Pembrolizumab (Keytruda, MK3475), humanized immunoglobulin G4, was approved to treat refractory melanoma in 2015 by the US Food and Drug Administration (FDA) and in Europe by the European Medicines Agency (EMA).27 That year, the FDA and EMA also approved pembrolizumab to treat metastatic NSCLC expressing PDL1 (>1%) and progressing after platinum-based chemotherapy (or after chemotherapy and tyrosine-kinase inhibitors targeting EGFR mutations or ALKrearrangement).6,13 More recently, pembrolizumab was approved by the FDA, EMA, and the Japanese Pharmacological and Medical Devices Agency to treat patients with chemotherapy-naïve, metastatic NSCLC expressing high PDL1 (≥50%) levels without EGFR mutation or ALK rearrangement. Finally, in January 2017, the FDA accelerated approval of the combination of pembrolizumab and pemetrexed–carboplatin chemotherapy (based on KEYNOTE-021 results) as first-line treatment for metastatic or advanced nonsquamous-cell NSCLC without EGFR mutation or ALK rearrangement, regardless of PDL1 expression.
Pembrolizumab’s clinical activity and safety were evaluated in the international multicenter Phase I KEYNOTE-001 trial (NCT01295827) of 495 untreated (101 first-line therapy) or previously treated (394 second-line or more) patients.28 All patients received pembrolizumab intravenously (IV): 2 or 10 mg/kg every 3 weeks (Q3W) or 10 mg/kg Q2W. Primary end points were safety and efficacy: objective response rate (ORR), OS, progression-free survival (PFS), and duration of response (DOR).28Observation during the trial of an association between tumor-cell PDL1 expression and pembrolizumab effectiveness led to a protocol amendment, with the adjunction of another principal outcome criterion: efficacy in patients with high PDL1-expressing tumors.29
In that study, the safety profile of pembrolizumab was acceptable. All-grade treatment-associated adverse events (AEs) combined occurred in 71% of the patients, without any difference among doses given or frequency; grade ≥3 AEs were reported in 10% of patients. ORR for the entire population was 19% (95% CI 16%–23%), with 22% of patients with stable disease. ORR for previously treated patients was 18% (95% CI 14%–22%). Median PFS and OS, respectively, were 3.7 (95% CI 2.9–4.1) and 12.0 (95% CI 9.3–14.7) months at the time of analysis.
After median follow-up of 22 months, median OS rates for treatment-naïve and previously treated patients, respectively, were 22.1 (95% CI 16.8–27.2) and 10.6 (95% CI 8.6–13.3) months.30 ORR, 12-month PFS, and 12-month OS, respectively, were 51.9%, 54%, and 85% for patients with ≥50% PDL1-expressing tumor cells compared to 26.7%, 35%, and 71% for the overall population. Those updated results confirmed pembrolizumab safety, with only 12 (11.9%) patients experiencing treatment-associated grade 3/4 AEs and no deaths.28,30,31 Those findings, promising in terms of safety and efficacy, led to several Phase I–III trials (Table 1) to evaluate pembrolizumab alone or in combination with other treatments for patients with advanced NSCLC.10–12
(To view a larger version of Table 1, click here.)
Pembrolizumab was approved as second-line therapy for NSCLC based on the results of the open Phase II–III randomized KEYNOTE-010 trial,10 which evaluated its efficacy and safety vs docetaxel for patients with advanced NSCLC whose cells expressed low PDL1 (≥1%) levels. Patients were randomized to receive one of three treatments Q3W: pembrolizumab 2 (n=345) or 10 mg/kg (n=346), or docetaxel 75 mg/m2 (n=343). Primary end points were OS and PFS for all patients and the subgroup with tumors expressing ≥50% PDL1. Using PDL1 expression as a companion test, and its pertinence, will be discussed in the section on biomarkers. Secondary end points included tolerance, ORR, and DOR. The trial included 1,034 patients. Compared to docetaxel, the 2 and 10 mg/kg pembrolizumab doses, respectively, prolonged OS – 10.4 (95% CI 9.4–11.9), 12.7 (95% CI 10.0–17.3) vs 8.5 (95% CI 7.5–9.8) months – with no difference between the two pembrolizumab doses (hazard ratio [HR] 1.17, 95% CI 0.94–1.45), but PFS rates did not differ significantly. After subgroup analyses, no significant OS benefit was observed with pembrolizumab compared to docetaxel for women (HR 1.02, 95% CI 0.78–1.32), patients with EGFR-mutated tumors (HR 1.79, 95% CI 0.94–3.42), or those whose tumors had 1%–49% PDL1 expression (HR 1.04, 95% CI 0.85–1.27).
Pembrolizumab efficacy against tumors with ≥50% PDL1 expression was a principal end-point cocriterion. Among the 2,222 patients screened, 1,475 (66%) had ≥1% PDL1 expression, with 633 (28%) having ≥50% PDL1 and thus were assessable for the cocriterion. Those patients’ median OS rates for the 2 and 10 mg/kg pembrolizumab and docetaxel groups, respectively, were 14.9 (95% CI 10.4–not reached), 17.3 (11.8–not reached), and 8.2 (6.4–10.7) months. OS was comparable for the two pembrolizumab arms. Pembrolizumab’s safety profile was better than that of docetaxel, with fewer grade 3/4 AEs, in agreement with Phase I trial results.28 Grade ≥3 AEs were reported in 13% (43 of 339), 16% (55 of 343), and 35% (109 of 309) of patients given 2 or 10 mg/kg pembrolizumab or docetaxel, respectively. Among the eleven treatment-attributed deaths, three occurred in the 2 mg/kg pembrolizumab arm (two interstitial and one infectious pneumopathy), three in 10 mg/kg pembrolizumab recipients (one myocardial infarction and one interstitial or one infectious pneumopathy), and five in the docetaxel group. Immune-related AEs (IRAEs), regardless of grade, occurred in 20% (69 of 339) and 19% (64 of 343) of patients who received 2- or 10 mg of kg pembrolizumab, respectively. The most common IRAEs were hypothyroidism, hyperthyroidism and interstitial pneumopathy.
Several clinical trials have evaluated first-line pembrolizumab. An analysis of 101 treatment-naïve patients included in the Phase-1 (KEYNOTE-001) trial was published in 2017.30 ORR (51.9%), 12-month PFS (54%), and 12-month OS (84%) for patients with PDL1 ≥50% NSCLC were better than for the entire population (26.7%, 35%, and 71%, respectively). Those findings led to Phase I–III trials testing the molecule as first-line therapy.
The Phase III randomized KEYNOTE-024 trial (NCT02142738) compared pembrolizumab (200 mg IV, 3QW) until disease progression or unacceptable toxicity for a maximum of 2 years vs standard platinum-based doublet chemotherapy (four to six cycles with maintenance therapy possible) as first-line therapy for advanced NSCLC without EGFR mutation or ALK rearrangement, but with PDL1 ≥50% determined by immunohistochemistry (IHC).11 The primary end point was PFS and secondary-outcome criteria were ORR, OS, and safety. Patients with untreated brain metastases, active autoimmune disease requiring systemic therapy, or performance status (PS) ≥1 were excluded. Patients who received platinum-based chemotherapy (cisplatin 75 mg/m2 IV or carboplatin area under the carboplatin area under curve (AUC) 5–6 IV Q3W) in combination with pemetrexed (500 mg/m2 IV Q3W), gemcitabine (1,250 mg/m2 IV Q3W), or paclitaxel (200 mg/m2 IV Q3W with carboplatin) served as controls. Among 1,934 patient tumors screened for PDL1, 500 (25.9%) showed PDL1 ≥50%. Finally, 305 patients were included and randomized (154 to the pembrolizumab arm and 151 to the control group).11 In the case of disease progression in the control group, crossover was authorized.
At the time of analysis, 50% of chemotherapy-treated controls had switched to pembrolizumab because of progression vs 23% of the pembrolizumab-arm patients. The main results are summarized in Table 1. Median PFS rates were 10.4 vs 6.0 months for the pembrolizumab and chemotherapy arms, respectively (HR 0.5, 95% CI 0.37–0.68; P<0.001). OS was also significantly better for patients in the pembrolizumab arm, with the median not reached for either arm. Estimated OS rates at 6 months were 80.2% and 72.4%, respectively, for the pembrolizumab and control arms (HR 0.60, 95% CI 0.41–0.89; P=0.005), with respective ORRs of 45% and 28% (P=0.0011). A PFS benefit was observed for all the predefined subgroups.
The data were updated at the 2017 World Conference on Lung Cancer.32 At data cutoff (July 10, 2017) after a median follow-up of 25.2 months, the HR for OS was 0.63 (95% CI 0.47–0.86, P=0.002). Median OS was 30.0 (18.3–not reached) months for the pembrolizumab arm and 14.2 (9.8–19.0) months for the chemotherapy arm. Those groups’ respective 12-month OS rates were 70.3% (95% CI 62.3%–76.9%) vs 54.8% (95% CI 46.4%–62.4%).32
All-grade AEs combined occurred in 73.4% and 90.0% of patients in the pembrolizumab and control groups, respectively.11 Grade 3/4 AEs were observed in 26% of pembrolizumab recipients and 51% of chemotherapy-treated controls. The main grade 3/4 non-IRAEs associated with pembrolizumab were diarrhea (3.9%), anemia (1.9%), and fatigue (1.3%). The overall IRAE frequency under pembrolizumab was 29.2%, among which 9.7% was grade ≥3, mainly pulmonary (2.6%), cutaneous (3.9%), and digestive (1.3%).
In that trial,33 patients’ quality of life was evaluated with the Quality of Life Questionnaire – cancer (QLQ-C)-30 and QLQ – lung cancer (QLQ LC)-13. Mean variations in QLQ-C30 scores compared to their initial values were 6.9 (95% CI 0.3–10.6) points for pembrolizumab and −0.9 (95% CI −4.8 to 3.0) for the controls, for a difference of 7.8 (95% CI 2.9–12.8, P=0.002). Fewer pembrolizumab-treated patients than chemotherapy controls (31% vs 39%, respectively) had poorer QLQ-LC13 composite scores, with longer time to deterioration under pembrolizumab than chemotherapy (HR 0.66, 95% CI 0.44–0.97; P=0.029).33 These findings led to premature termination of the trial by the independent monitoring committee, with a recommendation to offer chemotherapy-treated patients the option to receive pembrolizumab.11 Based on these results, the FDA, EMA, and Japanese Pharmacological and Medical Devices Agency approved pembrolizumab for first-line treatment of metastatic NSCLC with elevated PDL1 (≥50%) expression.