Efficacy and safety of serplulimab plus nab-paclitaxel in previously treated patients with PD-L1-positive advanced cervical cancer: a phase II, single-arm study

Cervical cancer remains a major global health burden, with approximately 604,000 new cases and 342,000 deaths in 2020 worldwide. Prognosis for metastatic or recurrent disease is poor; first-line regimens yield a median progression-free survival (PFS) of less than 1 year, and there is no established standard of care beyond first line. PD-1/PD-L1 pathway inhibition is a promising strategy in cervical cancer, supported by HPV-associated upregulation of PD-L1 and immune evasion. Single-agent PD-1 inhibitors (pembrolizumab, nivolumab, balstilimab) have shown activity in PD-L1-positive or unselected cervical cancer but with modest objective response rates (generally under 30%). Cemiplimab has demonstrated a survival advantage over chemotherapy in the recurrent setting, further validating PD-1 blockade. Nab-paclitaxel has shown one of the higher response rates among chemotherapies for drug-resistant cervical cancer and possesses antiangiogenic properties. Combining chemotherapy with immune checkpoint inhibitors has improved outcomes in several solid tumors and may enhance response in cervical cancer. Serplulimab is a humanized anti-PD-1 monoclonal antibody with demonstrated antitumor activity and acceptable safety across cancers, including improved survival when combined with chemotherapy in extensive-stage small cell lung cancer. This study evaluates serplulimab plus nab-paclitaxel in patients with advanced PD-L1-positive cervical cancer who progressed after or were intolerant to first-line therapy.

Prior studies in previously treated cervical cancer show modest activity with PD-1 inhibitors: pembrolizumab in KEYNOTE-158 reported ORRs around 12–15% overall and ~14.6% in PD-L1-positive tumors; nivolumab showed an ORR of 26.3% in a small cohort; balstilimab achieved a 15% ORR with a median DOR of 15.4 months; cemiplimab improved survival vs chemotherapy with an ORR of 16.4%. Chemotherapy options yield limited benefit; nab-paclitaxel monotherapy produced an ORR of 28.6% with PFS 5.0 months and OS 9.4 months in resistant disease. In other malignancies, combining PD-1/PD-L1 inhibitors with chemotherapy has improved response and outcomes, suggesting potential synergy. Serplulimab has demonstrated efficacy and a manageable safety profile across indications, including survival benefit with chemotherapy in small cell lung cancer and activity in MSI-H tumors. These data support investigating serplulimab plus nab-paclitaxel in cervical cancer to potentially enhance tumor responses beyond single-agent PD-1 inhibition or chemotherapy alone.

Design: Single-arm, open-label, multicenter phase II study conducted across 11 sites in China (ClinicalTrials.gov NCT04150575). Population: Adults 18–75 years with histologically or cytologically confirmed PD-L1-positive (CPS ≥1) advanced cervical cancer, measurable disease per RECIST v1.1, ECOG PS 0–1, adequate organ function, and life expectancy ≥12 weeks. Patients had progressed after standard treatment or were intolerant to first-line chemotherapy. Key exclusions: prior nab-paclitaxel, prior immune checkpoint inhibitor therapy, active second malignancy, CNS/leptomeningeal metastasis, planned or prior organ/bone marrow transplant, significant cardiovascular disease, or autoimmune disease. Treatment: Serplulimab 4.5 mg/kg IV every 3 weeks for up to 2 years (35 cycles) plus nab-paclitaxel 260 mg/m2 IV every 3 weeks for up to six cycles. Treatment continued until progression, new antitumor therapy, death, unacceptable toxicity, consent withdrawal, or other protocol-specified reasons. Assessments: Tumor imaging every 6 weeks for 48 weeks, then every 12 weeks. Efficacy assessed by independent radiological review committee (IRRC) and investigators per RECIST v1.1. Safety assessed by AE monitoring (MedDRA v23.1; NCI CTCAE v5.0), including treatment-emergent AEs (TEAEs), serious AEs, adverse drug reactions (ADRs), infusion-related reactions, and immune-related AEs (irAEs). Follow-up: Safety at 30 and 90 days post last dose; survival every 12 weeks post-treatment. Endpoints: Primary—safety and ORR by IRRC (CR+PR). Secondary—ORR by investigator, PFS (IRRC and investigator), 6-month PFS rate, OS, 6-month OS rate, duration of response (DOR), and disease control rate (DCR: CR+PR+SD). Statistics: No formal hypothesis testing. Full analysis set (FAS) included all dosed patients and was primary for efficacy and safety; per-protocol set (PPS) supported FAS; safety set (SS) included all treated patients. ORR CIs via Exact (Clopper-Pearson); time-to-event endpoints via Kaplan-Meier. Descriptive statistics summarized patient flow and baseline characteristics.

Enrollment and follow-up: 52 screened; 21 enrolled between Dec 27, 2019 and Jun 30, 2020. Median follow-up 14.6 months (range 0.2–21.7) at data cutoff (Dec 31, 2021). Baseline: Mean age 50.8 years (31–64); ECOG 1 in 71.4%; squamous histology 95.2%; CPS ≥20 in 66.7%; ≥2 prior lines in 28.6%; prior radiotherapy 85.7%; lung metastasis 42.9%. Efficacy (IRRC, RECIST v1.1): ORR 57.1% (12/21; 95% CI 34.0–78.2) with CR 14.3% (3/21) and PR 42.9% (9/21); DCR 76.2% (16/21; 95% CI 52.8–91.8). Among 19 evaluable post-baseline, 89.5% had tumor shrinkage; 78.9% had ≥30% reduction. DOR: Median not reached (95% CI 4.1–NR); 12-month DOR rate 70.0% (95% CI 32.9–89.2). PFS (IRRC): Median 5.7 months (95% CI 3.0–NR); 6- and 12-month PFS rate 48.2% (95% CI 23.3–69.5). OS: Median 15.5 months (95% CI 10.5–NR). Investigator assessment: ORR 47.6% (95% CI 25.7–70.2); median PFS 5.6 months (95% CI 3.0–13.9). PD-L1 subgroup: IRRC-assessed ORR 64.3% in CPS ≥20 (Supplementary Table 2). Safety: All patients experienced TEAEs; grade ≥3 TEAEs in 81.0% (17/21). Common TEAEs (any grade): anemia 76.2% (grade ≥3: 23.8%), decreased WBC 71.4% (28.6%), decreased neutrophils 66.7% (33.3%), constipation 52.4%, alopecia 47.6%, asthenia 42.9%, decreased appetite 33.3%, elevated g-glutamyltransferase 28.6% (14.3%). Serious AEs in 38.1% (8/21). TEAE-related deaths in 4 patients; one death (multiple organ dysfunction syndrome) possibly treatment-related. ADRs in 90.5% (19/21); grade ≥3 ADRs in 33.3% (7/21). No infusion-related reactions. irAEs in 57.1% (12/21); grade ≥3 irAEs in 2 patients (elevated creatine phosphokinase in one; diarrhea and anemia in another); most common irAEs: hypothyroidism 28.6%, hyperthyroidism 14.3%, pruritus 14.3%.

Combining serplulimab with nab-paclitaxel demonstrated a high ORR (57.1% by IRRC), durable responses (median DOR not reached), and encouraging survival (median OS 15.5 months) in previously treated PD-L1-positive advanced cervical cancer. These outcomes compare favorably, with caution due to cross-trial differences, to historic results from single-agent PD-1 inhibitors and chemotherapy, suggesting potential synergy between chemotherapy and PD-1 blockade. Safety was manageable and consistent with known profiles of PD-1 inhibitors and taxanes; cytopenias were expected with nab-paclitaxel, and irAEs were mostly grade 1–2 with no infusion reactions or serplulimab discontinuations due to TEAEs. Exploratory findings suggest higher PD-L1 expression (CPS ≥20) may be associated with higher response rates, though the small sample precludes definitive conclusions. Overall, the regimen offers meaningful clinical activity and manageable toxicity in the second-line and later setting where effective options are limited.

Serplulimab plus nab-paclitaxel provided durable antitumor activity with a manageable safety profile in previously treated patients with PD-L1-positive advanced cervical cancer. The high ORR, prolonged DOR, and favorable survival metrics support this combination as a clinically meaningful option in the second-line setting. Further larger, randomized studies are warranted to confirm efficacy, better define safety, and evaluate predictive biomarkers such as PD-L1 CPS. A phase II study of serplulimab plus bevacizumab with chemotherapy in the first-line setting for recurrent or metastatic cervical cancer is planned (NCT05444374).

Single-arm, open-label design without a control arm and a small sample size (N=21) limit the robustness and generalizability of the findings and preclude definitive comparative efficacy conclusions. The study was not powered for hypothesis testing. Subgroup analyses, including the relationship between PD-L1 CPS and efficacy (ORR, PFS), are exploratory and inconclusive. Potential selection bias and the multicenter nature with varying prior treatments may influence outcomes.