Modest survival benefits were observed in patients with extensive-stage small cell lung cancer who received the combination of pembrolizumab and etoposide plus platinum compared with patients who received EP and placebo.
Charles M. Rudin, MD
Modest survival benefits were observed in patients with extensive-stage small cell lung cancer (ES-SCLC) who received the combination of pembrolizumab (Keytruda) and etoposide plus platinum (EP) compared with patients who received EP and placebo. Although progression-free survival (PFS) rates reached the threshold for significance, OS rates failed to reach the prespecified threshold, according to data from the phase 3 KEYNOTE-604 trial (NCT03066778).1
Results from KEYNOTE-604 were presented at the 2020 ASCO Virtual Scientific Program by Charles M. Rudin, MD, and were simultaneously released in the Journal of Clinical Oncology.2 Prior data from the KEYNOTE-028 and KEYNOTE-158 trials signaled that pembrolizumab immunotherapy could elicit durable responses with a tolerable safety profile, which addresses an issue of highly aggressive disease at diagnosis that is often observed with ES-SCLC.
Rudin stated that "SCLC remains an exceptional aggressive disease. About two-thirds of patients present with metastatic disease at the time of diagnosis and despite a high response rate to initial chemotherapy, these are rarely durable."
In addition to pembrolizumab, combination regimens like atezolizumab (Tecentriq), an anti-PD-L1 therapy plus EP, as seen in the phase 3 IMpower 133 study (NCT02763579), as well as durvalumab (Imfinzi) in combination with EP, as assessed in the phase 3 CASPIAN trial (NCT03043872), significantly improved OS compared with EP alone.
Rudin et al aimed to improve upon the efficacy of immunotherapy in newly diagnosed ES-SCLC with the combination of pembrolizumab and EP in KEYNOTE-604.
A total of 453 patients were randomized 1:1. In the experimental arm, 228 patients received pembrolizumab 220 mg on day1 plus EP 100mg/m2 on days 1 and 2 and carboplatin AUC 5 on day 1 or cisplatin 75mg/m2 on day 1. In the control arm, 225 patients received matching doses of placebo, matching EP, and carboplatin or cisplatin. Pembrolizumab and placebo were continued for up to 31 cycles.
Patients were eligible to enroll in the study if they had stage IV SCLC with an ECOG performance status of 0 or 1, a sample for use in biomarker testing, adequate organ function, and a life expectancy of at least 3 months. Patients could not have prior systemic therapy or unstable brain metastases. To receive platinum-based chemotherapy, patients were required to have an ECOG performance status of 0 versus 1 and a lactate dehydrogenase (LDH) level of greater than or equal to the upper limits of normal (ULN) versus less than or equal to the ULN.
The trial is investigating the co primary end points of PFS per RECIST v1.1 by blinded independent central review and OS. The secondary end points are overall response rate (ORR) and duration of response (DOR) per RECIST v1.1 by BICR as well as safety.
KEYNOTE-604 analyzed efficacy in an intention-to-treat (ITT) population and safety in an as-treated population. The analyses included a second interim analysis with a data cutoff of March 29, 2019 and a final analysis with a data cutoff date of December 2, 2019.
Patients were screened at baseline and of the 228 patients in the pembrolizumab arm, the median age was 64 years (range, 24-81). The majority of patients (66.7%) were male. In terms of fitness for treatment and medical history, 73.7% (168 of 228) of patients had an ECOG performance status of 1, 96.5% (220 of 228) were former smokers, 55.7% (127 of 228) had an LDH level above ULN, 14.5% (33 of 228) had brain metastases, 41.7% (95 of 228) had liver metastases, the sum of the largest diameters of target lesion was 134.8 (range, 24.4-431.7), and 38.6% (88 of 228) of patients had a PD-L1 complete positive score (CPS) of at least 1.
Of the 225 patients in the placebo arm, the median age was 65 years (range, 37-83) and 63.1% of patients were male. An ECOG performance status of 1 was seen in 75.1% of patients (169 of 225), 96.4% (217 of 225) were former or current smokers, 57.3% had an LDH greater than ULN, 9.8% (22 of 225) had brain metastases, 40.9% (92 of 225) had liver metastases. The sum of the largest diameters of the target lesion in the placebo arm was 126.6 (range, 20.8-408.8), and 43.1% (97 of 225) of patients had a PD-L1 CPS of 1 or above. For the option of carboplatin or cisplatin, 70.6% of participants in the pembrolizumab group received carboplatin, as did 69.3% of patients (n = 156) in the placebo group.
The initial PFS analysis showed that the median PFS among patients who received pembrolizumab was higher than in the placebo arm (HR, 0.75; 95% CI, 0.61-0.91; P = .0023) in the intention-to-treat (ITT) population. Throughout the period that chemotherapy was administered the survival curves overlapped, said Rudin, which resulted in similar median PFS. After chemotherapy administration, the curves diverged in favor of pembrolizumab. The median PFS was 4.5 months (95% CI, 4.3-5.4) in the pembrolizumab arm versus 4.3 months (95% CI, 4.2-4.4) in the placebo arm. At 6 months, the PFS rate was 34.1% in the pembrolizumab arm versus 23.8% in the placebo arm. At 12 months, the PFS rate was 13.6% in the pembrolizumab arm compared with 3.1% in the placebo arm.
In the final PFS analysis, the separation of the survival curves was maintained (HR, 0.73; 95% CI, 0.60-0.88). The median PFS was 4.8 months (95% CI, 4.3-5.4) compared with 4.3 months (95% CI, 4.2-4.5) in the placebo arm. The 12-month PFS rate observed with the pembrolizumab combination was 15.9% versus 5.0% with the placebo combination. Even at 18 months, the PFS rate in the pembrolizumab arm was higher than the placebo arm at 10.8% versus 2.1%.
In terms of overall survival in the ITT population, pembrolizumab/EP prolonged OS compared with the control combination (HR, 0.80; 95% CI, 0.64-0.98; P = .0164). Rudin noted, "While the 95% CIs did not cross 1, the P-value of 0.0164 narrowly missed the superiority threshold, which was 0.0128." The median OS was 10.8 months (95% CI, 9.2-12.9) with the addition of pembrolizumab to EP. On the placebo arm, the median OS was 9.7 months (range, 8.6-10.7 months). The 12-month OS rate was 45.1% in the pembrolizumab arm compared with 39.6% in the placebo arm. At 24 months, the OS rate was 22.5% in the pembrolizumab arm compared with 11.2% in the placebo arm.
In the as-treated population, the OS curves were identical, but the HR decreased to 0.78 (95% CI, 0.63-0.97; nominal P = .0124), according to an exploratory analysis. This result in particular further supports the benefit of adding pembrolizumab to EP, Rudin stated.
Survival outcomes in the various subgroups in the study were similar between the 2 arms, with the brain metastases population being the only outlier, by favoring placebo.
The ORR in the ITT population for pembrolizumab/EP was 70.6% (95% CI, 64.2%-76.4%), which included complete responses (CRs) in 1.8% (4) of patients, partial responses (PRs) in 68.9% (157) of patients, stable disease (SD) in 17.5% (40) of patients, and progressive disease (PD) in 3.5% (8) of patients. The best response was not evaluable (NE) in 2.6% (5) of patients in the pembrolizumab arm, and not applicable (NA) in 5.7% (13) of patients . The ORR observed with placebo/EP was 61.8% (95% CI, 55.1%-68.2%) of which best response were, CRs in 0.9% (2), PRs in 60.9% 137) of patients, SD in 24.9% (56) of patients, and PD in 5.3% of patients (12). Best responses were NE in 2.2% of patients who received the placebo combination (5) and NA in 5.8% (13).
Responses were more durable among subjects who received pembrolizumab. Similar to the survival activity, the DOR curves overlapped at first, then separated in favor of pembrolizumab after about 5 months. The median DOR in the pembrolizumab/EP arm was 4.2 months (95% CI, 1.0-26.0) and the median DOR in the placebo/EP arm was 3.7 months (95% CI, 1.4-25.8). The DOR rate at 12 months was 19.3% with pembrolizumab versus 3.3% with the control and at 18 months, the DOR rate was 16.3% versus 1.3%, respectively.
The safety analysis showed that adverse events (AEs) of any-grade occurred in 100% of patients in the pembrolizumab arm and 99.6% of patients in the placebo arm, in the as-treated population. AEs were grade 3/4 in 76.7% of subjects who received the addition of pembrolizumab to EP compared with 74.9% of those who received the placebo combination. Grade 5 AEs, which led to death were also observed in 6.3% of patients in the pembrolizumab arm versus 5.4% in the control arm. Additionally, a higher percentage of patients in the pembrolizumab arm discontinued treatment on any study drug (14.8% versus 6.2%), and the percentage of patients who discontinued all treatments was 4.0% versus 3.6%, respectively.
The most common AEs were hematologic and were not intensified by the addition of pembrolizumab. AEs in the pembrolizumab arm versus the control arm, respectively were neutropenia (57.0% vs. 53.4%), anemia (48.4% vs. 46.6%), nausea (38.6% vs. 43.0%), alopecia (33.6% vs. 37.7%) decreased appetite (30.9% vs. 24.7%), constipation (29.6% vs. 26.5%), fatigue (27.4% vs. 27.4%), thrombocytopenia (26.5% vs. 22.0%), leukopenia (22.4% vs. 20.6%), diarrhea (21.1% vs. 18.8%), and cough (19.7% vs. 20.2%).
Any grade immune-related AEs in the as-treated population were experienced by 24.7% of patients in the pembrolizumab arm and 10.3% of those in the placebo arm. Immune-related AEs were grade 3 or 4 in severity for 7.2% of the pembrolizumab group, and 1.3% of the placebo group. No events caused death in the pembrolizumab group, but 0.4% of immune-related AEs led to death in the placebo arm. In terms of treatment discontinuation, 5.8% of patients who received pembrolizumab discontinued any study treatment versus 0.9% of patients who received the placebo combination. Discontinuation of all treatment in the study occurred in 0.9% of patients in the pembrolizumab arm and no patients in the placebo arm.
Of the immune-related AEs, the most common were hypothyroidism (10.3% vs. 2.2%), hyperthyroidism (6.7% vs. 2.7%), and pneumonitis (4.0% vs. 2.2%).
Overall, Rudin concluded that this safety profile was expected with pembrolizumab and the toxicities were manageable.
The efficacy and safety data taken together support the use of pembrolizumab to improve upon the capabilities of EP as treatment of patients with ES-SCLC.