30, EORTC QLQ-LC13 and EuroQoL-5D-3L instruments [8]. For changes from baseline to week 12, there were either numeric or significant improvements in the EORTC QLQ-C30 global health status/ quality of life scores for pembrolizumab compared to docetaxel. Compared with docetaxel, pembrolizumab also prolonged the time to deterioration for the EORTC QLQ-LC13 composite endpoint of cough, dyspnoea and chest pain. Along with results from the supportive patient-reported outcome analyses, these findings suggest that the patient health-related quality of life and symptoms were maintained or improved to a greater degree with pembrolizumab than with docetaxel in this population.

Two-year data: CheckMate 017 and 057

Nivolumab is a standard of care for previously treated NSCLC patients based on the results of the global, randomised, open-label, phase III CheckMate 017 and 057 trials. In both studies, nivolumab significantly prolonged OS compared with docetaxel in previously treated patients with squamous NSCLC (CheckMate 017) [9] or with non-squamous NSCLC (CheckMate 057) [10].
The updated efficacy and safety data after ≥ 2 years of follow-up were presented at the ESMO Congress [11]. These showed that in both trials, the improved OS rates for nivolumab over docetaxel remained consistent from year 1 to year 2. Among responders, approximately one third of the nivolumabtreated patients (but none of the docetaxel-treated patients) had ongoing responses. Durable responses occurred regardless of PD-L1 expression levels. No new safety signals were identified for nivolumab therapy. Treatment-related selected AEs were managed using protocol-defined toxicity management algorithms, and these were resolved in the majority of patients.
Reck et al. presented data on the impact of nivolumab versus docetaxel on the overall health status of the patients treated in CheckMate 057 [12]. Both the EQ-5D VAS and the Lung Cancer Symptom Scale indicated better preservation of health status, health-related quality of life, and symptom control with nivolumab compared to docetaxel. Also, both of these assessments hinted at relative improvements in patient-reported outcomes for nivolumab over docetaxel, and suggested that the onset of the benefit occurred prior to the separation of the survival curves, which again favoured nivolumab.

Hardly any first-line benefits in CheckMate 026

Negative results were obtained for firstline nivolumab in patients with stage IV or recurrent PD-L1–positive NSCLC. The open-label, international, phase III, CheckMate 026 study compared firstline nivolumab with platinum-based doublet chemotherapy in this population [13]. PD-L1 expression of ≥ 1 % was mandatory. Crossover to nivolumab in the case of progression was optional. The results for the primary endpoint, which was PFS by independent radiological review in the ≥ 5 % PD-L1–positive population, did not differ significantly between the two regimens (4.2 vs. 5.9 months, for nivolumab and chemotherapy, respectively). This also applied to OS (14.4 vs. 13.2 months). Progressive disease was more common in the nivolumab arm (27.5 % vs. 9.9 %), but when responses were seen, they lasted more than twice as long with nivolumab than in the chemotherapy-treated population (12.1 vs. 5.7 months).
In general, the subgroups mirrored the overall study population. Whereas patients with squamous histology appeared to fare better with regard to PFS and OS when treated with nivolumab, the opposite appeared to be the case for the non-squamous population; however, definite conclusions cannot be drawn due to the overlapping confidence intervals. The CheckMate 227 trial continues to evaluate the role of nivolumab as monotherapy and in combination with ipilimumab or standard chemotherapy in the first-line setting of stage IV or recurrent NSCLC.

Neoadjuvant use of nivolumab

Preliminary, but aspirational, data have been generated with neoadjuvant nivolumab in a trial that enrolled 18 patients with newly diagnosed, resectable stage I (> 2 cm)/II/IIIA NSCLC [14]. The rationale for neoadjuvant use of anti– PD-1 strategies in early-stage NSCLC results from the fact that stage I to III NSCLC, albeit considered early-stage disease, has poor prognosis, and only modest benefits are seen with adjuvant chemotherapy. Nivolumab was administered at a dose of 3 mg/kg at 4 weeks and 2 weeks prior to surgical resection. The primary endpoint was safety and tolerability. Exploratory endpoints included various correlative analyses of blood and the tumour, as well as other clinical outcome parameters, such as pathological response.
These two neoadjuvant doses of nivolumab did not delay or interfere with the surgical resection in any of the patients. According to exploratory analyses of the responses, 22 % of the patients had radiographic response, and seven patients showed pathological down-staging from the pre-treatment clinical stage. Major pathological response was defined as < 10 % residual viable tumour cells at resection. One of the seven patients with a major pathological response experienced a pathological complete response (Figure 4). The tumours of these patients showed immune cell infiltration. Toxicity was consistent with the safety profile observed in other studies with nivolumab, and the treatment was well tolerated. One third of patients experienced treatment- related AEs of any grade, but there was only one grade 3/4 AE. Comprehensive studies on aspects such as genomics and functionality of tumourinfiltrating lymphocytes are ongoing, and larger follow-up clinical studies are planned.

Figure 4: Pathological responses in tumours of 17 patients with early NSCLC after neoadjuvant administration of two nivolumab doses

Figure 4: Pathological responses in tumours of 17 patients with early NSCLC after neoadjuvant administration of two nivolumab doses


  1. Reck M et al., KEYNOTE-024: pembrolizumab vs platinum-based chemotherapy as first-line therapy for advanced NSCLC with a PD-L1 TPS
    ≥50 %. ESMO 2016, abstract LBA8_PR
  2. Langer CJ et al., Randomized phase 2 study of carboplatin and pemetrexed ± Pembrolizumab as first-line therapy for advanced
    NSCLC: KEYNOTE-021 cohort G. ESMO 2016, LBA46_PR
  3. Fehrenbacher L et al., Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): a multicentre,
    open-label, phase 2 randomised controlled trial. Lancet 2016; 387(10030):1837-46
  4. Smith DA et al., Updated survival and biomarker analyses of a randomized phase II study of atezolizumab vs docetaxel in 2L/3L NSCLC
    (POPLAR). J Clin Oncol 34, 2016 (suppl; abstr 9028)
  5. 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
  6. Herbst RS et al., Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-
    010): a randomised controlled trial. Lancet 2016; 387: 1540-1550
  7. Herbst RS et al., Pembrolizumab versus docetaxel for previously treated, PD-L1-expressiong NSCLC: updated outcomes of KEYNOTE-010.
    ESMO 2016, abstract LBA48
  8. Barlesi F et al., Assessment of health-related quality of life in KEYNOTE-010: a phase 2/3 study of pembrolizumab versus docetaxel in patients
    with previously treated advanced NSCLC. ESMO 2016, abstract 1219P
  9. Brahmer J et al., Nivolumab versus Docetaxel in Advanced Squamous-Cell Non-Small-Cell Lung Cancer. N Engl J Med 2015; 373: 123-135
  10. Borghaei H et al., Nivolumab versus Docetaxel in Advanced Nonsquamous Non-Small-Cell Lung Cancer.N Engl J Med 2015; 373:
  11. Barlesi F et al., Long-term outcomes with nivolumab vs docetaxel in patients with advanced NSCLC: CheckMate 017 and Check-
    Mate 057 2-y update. ESMO 2016, abstract 1215PD
  12. Reck M et al., Overall health status in patients with advanced non-squamous NSCLC treated with nivolumab or docetaxel in Check-
    Mate 057. ESMO 2016, abstract 1217PD
  13. Socinski MA et al., CheckMate 026: A phase 3 trial of nivolumab vs investigator’s choice of platinum-based doublet chemotherapy as firstline
    therapy for stage IV/recurrent programmed death ligand 1-positive NSCLC. ESMO 2016, abstract LBA7_PR
  14. Forde PM et al., Neoadjuvant anti-PD-1, nivolumab, in early stage resectable NSCLC. ESMO 2016, abstract LBA41-PR