Pembrolizumab or Placebo Plus Chemotherapy With or Without Bevacizumab for Persistent, Recurrent, or Metastatic Cervical Cancer: Subgroup Analyses From the KEYNOTE-826 Randomized Clinical Trial

Abstract

Importance: The KEYNOTE-826 randomized clinical trial showed statistically significant and clinically meaningful survival benefits with the addition of pembrolizumab to chemotherapy with or without bevacizumab in patients with persistent, recurrent, or metastatic cervical cancer. Treatment effects in patient subgroups of the study population are unknown.

Objective: To assess efficacy outcomes in patient subgroups of KEYNOTE-826.

Design, setting, and participants: Exploratory subgroup analyses were conducted in a global, phase 3, randomized, double-blind, placebo-controlled clinical trial. Participants included women with persistent, recurrent, or metastatic adenocarcinoma, adenosquamous carcinoma, or squamous cell carcinoma of the cervix that had not been treated with systemic chemotherapy and was not amenable to curative treatment. This subanalysis was conducted from November 20, 2018, to May 3, 2021.

Interventions: Pembrolizumab, 200 mg, every 3 weeks or placebo for up to 35 cycles plus chemotherapy (paclitaxel, 175 mg/m2, plus cisplatin, 50 mg/m2, or carboplatin AUC 5 [area under the free carboplatin plasma concentration vs time curve]) with or without bevacizumab, 15 mg/kg.

Main outcomes and measures: Overall survival (OS) and progression-free survival (PFS) by investigator assessment per Response Evaluation Criteria in Solid Tumors version 1.1 in subgroups defined by use of bevacizumab (yes or no), choice of platinum (carboplatin or cisplatin), prior chemoradiotherapy (CRT) exposure only (yes or no), and histologic type (squamous or nonsquamous) in patients with programmed cell death ligand 1-positive tumors (defined as a combined positive score [CPS] ≥1) and in the intention-to-treat population.

Results: A total of 617 patients (median age, 51 years; range, 22-82 years) were enrolled in the trial. In the CPS greater than or equal to 1 population, hazard ratios (HRs) for OS favored the pembrolizumab group in all subgroups: with bevacizumab (HR, 0.62; 95% CI, 0.45-0.87) and without bevacizumab (HR, 0.67; 95% CI, 0.47-0.96), use of carboplatin (HR, 0.65; 95% CI, 0.50-0.85) and cisplatin (HR, 0.53; 95% CI, 0.27-1.04), with prior CRT only (HR, 0.56; 95% CI, 0.39-0.81) and without prior CRT only (HR, 0.72; 95% CI, 0.52-1.00), and squamous (HR, 0.60; 95% CI, 0.46-0.79) and nonsquamous (HR, 0.70; 95% CI, 0.41-1.20) histologic type. In the intention-to-treat population, HRs for OS also favored the pembrolizumab group in all subgroups: with bevacizumab (HR, 0.63; 95% CI, 0.47-0.87) and without bevacizumab (HR, 0.74; 95% CI, 0.53-1.04), use of carboplatin (HR, 0.69; 95% CI, 0.54-0.89) or cisplatin (HR, 0.59; 95% CI, 0.32-1.09), with prior CRT only (HR, 0.64; 95% CI, 0.45-0.91) and without prior CRT only (HR, 0.71; 95% CI, 0.53-0.97), and squamous (HR, 0.61; 95% CI, 0.47-0.80) and nonsquamous (HR, 0.76; 95% CI, 0.47-1.23) histologic type. Similar to OS, the addition of pembrolizumab prolonged PFS across all subgroups in the CPS greater than or equal to 1 and intention-to-treat populations.

Conclusions and relevance: The findings of this trial suggest that adding pembrolizumab to chemotherapy with or without bevacizumab improved OS across subgroups of patients with persistent, recurrent, or metastatic cervical cancer.

Trial registration: ClinicalTrials.gov Identifier: NCT03635567.

Figure 1.. Trial Profile in the Intention-to-Treat Population

CRT indicates chemoradiotherapy. ^aTreatment selection was per investigator’s choice.

Figure 2.. Overall Survival by Bevacizumab Use

Estimates of overall survival in subgroups of the programmed cell death ligand 1 combined positive score (CPS) greater than or equal to 1 population with concomitant bevacizumab use (A) and without concomitant bevacizumab use (B), and of overall survival with use of bevacizumab (C). The assigned regimen in both groups also included paclitaxel and the investigator’s choice of cisplatin or carboplatin. Tick marks in panels A and B indicate censored data. HR indicates hazard ratio.

Figure 3.. Overall Survival by Choice of Platinum

Estimates of overall survival in subgroups of the programmed cell death ligand 1 combined positive score (CPS) greater than or equal to 1 population with carboplatin (A) and cisplatin (B), and of overall survival by choice of platinum (C). The assigned regimen in both groups also included paclitaxel and per investigator discretion, bevacizumab. Tick marks in panels A and B indicate censored data. HR indicates hazard ratio.

Figure 4.. Overall Survival by Prior Chemoradiotherapy (CRT) Only

Estimates of overall survival in subgroups of the programmed cell death ligand 1 combined positive score (CPS) greater than or equal to 1 population with prior CRT only (A) and without prior CRT only (B), and of overall survival by receipt of prior CRT only (C). The assigned regimen in both groups also included paclitaxel, the investigator’s choice of cisplatin or carboplatin, and per investigator discretion, bevacizumab. Tick marks in panels A and B indicate censored data. HR indicates hazard ratio.

Figure 5.. Overall Survival by Histologic Type

Estimates of overall survival in subgroups of the programmed cell death ligand 1 combined positive score (CPS) greater than or equal to 1 population with squamous histologic type (A) and nonsquamous histologic type (B), and analysis of overall survival by histologic type (C). The assigned regimen in both groups also included paclitaxel, the investigator’s choice of cisplatin or carboplatin, and per investigator discretion, bevacizumab. Tick marks in panels A and B indicate censored data. HR indicates hazard ratio.