Vibostolimab Plus Pembrolizumab Fails to Boost Efficacy in Pretreated PD-L1+ Cervical Cancer

Treatment with a coformulation of vibostolimab (MK-7684) and pembrolizumab (Keytruda) did not lead to improvements in overall response rate (ORR), progression-free survival (PFS), or overall survival (OS) compared with pembrolizumab alone in previously treated patients with metastatic cervical cancer with a PD-L1 combined positive score (CPS) of at least 1, according to data from the phase 2 KEYVIBE-005 trial (NCT50007106).¹

Findings presented at the 2024 ESMO Gynecological Cancers Congress showed that patients enrolled in cohort A1 who received vibostolimab plus pembrolizumab (n = 85) experienced an ORR of 20% (95% CI, 12%-30%) per blinded independent central review (BICR) assessment compared with 16% (95% CI, 9%-25%) for those given pembrolizumab alone (n = 84; difference, 5%; 95% CI, –7% to 16%; P = .2215). The complete response (CR) and partial response (PR) rates in the combination arm were 7% and 13%, respectively. These respective rates were 7% and 8% in the pembrolizumab monotherapy arm. The median duration of response (DOR) was 10.9 months (range, 5.3–not reached [NR]) in the combination arm vs NR (95% CI, NR-NR) in the single-agent arm.

The study also included patients with a PD-L1 CPS of less than 1, and they were enrolled to cohort A2 to receive vibostolimab plus pembrolizumab. In this patient population, the combination elicited an investigator-assessed ORR of 16% (95% CI, 6%-34%), which was comprised exclusively of PRs. The median DOR was 10.8 months (range, 9.6-NR).

“Importantly, if [patients] do respond—regardless of what [treatment] you give them—those who [do so] respond for a long time,” lead study author Alexandra Leary, MD, PhD, of the Medicine Department at Gustave Roussy in Villejuif, France, said in a presentation of the data.

In cohort A1, patients with a PD-L1 CPS of at least 1 treated with vibostolimab plus pembrolizumab achieved a BICR-assessed median PFS of 2.2 months (95% CI, 2.1-4.2) compared with 2.1 months (95% CI, 2.1-2.3) for those administered pembrolizumab alone (HR, 0.99; 95% CI, 0.70-1.38). The 12-month PFS rates were 16.8% and 19.8%, respectively.

Patients in cohort A2 experienced a median PFS of 2.2 months (95% CI, 2.0-4.2) per investigator assessment and a 12-month PFS rate of 19.4%.

Vibostolimab is a humanized IgG1 anti-TIGIT antibody designed to engage with Fcγ receptors on myeloid cells. Prior data from the phase 1 KEYVIBE-001 trial (NCT02964013) showed that the combination of vibostolimab and pembrolizumab was safe and produced antitumor activity in patients with advanced cervical cancer, irrespective of PD-L1 status.²

KEYVIBE-005 was an open-label study further evaluating the combination in patients with PD-1/PD-L1 inhibitor–naive, unresectable or metastatic cervical cancer who received at least 1 prior line of therapy. Patients underwent biomarker screening to determine PD-L1 status; those with a CPS of at least 1 were assigned to cohort A1, and those with a CPS of less than 1 were enrolled into cohort A2.¹

Within cohort A1, patients were randomly assigned 1:1 to receive coformulated vibostolimab at 200 mg plus pembrolizumab at 200 mg or pembrolizumab alone at 200 mg. In cohort A2, all patients received the same coformulated regimen of vibostolimab plus pembrolizumab. Notably, prior bevacizumab (Avastin) use (yes vs no) served as a stratification factor in cohort A1. Treatment in all arms continued until unacceptable toxicity, disease progression, or patient withdrawal.

BICR-assessed ORR and PFS per RECIST 1.1 criteria in cohort A1, as well as investigator-assessed ORR per RECIST 1.1 criteria in cohort A2, served as the trial’s primary end points.

In cohort A1, the median age was 53.0 years (range, 23-83) in the combination arm and 48.5 years (range, 21-42) in the single-agent arm. Other key baseline characteristics included an ECOG performance status of 0 (vibostolimab/pembrolizumab, 45%; pembrolizumab, 51%) or 0 (55%; 49%); 0 prior lines of therapy (5%; 2%), 1 prior line of therapy (59%; 62%), 2 prior lines of therapy (27%; 20%), or 3 or more prior lines of therapy (9%; 15%); prior neoadjuvant or adjuvant therapy (38%; 30%); prior bevacizumab (48%; 46%); a PD-L1 CPS of at least 10 (75%; 65%) or 1 to 9 (25%; 35%); a histology of adenocarcinoma (28%; 37%), adenosquamous (2%; 6%), or squamous cell (69%; 57%); and liver metastases (26%; 20%).

Patients in cohort A2 had a median age of 52.0 years (range, 30-83), and 58% had an ECOG performance status of 0. Prior lines of therapy included 1 (52%), 2 (23%), or 3 or more (26%). Prior neoadjuvant/adjuvant therapy and prior bevacizumab was reported in 58% and 48% of patients, respectively. Regarding histology, patients had either adenocarcinoma (74%), squamous cell (23%), or epidermoid (3%), and 13% of patients had liver metastases at baseline.

In cohort A1, 18% of patients receiving vibostolimab/pembrolizumab were ongoing treatment at data cutoff at a median follow-up of 18.2 months (range, 1.6-17.0). Eighty-two percent of patients discontinued treatment due to adverse effects (AEs; 13%), disease progression (65%), or patient/physician withdrawal (5%). Seventeen percent of patients receiving pembrolizumab alone were still on treatment, and reasons for discontinuation among the 83% who stopped treatment included AEs (4%), disease progression (75%), and patient/physician withdrawal (5%).

Ten percent of patients in cohort A2 were still on treatment at data cutoff at a median follow-up of 15.7 months (range, 12-22.2). Reasons for discontinuation included AEs (10%), disease progression (74%), patient/physician withdrawal (3%), and non-study anticancer therapy (3%).

Additional data from cohort A1 showed that vibostolimab/pembrolizumab elicited a median OS of 10.2 months (95% CI, 7.2-15.2) compared with 10.3 months (95% CI, 8.4-14.7) for pembrolizumab alone (HR, 1.00; 95% CI, 0.69-1.45). The 12-month PFS rates were 48.2% and 42.1%, respectively. In cohort A2, the median OS was 12.8 months (95% CI, 7.9-17.0) with a 12-month OS rate of 50.6%.

Findings from a biomarker subgroup analysis in cohort A1 demonstrated that a higher ORR was observed for both treatment regimens in patients with high PD-L1, TIGIT immunohistochemistry, tumor mutational burden, and inflamed T cells. Notably, trends toward an improved ORR were observed in the high subgroups for vibostolimab/pembrolizumab vs pembrolizumab alone.

Notably, a circulating tumor DNA (ctDNA) analysis revealed that ctDNA was reduced by 32% from baseline to cycle 3 in patients treated with vibostolimab/pembrolizumab compared with a 6% reduction in those given pembrolizumab alone.

Regarding safety data from cohort A1, the rates of any-grade treatment-related AEs (TRAEs) were 75% for vibostolimab/pembrolizumab vs 58% for pembrolizumab alone. The rates of grade 3/4 TRAEs were 18% and 12%, respectively. One grade 5 TRAE (1%) was reported in the combination arm, and none occurred in the monotherapy arm. Notably, 6% of patients discontinued treated due to TREAs in the vibostolimab/pembrolizumab arm vs none in the pembrolizumab arm.

The rates of any-grade and grade 3/4 immune-mediated AES were 35% and 8%, respectively, for vibostolimab/pembrolizumab. Those respective rates were 31% and 4% for pembrolizumab alone. The most common any-grade immune-mediated AEs included hypothyroidism (vibostolimab/pembrolizumab, 17%; pembrolizumab, 13%), hyperthyroidism (17%; 19%), hypophysitis (4%; 0%), colitis (1%; 5%), and pneumonitis (1%; 2%).

In cohort A2, 42.4% of patients experienced any-grade TRAEs, including 6.1% who had grade 3/4 TRAEs. No grade 5 TRAEs were reported, and 3.0% of patients discontinued treatment due to TRAEs.

The rates of any-grade and grade 3/4 immune-mediated AEs were 36% and 16%, respectively, with the most common including hypothyroidism (10%), hyperthyroidism (16%), hypophysitis (7%), colitis (3%), and pneumonitis (7%).

References

1. Leary A, Rojas C, Yonemori K, et al. KEYVIBE-005: efficacy and safety of coformulated vibostolimab/pembrolizumab in patients with previously treated metastatic cervical cancer. Presented at: 2024 ESMO Gynecological Cancers Congress; June 20-22, 2024; Florence, Italy. Abstract 21O.
2. Shapira-Frommer R, Perets R, Voskoboynik M, et al. Safety and efficacy of vibostolimab (vibo) plus pembrolizumab (pembro) in patients (pts) with cervical cancer naive to PD-1/PD-L1 inhibitors. Cancer Res2022;82(suppl 12):CT508. doi:10.1158/1538-7445.AM2022-CT50