Cadonilimab Plus Chemotherapy Provides Frontline Survival Benefit in Metastatic Cervical Cancer

The combination of cadonilimab (AK104) and platinum-based chemotherapy with or without bevacizumab (Avastin) improved survival outcomes and was associated with a manageable safety profile compared with placebo plus chemotherapy in the frontline setting for patients with persistent, recurrent, or metastatic cervical cancer, according to interim findings from the phase 3 COMPASSION-16 trial (NCT04982237; AKT104-303), which were published in The Lancet.¹

At a data cutoff date of September 4, 2023, and a median follow-up of 17.9 months (interquartile range [IQR], 15.9-20.2), the median progression-free survival (PFS) per blinded independent central review (BICR) was 12.7 months (95% CI, 11.6-16.1) among patients who received cadonilimab (n = 222) vs 8.1 months (95% CI, 7.7-9.6) among those who received placebo (n = 223; HR, 0.62; 95% CI, 0.49-0.80; P < .0001). Investigator-assessed PFS outcomes were consistent with the BICR-assessed results.

At a data cutoff of April 30, 2024, and a median follow-up of 25.6 months (IQR, 23.6-28.0), the median overall survival (OS) was not reached (NR; 95% CI, 27.0 months-not evaluable [NE]) vs 22.8 months (95% CI, 17.6-29.0) in the cadonilimab and placebo arms, respectively (HR, 0.64; 95% CI, 0.48-0.86; P = .0011). The 24-month OS rates were 62.2% (95% CI, 55.2%-68.5%) in the cadonilimab arm and 48.4% (95% CI, 41.1%-55.4%) in the placebo arm.

“COMPASSION-16 (AKT104-303) is the first phase 3 study to show that adding cadonilimab to chemotherapy with or without bevacizumab significantly and clinically meaningfully improved PFS and OS in first-line advanced cervical cancer,” lead study author Xiaohua Wu, MD, and coauthors, wrote in the paper.

Wu is a professor and chair of the Department of Gynecologic Oncology at Fudan University Shanghai Cancer Center in China.

Trial Background and Rationale 

Cadonilimab is a PD-1–and CTLA-4–directed bispecific antibody that has generated clinical benefits in patients with advanced cervical cancer. In the phase 2 COMPASSION-13 trial, patients with recurrent or metastatic cervical cancer who received cadonilimab plus chemotherapy with or without bevacizumab in the frontline setting (n = 30) achieved an objective response rate (ORR) of 79.3% (95% CI, 60.3%-92.0%); this rate was 70.6% in patients with a PD-L1 combined positive score (CPS) lower than 1 (n = 17).²

COMPASSION-16 Trial Design and Patient Characteristics

Between September 11, 2021, and June 23, 2022, the randomized, double-blind, multicenter, placebo-controlled COMPASSION-16 trial enrolled 445 women ages 18 to 75 years across 59 clinical sites in China.¹ Patients need to have previously untreated persistent, recurrent, or metastatic (stage IVB) cervical cancer of squamous cell carcinoma, adenocarcinoma, or adenosquamous carcinoma that was not amenable to curative surgery or concurrent chemoradiotherapy. Patients need to have received no prior systemic therapy for persistent, recurrent, or metastatic disease; at least 1 measurable lesion per RECIST 1.1 criteria; an ECOG performance status of 0 or 1; and adequate organ function.

Patients were excluded if they had received prior concurrent chemoradiotherapy for curative intent or as neoadjuvant or adjuvant treatment within 90 days before to randomization; had received previous treatment directed against the mechanism of tumor immunity; or had nervous system metastases or carcinomatous meningitis. Patients were not permitted to receive bevacizumab if they had serious vascular diseases, tumor invasion of the rectum or bladder, gastrointestinal perforation within 6 months of randomization, active bleeding, or high-risk factors for bleeding.

Patients were randomly assigned 1:1 to receive cadonilimab at 10 mg/kg or placebo in combination with platinum-based chemotherapy (cisplatin at 50 mg/m2 or carboplatin at an area under the curve of 4-5 plus paclitaxel at 175 mg/m2) with or without bevacizumab at 15 mg/kg every 3 weeks for 6 cycles, followed by maintenance therapy with cadonilimab or placebo with or without bevacizumab every 3 weeks. Patients were stratified by bevacizumab use (yes vs no) and prior exposure to concurrent chemoradiotherapy (yes vs no). Prior to randomization, investigators decided whether to use cisplatin or carboplatin and bevacizumab. Treatment crossover and changes to the chemotherapy regimen were not permitted.

Patients continued treatment until disease progression, unacceptable toxicity, no further clinical benefit, withdrawal of consent, or receipt of a maximum of 2 years of cadonilimab or placebo. After 2 years of treatment, medical discussions could be conducted with the sponsor to determine whether a patient could benefit from continuing their assigned regimen. 

The trial’s dual primary end points were PFS per BICR and OS in the full analysis set. Secondary end points included investigator-assessed PFS, ORR, duration of response (DOR), disease control rate (DCR), time to response (TTR) per RECIST 1.1 criteria, the association between PD-L1 expression and efficacy, and safety.

Baseline characteristics were balanced between the 2 arms. In the overall population, the median age was 56 years (IQR, 50-62), 64% of patients had an ECOG performance score of 1, 83% of patients had squamous cell carcinoma, 48% of patients had received prior concurrent chemoradiotherapy, and 73% of patients had metastatic disease at baseline. All patients were ethnically Chinese. Sixty percent of patients received bevacizumab. In total, 26% of patients had a PD-L1 CPS lower than 1, 70% of patients had a PD-L1 CPS of at least 1, and 40% of patients had a PD-L1 CPS of at least 10.

At the data cutoff date for the OS interim analysis in the full analysis set, 21% of patients in the cadonilimab arm and 14% of those in the placebo arm were continuing to receive at least 1 study drug. Four patients who were randomly assigned to the placebo arm accidentally received 1 dose of cadonilimab and were included in the cadonilimab arm for the safety analysis.

Additional BICR-Assessed Efficacy Findings

At the data cutoff date for the PFS interim analysis, the BICR-assessed PFS outcomes favored the cadonilimab arm vs the placebo arm across all prespecified subgroups. Among patients with a PD-L1 CPS of at least 1, the median PFS was 13.5 months (95% CI, 10.8-17.3) vs 8.3 months (95% CI, 7.5-9.7 [HR, 0.62; 95% CI, 0.46-0.83]). Among patients with a PD-L1 CPS lower than 1, the respective median PFS values in these 2 arms were 12.0 months (95% CI, 9.5-14.5) vs 8.2 months (95% CI, 7.8-11.8 [HR, 0.73; 95% CI, 0.46-1.17]).

Among patients who received bevacizumab, the median PFS was 15.1 months (95% CI, 11.8-19.7) in the cadonilimab arm vs 11.5 months (95% CI, 9.5-14.8) in the placebo arm (HR, 0.81; 95% CI, 0.58-1.13). Among those who did not receive bevacizumab, the median PFS values in these respective arms were 11.7 months (95% CI, 9.5-14.5) and 6.9 months (95% CI, 5.6-7.8 [HR, 0.46; 95% CI, 0.32-0.66]).

“We speculate that the efficacy shown in the subgroup without bevacizumab use could be attributable to the combined anti–PD-1 and anti–CTLA-4 activity of cadonilimab,” the authors noted. “Cadonilimab might offer a suitable treatment option for individuals who are not eligible for bevacizumab therapy.”

Among patients who had received concurrent chemoradiotherapy, the median PFS was 15.4 months (95% CI, 11.6-20.0) in the cadonilimab arm vs 7.9 months (95% CI, 7.5-11.5) in the placebo arm (HR, 0.55; 95% CI, 0.39-0.79). Among those who did not receive concurrent chemoradiotherapy, the median PFS values in these respective arms were 11.8 months (95% CI, 9.7-14.4) and 8.5 months (95% CI, 7.7-10.2 [HR, 0.72; 95% CI, 0.52-1.01]).

At the data cutoff for the prespecified OS interim analysis, the OS outcomes favored the cadonilimab arm vs the placebo arm across all prespecified subgroups. Among patients with a PD-L1 CPS of at least 1, the median OS was NR (95% CI, 26.6-NE) vs 22.7 months (95% CI, 15.5-NE [HR, 0.69; 95% CI, 0.49-0.97]). Among patients with a PD-L1 CPS lower than 1, the respective median OS values in these 2 arms were NR (95% CI, 23.0-NE) vs 25.3 months (95% CI, 17.1-NE [HR, 0.77; 95% CI, 0.44-1.34]).

Among patients who received bevacizumab, the median OS was NR (95% CI, 27.8-NE) in the cadonilimab arm vs NR (95% CI, 25.7-NE) in the placebo arm (HR, 0.84; 95% CI, 0.56-1.26). Among those who did not receive bevacizumab, the median OS values in these respective arms were 28.2 months (95% CI, 17.5-NE) and 15.1 months (95% CI, 12.2-17.6 [HR, 0.50; 95% CI, 0.33-0.75]). Among patients who had received prior concurrent chemoradiotherapy, the median OS was NR (95% CI, 28.2-NE) in the cadonilimab arm vs 22.8 months (95% CI, 16.9-NE) in the placebo arm (HR, 0.54; 95% CI, 0.35-0.82). Among those who did not receive prior concurrent chemoradiotherapy, the median OS values in these respective arms were 27.0 months (95% CI, 23.2-NE) and 24.5 months (95% CI, 15.7-NE [HR, 0.76; 95% CI, 0.52-1.12]).

The study authors noted that the updated median PFS results per BICR at the data cutoff date for the prespecified interim OS analysis were consistent with those seen in the previous interim analysis, including in each PD-L1 expression subgroup. At the data cutoff date for the prespecified interim OS analysis, 83% of patients in the cadonilimab arm had a CR or PR per BICR vs 69% of those in the placebo arm. Among these 2 arms, 36% and 23% of patients had a CR.

At the data cutoff date for the prespecified interim OS analysis, among 208 BICR-evaluable patients in the cadonilimab arm, the DCR was 94% vs 92% among 205 BICR-evaluable patients in the placebo arm. In the cadonilimab arm, the median DOR was 13.2 months (IQR, 6.2-NE), and the median TTR was 1.5 months (IQR, 1.4-2.7). In the placebo arm, the median DOR and median TTR were 8.2 months (IQR, 4.7-25.7) and 1.5 months (IQR, 1.4-2.8), respectively.

Investigator-Assessed Efficacy Findings

The investigator-assessed ORR was 79% in the cadonilimab arm vs 68% in the placebo arm. The respective investigator-assessed DCRs were 92% and 93%. The respective median DORs were 15.7 months (IQR, 6.2-27.4) and 8.3 months (IQR, 4.4-21.2). The median TTRs in these respective arms were 1.5 months (IQR, 1.4-2.8) and 1.5 months (IQR, 1.4-2.7). The median duration of treatment was 10.4 months (IQR, 4.2-22.5) in the cadonilimab arm vs 8.5 months (IQR, 4.6-15.0) in the placebo arm.

Understanding the Safety

Treatment-emergent adverse effects (TEAEs) were observed in over 99% of treated patients in the cadonilimab arm (n = 226) and 100% of those in the placebo arm (n = 219). Grade 3 or higher TEAEs occurred in 85% and 80% of patients in these respective arms. In the cadonilimab arm, the most common AEs were anemia (any-grade, 68%; grade ≥ 3, 17%), decreased white blood cell (WBC) counts (any-grade, 67%; grade ≥ 3, 28%), and decreased neutrophil counts (any-grade, 61%; grade ≥ 3, 41%). In the placebo arm, the most common AEs were anemia (any-grade, 74%; grade ≥ 3, 26%), decreased WBC counts (any-grade, 74%; grade ≥ 3, 36%), and decreased neutrophil counts (any-grade, 66%; grade ≥ 3, 46%).

AEs led to discontinuation of any study drugs in 28% of patients in the cadonilimab arm vs 11% of those in the placebo arm. AEs led to death in 12 patients in the cadonilimab arm; 9 of these AEs were treatment related. AEs led to death in 7 patients in the cadonilimab arm; 6 of these AEs were treatment related.

Treatment-related AEs (TRAEs) were reported in over 99% of patients in the cadonilimab arm and 100% of those in the placebo arm. Grade 3 or higher TRAEs occurred in 82% and 79% of patients in these respective arms. Immune-related AEs were reported in 46% of patients in the cadonilimab arm vs 7% of those in the placebo arm; 10% and less than 1% of patients in these respective arms had grade 3 or higher TRAEs. No grade 5 immune-related AEs were reported in either arm.

The authors noted that one limitation of this trial was the homogeneity of the patient demographics, since the trial was conducted in China and only included patients who were ethnically Chinese.

“COMPASSION-16 is the largest first-line cervical cancer trial conducted in an Asian population to date, and the findings provided valuable evidence in the treatment of advanced cervical cancer in this population,” the authors concluded.

References

1. Wu X, Sun Y, Yang H, et al. Cadonilimab plus platinum-based chemotherapy with or without bevacizumab as first-line treatment for persistent, recurrent, or metastatic cervical cancer (COMPASSION-16): a randomised, double-blind, placebo-controlled phase 3 trial in China. Lancet. 2024;404(10463):1668-1676. doi:10.1016/S0140-6736(24)02135-4
2. Lou H, Cai H, Huang X, et al. Cadonilimab combined with chemotherapy with or without bevacizumab as first-line treatment in recurrent or metastatic cervical cancer (COMPASSION-13): a phase 2 study. Clin Cancer Res. 2024;30(8):1501-1508. doi:10.1158/1078-0432.CCR-23-3162