Comparative effectiveness of immune checkpoint inhibitors in squamous vs. non-squamous cervical carcinoma: a systematic review and meta-analysis of randomized controlled trials Authors: Li-Fang Lan, Yu-Jie Chen, Dun-Chang Mo, Wen-Yi Li, Fei Yu Published: 2026 ABSTRACT AbstractBackgroundThe differential therapeutic efficacy of immune checkpoint inhibitor (ICI) therapy in patients with cervical squamous cell carcinoma (SCC) and non-squamous cell carcinoma (non-SCC) remains unclear. This study aimed to evaluate and compare the efficacy of ICI therapy in patients with metastatic cervical SCC and non-SCC.MethodsA systematic literature search was conducted in PubMed, Embase, and the Cochrane Library to identify randomized controlled trials (RCTs) investigating the efficacy of ICI therapy for metastatic cervical cancer. The primary efficacy endpoints were overall survival (OS), progression-free survival (PFS), and objective response rate (ORR).ResultsFour eligible RCTs were included in this meta-analysis. ICI therapy significantly improved PFS and OS in both SCC and non-SCC patients compared with the control treatment. Anti-programmed cell death protein 1 (anti-PD-1) monotherapy significantly increased the ORR in SCC patients (P<0.001), while no significant improvement was observed in non-SCC patients (P=0.23). Subgroup analysis showed that anti-PD-1 and anti-PD-1/CTLA-4 combination therapies significantly prolonged PFS in SCC patients (both P<0.001), but not in non-SCC patients (P=0.13 and P=0.83, respectively). In contrast, anti-programmed cell death ligand 1 (anti-PD-L1) therapy failed to improve PFS in SCC patients (P=0.22) but significantly enhanced PFS in non-SCC patients (P<0.001). OS subgroup analysis revealed that anti-PD-1, anti-PD-L1, and anti-PD-1/CTLA-4 therapies all significantly prolonged OS in SCC patients. In non-SCC patients, only the anti-PD-1 subgroup exhibited a significant OS benefit (P=0.006), with no statistically significant differences observed in the anti-PD-L1 and anti-PD-1/CTLA-4 subgroups (P=0.08 and P=0.83, respectively).ConclusionsICI therapy improves clinical outcomes in patients with metastatic cervical cancer, but its efficacy varies significantly between SCC and non-SCC subtypes. These findings warrant high attention in clinical practice and provide a reference for individualized therapeutic decision-making.Keywords: Cervical squamous cell carcinoma, cervical non-squamous cell carcinoma, immune checkpoint inhibitors (ICIs), meta-analysis Highlight box Key findings • This meta-analysis of four eligible randomized controlled trials (RCTs) confirmed distinct survival and response benefits of immune checkpoint inhibitors (ICIs) in metastatic cervical squamous cell carcinoma (SCC) versus non-SCC. Anti-PD-1 monotherapy improved progression-free survival (PFS) and overall survival in both subgroups; anti-PD-1/CTLA-4 combination only benefited SCC patients, while anti-PD-L1 therapy only improved PFS in non-SCC patients. What is known and what is new? • Previous clinical trials have proven that ICIs exert anti-tumor activity in patients with metastatic cervical carcinoma. • Prior researches lacked subgroup analysis stratified by tumor histology; this study first quantified differentiated ICI efficacy between SCC and non-SCC cervical cancer. What is the implication, and what should change now? • Clinicians should select individualized ICI regimens based on cervical cancer histotype to maximize treatment benefits; stratified prospective trials are required to further verify the subgroup differences we observed. Introduction Cervical cancer is a common malignant tumor threatening women’s health and ranks as the second most prevalent gynecological malignancy worldwide in terms of incidence and mortality (1). Over the past decade, the global implementation of human papillomavirus (HPV) vaccination programs and cervical cancer screening strategies has led to a gradual decline in the incidence and mortality of cervical cancer (2). However, disease recurrence and metastasis remain major challenges in clinical management, and patients with recurrent or metastatic cervical cancer have a poor prognosis, which poses a persistent therapeutic dilemma in gynecological oncology (3,4). In recent years, the rapid development of immunotherapy, especially immune checkpoint inhibitors (ICIs), has shown great therapeutic potential in the treatment of cervical cancer and significantly improved the long-term survival of patients (5). ICIs can modulate the immune system to enhance the efficacy of conventional radiotherapy and chemotherapy (6-8). Clinical evidence has confirmed that the combination of ICIs with radiotherapy and chemotherapy achieves better therapeutic outcomes in patients with locally advanced cervical cancer compared with radiotherapy and chemotherapy alone (8). For recurrent or metastatic cervical cancer, multiple studies have demonstrated that first-line therapy with ICIs combined with chemotherapy (with or without bevacizumab) significantly improves long-term survival (5,9), and this combination regimen has thus become the recommended standard treatment for recurrent or metastatic cervical cancer (5,9,10). Despite the widespread clinical recommendation of ICIs for recurrent or metastatic cervical cancer, there is a paucity of high-quality evidence regarding the differential efficacy of ICIs in different histological subtypes of cervical cancer, particularly between SCC and non-SCC. To address this knowledge gap, we conducted a comprehensive meta-analysis of randomized controlled trials (RCTs) to systematically evaluate and compare the efficacy of ICI therapy in patients with recurrent or metastatic cervical SCC and non-SCC, aiming to provide more refined clinical evidence for individualized treatment decision-making in cervical cancer. We present this article in accordance with the PRISMA reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0345/rc). Methods Literature screening A comprehensive and systematic electronic literature search was performed in PubMed, EMBASE, and the Cochrane Library from the inception of each database to March 2025. The search strategy was constructed using the following Boolean terms: (“immunotherapy” OR “immune checkpoint inhibitors” OR “PD-1” OR “PD-L1” OR “pembrolizumab” OR “nivolumab” OR “atezolizumab” OR “avelumab” OR “tremelimumab” OR “durvalumab” OR “spartalizumab” OR “toripalimab” OR “dostarlimab” OR “camrelizumab” OR “sintilimab” OR “cemiplimab”) AND (“cervical squamous cell carcinoma” OR “cervical adenocarcinoma” OR “cervical adenosquamous carcinoma”). In addition, a manual search of the reference lists of all included studies and relevant review articles was conducted to identify potentially eligible studies that might have been missed by the electronic search. Inclusion and exclusion criteria The inclusion criteria were restricted to RCTs that assessed the efficacy of ICI therapy in patients with recurrent or metastatic cervical cancer. Non-RCTs, review articles, meta-analyses, case reports, and studies published in languages other than English were systematically excluded. Additionally, studies with ambiguous or insufficient data were not considered for inclusion. In instances of duplicate publications, only the study presenting the most comprehensive and up-to-date dataset was retained for analysis. Data extraction and quality assessment Two independent reviewers (L.F.L. and D.C.M.) conducted the initial literature screening, title and abstract evaluation, and full-text assessment to determine study eligibility. Any discrepancies between the two reviewers were resolved through discussion and consensus with a third senior reviewer (Y.J.C.). Two other independent reviewers (F.Y. and W.Y.L.) extracted the following data from the included studies: basic study characteristics (study name, first author, publication year, study phase), patient characteristics (sample size, age range), treatment regimens (ICI type, combination therapy, control regimen), and clinical outcome indicators [overall survival (OS), progression-free survival (PFS), objective response rate (ORR)]. Hazard ratios (HRs) with 95% confidence intervals (CIs) were extracted for OS and PFS, and odds ratios (ORs) with 95% CIs were extracted for ORR. The methodological quality and risk of bias of the included RCTs were independently assessed by F.Y. and W.Y.L. using the Cochrane Risk of Bias Tool 2.0 (11), which evaluates seven key domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other potential biases. Each domain was rated as low risk, unclear risk, or high risk of bias according to the standard assessment criteria. Publication bias was evaluated using funnel plots, and any disagreements in the risk of bias assessment were resolved by Y.J.C. Statistical analysis All statistical analyses were performed using Review Manager 5.4 software (The Cochrane Collaboration, Copenhagen, Denmark). OS and PFS were synthesized using pooled HRs with 95% CIs, and ORR was synthesized using pooled ORs with 95% CIs. Heterogeneity among the included studies was quantitatively assessed using the I2 statistic, where I2<50% indicated low heterogeneity, 50%≤I2<75% indicated moderate heterogeneity, and I2≥75% indicated high heterogeneity. Given the limited number of included studies, a random-effects model was applied for all meta-analyses to reduce potential bias and improve the robustness of the results. A two-sided P value <0.05 was considered statistically significant for all analyses. Results Literature selection A total of 718 relevant publications were initially identified through the electronic database search. After removing duplicate publications, 327 studies were screened for titles and abstracts, and 45 studies were further assessed for full-text eligibility. Finally, four RCTs [BEATcc (9), EMPOWERCervical 1 (10), KEYNOTE-826 (5), and COMPASSION-16 (12)] involving 2,080 patients published between 2021 and 2024 were included in this meta-analysis. The detailed literature screening process and results are presented in the PRISMA flow diagram (Figure 1). Flowchart illustrating the literature review process. RCT, randomized controlled trial. Characteristics of included studies The four included RCTs were multicenter studies conducted in different geographical regions: BEATcc was conducted in 92 clinical research centers across Europe and Japan; EMPOWERCervical 1 included 111 centers in 14 countries worldwide; KEYNOTE-826 was carried out in 151 centers in 19 countries; and COMPASSION-16 was a multicenter study conducted in 59 clinical trial institutions in China. All enrolled patients were pathologically confirmed to have recurrent or metastatic cervical cancer and were randomly assigned to the experimental group (ICI monotherapy or ICI combined with chemotherapy with/without bevacizumab) or the control group (conventional chemotherapy with/without bevacizumab). The ICIs evaluated in these studies included two anti-PD-1 antibodies (pembrolizumab, cemiplimab), one anti-PD-L1 antibody (atezolizumab), and one anti-PD-1/CTLA-4 bispecific antibody (cadonilimab). The detailed basic characteristics and treatment regimens of the included studies are summarized in Table 1. Table 1Baseline characteristics of the studies included in the meta-analysisStudy [year]Study designPhase Number of patients Age (years), median [IQR]Number of patients with squamous cell carcinoma, n (%)InterventionMedian follow-up of the study group (months)EMPOWERCervical [2022] (10)RCT3608Exp: 51 [22–81]; Con: 50 [24–87]Exp: 240 (78.9); Con: 233 (76.6)Exp: cemiplimab; Con: chemotherapy16.8KEYNOTE-826 (2021) (5)RCT3617Exp: 51 [25–82]; Con: 50 [22–79]Exp: 235 (76.8); Con: 211 (68.3)Exp: pembrolizumab; Con: chemotherapy22BEATcc [2023] (9)RCT3410Exp: 51.0 [43.0–60.0]; Con: 52.5 [43.5–61.0]Exp: 164 (79.6); Con: 157 (77.0)Exp: atezolizumab plus bevacizumab plus chemotherapy; Con: bevacizumab plus chemotherapy32.9COMPASSION-16 [2024] (12)RCT3445Exp: 56 [23–75]; Con: 56 [23–75]Exp: 182 (82.0); Con: 188 (84.3)Exp: cadonilimab plus bevacizumab plus chemotherapy; Con: bevacizumab plus chemotherapy25.6Con, control; Exp, experiment; IQR, interquartile range; RCT, randomized controlled trial. Baseline characteristics of the studies included in the meta-analysis Con, control; Exp, experiment; IQR, interquartile range; RCT, randomized controlled trial. Therapeutic effect OS All four included studies reported OS outcomes stratified by SCC and non-SCC subtypes. Meta-analysis results (Figure 2) showed that ICI therapy significantly improved OS in both SCC patients (HR =0.68, 95% CI: 0.59–0.78, P<0.001; Figure 2A) and non-SCC patients (HR =0.68, 95% CI: 0.53–0.87, P=0.002; Figure 2B) compared with standard therapy. Subgroup analysis by ICI type revealed that anti-PD-1 therapy significantly prolonged OS in both SCC (HR =0.68, 95% CI: 0.57–0.80, P<0.001; Figure 2A) and non-SCC patients (HR =0.64, 95% CI: 0.46–0.88, P=0.006; Figure 2B). Anti-PD-L1 therapy significantly improved OS in SCC patients (HR =0.72, 95% CI: 0.54–0.97, P=0.03; Figure 2A), but no significant OS benefit was observed in non-SCC patients (HR =0.62, 95% CI: 0.36–1.06, P=0.08; Figure 2B). For the anti-PD-1/CTLA-4 dual blockade subgroup, combination therapy significantly prolonged OS in SCC patients (HR =0.64, 95% CI: 0.47–0.88, P=0.006; Figure 2A), while no statistically significant difference was found in non-SCC patients (HR =0.94, 95% CI: 0.52–1.69, P=0.83; Figure 2B). Forest plot of overall survival. (A) Forest plot of overall survival in SCC carcinoma. (B) Forest plot of overall survival in non-SCC carcinoma. CI, confidence interval; IV, inverse variance; PD-1, programmed cell death protein 1; PD-L1, programmed cell death ligand 1; SCC, squamous cell carcinoma; SE, standard error. PFS PFS data stratified by histological subtypes were available in all four included studies, and the meta-analysis results are presented in Figure 3. ICI therapy significantly improved PFS in both SCC patients (HR =0.66, 95% CI: 0.58–0.75, P<0.001; Figure 3A) and non-SCC patients (HR =0.72, 95% CI: 0.56–0.91, P=0.007; Figure 3B) compared with standard therapy. Subgroup analysis showed that anti-PD-1 therapy significantly enhanced PFS in SCC patients (HR =0.67, 95% CI: 0.58–0.78, P<0.001; Figure 3A), but the improvement was not statistically significant in non-SCC patients (HR =0.78, 95% CI: 0.57–1.08, P=0.13; Figure 3B). In contrast, anti-PD-L1 therapy failed to show a significant PFS benefit in SCC patients (HR =0.75, 95% CI: 0.47–1.19, P=0.22; Figure 3A) but significantly prolonged PFS in non-SCC patients (HR =0.58, 95% CI: 0.45–0.76, P<0.001; Figure 3B). Additionally, anti-PD-1/CTLA-4 combination therapy significantly improved PFS in SCC patients (HR =0.58, 95% CI: 0.44–0.76, P<0.001; Figure 3A), while no significant effect was observed in non-SCC patients (HR =0.94, 95% CI: 0.52–1.69, P=0.83; Figure 3B). Forest plot of progression-free survival. (A) Forest plot of progression-free survival in SCC carcinoma. (B) Forest plot of progression-free survival in non-SCC carcinoma. CI, confidence interval; IV, inverse variance; PD-1, programmed cell death protein 1; PD-L1, programmed cell death ligand 1; SCC, squamous cell carcinoma; SE, standard error. ORR ORR outcomes stratified by SCC and non-SCC subtypes were reported in only two of the four included studies [KEYNOTE-826 (5) and EMPOWERCervical 1 (10)], both of which evaluated anti-PD-1 monotherapy. Meta-analysis results (Figure 4) showed that anti-PD-1 monotherapy significantly increased the ORR in SCC patients compared with standard treatment (OR =2.96, 95% CI: 1.70–5.15, P<0.001; Figure 4A). In contrast, no statistically significant difference in ORR was observed between the anti-PD-1 and control groups in non-SCC patients (OR =1.54, 95% CI: 0.76–3.10, P=0.23; Figure 4B). Forest plot of objective response rate. (A) Forest plot of objective response rate in SCC carcinoma. (B) Forest plot of objective response rate in non-SCC carcinoma. CI, confidence interval; M‑H, Mantel–Haenszel; SCC, squamous cell carcinoma. Assessments of RCTs qualities The results of the risk of bias assessment for the four included RCTs are presented in Figure 5. All studies were rated as low risk of bias in all seven assessment domains of the Cochrane Risk of Bias Tool 2.0, indicating that all included studies were high-quality RCTs with reliable study results. The assessment of bias of included studies. +, low risk; ?, unclear risk. Discussion To the best of our knowledge, this is the first comprehensive meta-analysis of RCTs to systematically evaluate and compare the differential therapeutic efficacy of ICI therapy between cervical SCC and non-SCC subtypes in patients with recurrent or metastatic disease. Our results demonstrate a statistically significant heterogeneity in the efficacy of ICI therapy between metastatic cervical SCC and non-SCC patients: SCC patients exhibit a more favorable response to most ICI regimens, while non-SCC patients only derive a significant OS benefit from anti-PD-1 monotherapy. These findings are consistent with the results of recent clinical studies (8,13,14) and highlight the importance of selecting individualized ICI treatment strategies based on the histological subtype of cervical cancer in clinical practice. The distinct therapeutic responses of SCC and non-SCC to ICIs may be attributed to their inherent biological and molecular characteristics. First, cervical SCC has a significantly higher PD-L1 expression rate than adenocarcinoma (80% vs. 15–20%), and metastatic SCC shows an even higher PD-L1 positivity rate (15), which may render SCC more sensitive to ICI therapy targeting the PD-1/PD-L1 pathway. Second, SCC typically has a higher tumor mutation burden (TMB) compared with non-SCC; for example, head and neck SCC has a 3.5-fold higher TMB than adenocarcinoma (16), which can lead to the presentation of more tumor-specific antigens and enhanced activation of CD8+ T cells, thereby improving the anti-tumor immune response induced by ICIs. Furthermore, HPV-associated cervical SCC often exhibits an “immunologically hot” tumor microenvironment characterized by abundant CD8+ T cell infiltration (17), while cervical adenocarcinoma may develop immune evasion mechanisms by suppressing cytokines such as TGF-β (17), which further contributes to the differential sensitivity to ICIs. Clinical evidence has further confirmed the differential efficacy of ICIs in cervical SCC and non-SCC. The KEYNOTE-826 trial showed that in patients with a combined positive score (CPS) ≥1, pembrolizumab combined with chemotherapy achieved a more significant OS benefit in SCC (HR =0.60) than in non-SCC (HR =0.70), and this difference was consistent in both cisplatin-based and carboplatin-based chemotherapy regimens (5). Similarly, the CheckMate 358 study reported a higher ORR in SCC patients (45.8%) than in non-SCC patients (31.6%) receiving ICI therapy (14). The relatively low efficacy of ICIs in non-SCC may be related to its unique molecular profile: for example, HER2 overexpression is more common in cervical adenocarcinoma (approximately 15–20%) (18,19), and the HER2 signaling pathway can inhibit T cell function through the PI3K/AKT/mTOR axis (20,21), thereby reducing the responsiveness of non-SCC to immunotherapy. However, promising progress has been made in the treatment of non-SCC in recent years: a phase II study showed that the combination of cadonilimab (PD-1/CTLA-4 bispecific antibody) and disitamab vedotin (RC48) [HER2 antibody-drug conjugate (ADC)] achieved an ORR of 92.3% in HER2-positive recurrent/metastatic cervical non-SCC (22), highlighting the synergistic effect of targeted therapy and immunotherapy in overcoming immune resistance in non-SCC. In addition, the combination of oncolytic virus CAN-2409 with PD-1 inhibitors achieved a median OS of 25.4 months in non-squamous non-small cell lung cancer (23), which provides a new therapeutic direction for the treatment of cervical non-SCC by inducing tumor cell lysis, releasing tumor-associated antigens, and activating the systemic anti-tumor immune response. The predictive value of immune biomarkers for ICI therapy also varies significantly between cervical SCC and non-SCC subtypes. In non-SCC patients, the response rate to immunotherapy reaches 81.8% in those with PD-L1 CPS ≥10, while the predictive value of PD-L1 for ICI efficacy in SCC is relatively weak (24), suggesting that the efficacy of ICIs in SCC may need to be comprehensively evaluated by combining multiple biomarkers such as TMB and HPV status (25,26). Furthermore, the incidence of microsatellite instability-high/mismatch repair deficiency (MSI-H/dMMR) in cervical cancer is relatively low (3–11.3%), and its incidence in non-SCC subtypes is even lower (27,28), which limits the clinical application of MSI-H/dMMR as a universal biomarker for predicting ICI efficacy in cervical cancer (29,30). This study has several strengths: first, it is the first meta-analysis to focus on the differential efficacy of ICIs in cervical SCC and non-SCC subtypes, filling the current knowledge gap in this field; second, only high-quality multicenter RCTs were included, which ensures the reliability and clinical applicability of the study results; third, a detailed subgroup analysis by ICI type was performed, which provides a more refined reference for clinical treatment selection. However, several limitations of this study should also be acknowledged: first, the number of included studies is relatively small (only four studies), and only four ICIs (two anti-PD-1, one anti-PD-L1, one PD-1/CTLA-4 bispecific antibody) were evaluated, which may limit the generalizability of the pooled results and introduce potential selection bias; second, subgroup analyses for anti-PD-L1 and anti-PD-1/CTLA-4 therapies were each based on only one trial. Such single-trial representation limits the statistical power and stability of these subgroup estimates, and the corresponding findings may be easily affected by the specific design, population, and treatment regimen of the individual trial. Therefore, these results should be interpreted with extra caution, and no definitive clinical recommendations can be drawn from these limited subgroup data; third, standardized and paired PD-L1 expression data stratified simultaneously by histological subtype and treatment arm were not available across all included trials. Although we discussed that cervical SCC generally presents higher PD-L1 positivity than non-SCC, the absence of integrated stratification analysis prevented us from clarifying whether the superior efficacy of ICIs in SCC patients was attributed to histological type itself, or merely to the higher prevalence of PD-L1 positivity in this subtype. This represents a major limitation in interpreting the underlying mechanism of differential ICI response between SCC and non-SCC, and highlights an urgent need for future prospective studies to design combined stratification by histology and PD-L1 expression; fourth, no eligible trial provided adverse events stratified by SCC and non-SCC subtypes. As a result, the safety profiles and risk-benefit balance of ICIs could not be compared between the two histological groups, which is essential for clinical decision-making. Future large-scale, prospective RCTs with long-term follow-up, standardized biomarker testing, and subtype-stratified safety and efficacy data are warranted to validate our findings and optimize individualized ICI therapy for cervical cancer. Conclusions This meta-analysis confirms that ICI therapy can significantly improve the clinical outcomes of patients with metastatic cervical cancer, but its efficacy exhibits significant heterogeneity between SCC and non-SCC subtypes. Cervical SCC patients show a more sensitive response to various ICI regimens (anti-PD-1, anti-PD-L1, anti-PD-1/CTLA-4), while non-SCC patients only obtain a significant OS benefit from anti-PD-1 monotherapy and a PFS benefit from anti-PD-L1 therapy. These findings highlight the necessity of formulating personalized ICI treatment strategies based on the histological subtype of cervical cancer in clinical practice to optimize the therapeutic efficacy of immunotherapy. Given the limitations of this study, further large-scale multicenter RCTs and comprehensive meta-analyses are needed to validate these findings and explore more effective ICI-based combination therapy regimens for cervical non-SCC patients. Supplementary The article’s supplementary files as Acknowledgments The authors would like to thank Liuzhou People’s Hospital Affiliated to Guangxi Medical University. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Footnotes Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0345/rc Funding: None. Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0345/coif). The authors have no conflicts of interest to declare. References Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0345/prf Supplementary Materials The article’s supplementary files as