KEYNOTE-826: Sustained survival benefit with pembro + chemo in advanced cervical cancer

The addition of pembrolizumab immunotherapy to chemotherapy, either with or without bevacizumab, continues to confer survival benefit in patients with persistent, recurrent, or metastatic cervical cancer, regardless of their programmed death-ligand 1 (PD-L1) combined positive score (CPS), according to the updated results of the KEYNOTE-826 trial presented at ASCO 2023.

“The updated findings are consistent with the previous interim data and provide further support for first-line pembrolizumab + chemo, with or without bevacizumab, as a new standard of care for women with persistent, recurrent, or metastatic cervical cancer,” said lead author Dr Bradley Monk from Creighton University School of Medicine in Phoenix, Arizona, US.

“At the protocol-specified final analysis of KEYNOTE-826, with a median follow-up of more than 3 years, the addition of pembrolizumab + chemo, with or without bevacizumab, continued to show substantial and clinically meaningful overall survival (OS) and progression-free survival (PFS) improvements in patients with advanced cervical cancer,” Monk noted.

After a median follow-up of 39.1 months, median OS remained significantly longer in the pembrolizumab + chemo arm compared with the placebo + chemo arm across all prespecified populations (28.6 vs 16.5 months [PD-L1 CPS ≥1], 29.6 vs 17.4 months [PD-L1 CPS ≥10], and 26.4 vs 16.8 months [all-comer]). [ASCO 2023, abstract 5500]

The combination of pembrolizumab + chemo also significantly reduced the risk of death by 40 percent and 42 percent in patients with PD-L1 CPS ≥1 or ≥10, respectively, and by 37 percent in the all-comer population (p_nominal<0.0001 for all).

Similarly, the median PFS was maintained in the pembrolizumab + chemo arm vs the placebo + chemo arm in the PD-L1 CPS ≥1 (10.5 vs 8.2 months; hazard ratio [HR], 0.58; p_nominal<0.0001), PD-L1 CPS ≥10 (10.4 vs 8.1 months; HR, 0.52; p_nominal<0.0001), and all-comer populations (10.4 vs 8.2 months; HR, 0.61; p_nominal<0.0001).

The 12-month PFS rate was also higher among those treated with pembrolizumab + chemo compared with placebo + chemo across all populations (45.6 percent vs 33.7 percent [PD-L1 CPS ≥1], 44.7 percent vs 33.5 percent [PD-L1 CPS ≥10], and 44.7 percent vs 33.1 percent [all-comer]).

The pembrolizumab + chemo arm also continued to show improved objective response rate over the placebo arm in the PD-L1 CPS ≥1 (68.5 percent vs 50.9 percent), PD-L1 CPS ≥10 (66.2 percent vs 51.5 percent), and all-comer populations (69.6 percent vs 50.3 percent), with a longer median duration of response of 19.2, 18.0, and 28.3 months, respectively.

“Overall, the pembrolizumab + chemo benefit was seen regardless of bevacizumab use,” Monk said.

Adverse events

Grade ≥3 adverse events (AEs) occurred at a higher rate in the pembrolizumab + chemo arm vs the placebo + chemo arm (82.4 percent vs 75.4 percent), leading to death in 5.2 percent and 4.9 percent of the patients, respectively. Anaemia, alopecia, nausea, and diarrhoea were the most common grade ≥3 AEs reported in both treatment arms.

“Nevertheless, the safety profile for pembrolizumab + chemotherapy was manageable, consistent with prior observations, and [there were] no new early signals,” said Monk.

This double-blind, phase III trial involved 617 patients with persistent, recurrent, or metastatic cervical cancer who had not already received chemotherapy and were not eligible to receive surgery or radiation therapy. Participants were randomized to receive pembrolizumab 200 mg (n=308) or placebo (n=309) every 3 weeks for up to 35 cycles in addition to chemo (paclitaxel 175 mg/m² + cisplatin 50 mg/m² or carboplatin AUC 5), with or without bevacizumab 15 mg/kg. Majority of the patients had tumours that expressed PD-L1, with 88.8 percent having a PD-L1 CPS of ≥1 and 51.4 percent having a PD-L1 CPS of ≥10.

“Before KEYNOTE-826, the standard of care was a platinum-based paclitaxel chemotherapy combination with or without bevacizumab treatment for people with this diagnosis,” said Monk.