Pancreatic ductal adenocarcinoma (PDAC), as one of the most aggressive malignant tumors in the digestive system, has a continuously rising incidence and mortality rate. According to the latest statistics from the American Cancer Society, it is estimated that there will be 67,440 new cases of PDAC and 51,980 deaths in 2025, making it the third leading cause of cancer-related death, with a 5-year relative survival rate of only 13% (1). Currently, first-line treatment for advanced PDAC mainly consists of albumin-bound paclitaxel combined with gemcitabine (AG regimen) or FOLFIRINOX (fluorouracil, leucovorin, irinotecan, and oxaliplatin), but there are significant bottlenecks in efficacy: the median overall survival (mOS) for the AG regimen is 8.5 months, while for the FOLFIRINOX regimen it is 11.1 months, and over 80% of patients experience disease progression within 6–9 months after treatment (2, 3). Although immune checkpoint inhibitors such as PD-1/CTLA-4 have made breakthroughs in multiple cancer types, PDAC has not been able to overcome treatment challenges due to its highly immunosuppressive tumor microenvironment and inherent chemotherapy resistance. Based on this, exploring the synergistic enhancement strategy of AG chemotherapy combined with novel bispecific antibodies (such as the PD-1/CTLA-4 bispecific antibody Cadonilimab) has become an important direction to overcome treatment resistance in PDAC.

Immune checkpoint inhibitors (ICIs) have significantly improved patient survival in solid tumors such as melanoma and non-small cell lung cancer (4, 5). However, their monotherapy efficacy in PDAC remains difficult to overcome. Studies have shown that PDAC has a highly immunosuppressive microenvironment (rich in fibrous stroma, regulatory T cells [Tregs], and myeloid-derived suppressor cells) and multiple immune evasion mechanisms (such as antigen presentation defects and overexpression of co-inhibitory molecules), which are the core reasons for the failure of ICIs (4, 6, 7). Given that chemotherapy drugs have immune regulatory potential (such as inducing immunogenic cell death and enhancing tumor antigen release), researchers have attempted to combine ICIs with the AG regimen to achieve synergistic effects (8). Preliminary clinical data indicate that the PD-1 inhibitor pembrolizumab combined with the AG regimen can increase the median progression-free survival (mPFS) and overall survival (mOS) in metastatic PDAC to 9.1 months and 15.0 months, respectively (9). However, the PD-1 monoclonal antibody nivolumab combined with the AG regimen failed to surpass the efficacy of traditional chemotherapy (10), suggesting that blocking the PD-1 pathway alone may be insufficient to reverse the immune resistance of PDAC (6).

Cytotoxic T lymphocyte-associated antigen 4 (CTLA-4), as an early T cell activation negative regulator, forms a complementary immunosuppressive mechanism with the PD-1 pathway: CTLA-4 mainly inhibits T cell activation in lymph nodes, while PD-1 weakens the effector function of T cells in peripheral tissues (11, 12). Preclinical evidence suggests that dual pathway blockade can significantly enhance anti-tumor activity by promoting T cell clonal proliferation, reducing the proportion of regulatory T cells, and remodeling the tumor microenvironment (13). In solid tumors such as melanoma and hepatocellular carcinoma, PD-1/CTLA-4 dual antibody combination regimens have shown survival benefits superior to monotherapy (14–16). However, PDAC-related research still faces challenges, such as the CCTG PA.7 phase II trial, where the PD-L1/CTLA-4 dual inhibitor (Durvalumab + Tremelimumab) combined with the AG regimen failed to improve patient survival and was associated with an increased incidence of grade 3 or higher toxicities (47% vs 26%) (17), indicating an urgent need for innovative improvements to this treatment regimen. Despite encouraging signals from early-phase studies, PD-1–based regimens have not altered the therapeutic landscape of PDAC. Pembrolizumab plus gemcitabine/nab-paclitaxel achieved mPFS 9.1 and mOS 15.0 months, but the small, single-arm design precluded firm conclusions (9). A phase I trial of nivolumab plus chemotherapy similarly failed to exceed historical benchmarks (10), and the randomized CCTG PA.7 trial of durvalumab plus tremelimumab did not improve OS or PFS while increasing grade ≥3 toxicities (17). Cadonilimab, a tetravalent PD-1/CTLA-4 bispecific with an Fc-silent, pH-dependent design, enables high-avidity intratumoral engagement and has shown a lower incidence of high-grade irAEs than conventional doublets (18–20). On this basis, we hypothesized that cadonilimab plus AG could overcome PDAC’s immune-refractory microenvironment and warranted real-world evaluation.

Cadonilimab is a tetravalent bispecific antibody targeting PD-1 and CTLA-4, which binds to PD-1 (KD = 1.33 nM) and CTLA-4 (KD = 3.79 nM) with high affinity through steric hindrance effects, forming an “immune synergistic activation domain” locally in tumors, significantly enhancing CD8+ T cell infiltration and interferon-γ secretion (18, 21). Its unique structural design (Fc silent mutation and pH-dependent binding) can reduce peripheral tissue-targeted toxicity, and clinical data show that the incidence of immune-related adverse events (irAEs) (32.1%) is significantly lower than that of traditional dual drug combinations (≥60%) (22, 23). In recurrent/metastatic cervical cancer (23), unresectable hepatocellular carcinoma (aHCC), and advanced gastric cancer (G/GEJ), Cadonilimab monotherapy or combination therapy has demonstrated significant survival benefits (mOS reaching 17.5-22.1 months) (20, 21, 24), but its application in PDAC has not been reported.

Based on the potential synergistic mechanism of PD-1/CTLA-4 dual pathway blockade and chemotherapy, this study aims to systematically evaluate the treatment response and safety profile of Cadonilimab combined with albumin-bound paclitaxel/gemcitabine (AG) regimen in patients with unresectable locally advanced or metastatic pancreatic ductal adenocarcinoma (PDAC) through a single-center retrospective study. The core hypothesis is that PD-1/CTLA-4 dual blockade can enhance tumor antigen-specific T cell responses and inhibit regulatory T cell (Tregs) function, synergizing with the immunomodulatory effects of AG chemotherapy to overcome PDAC treatment resistance (25–27). By assessing the dynamic changes of CA19–9 and baseline inflammatory markers, the study will explore their correlation with clinical outcomes (ORR, PFS, OS) and screen for prognostic biomarkers (28, 29).

The highly immunosuppressive microenvironment of pancreatic ductal adenocarcinoma (PDAC) (dense fibrous stroma, regulatory T cell infiltration) results in minimal efficacy of traditional immune checkpoint inhibitors (ICIs) (response rate of less than 2%) (6).This study employs Cadonilimab (a PD-1/CTLA-4 bispecific antibody) in combination with AG chemotherapy, achieving synergistic effects through dual mechanisms: the tetravalent structure of Cadonilimab specifically binds to PD-1/CTLA-4, blocking both immune checkpoint pathways, promoting local T cell activation in the tumor, while the Fc silent mutation reduces peripheral immunotoxicity (18); AG chemotherapy induces immunogenic cell death in tumor cells, releasing tumor-associated antigens and damage-associated molecular patterns, enhancing the antigen presentation function of dendritic cells (8). This “chemotherapy sensitization - immune activation” combination strategy provides a new pathway to overcome the inherent resistance of PDAC.

Research shows that the disease control rate (DCR) of combination therapy reaches 85.7% (12/14), significantly better than the traditional AG regimen’s 65%-70% (2). Although the objective response rate (ORR) is 14.3%, the improvement in DCR is directly associated with survival benefits — the median progression-free survival (mPFS) is 7.87 months, and the median overall survival (mOS) is 11.45 months, extending 0.9 months and 3.0 months respectively compared to the AG monotherapy (2). The survival outcomes (mOS 11.45 months) should be interpreted in the context of consistent patient characteristics with historical cohorts:Von Hoff et al. (2013) (2) reported 68% of patients with ECOG PS 0-1, consistent with our cohort’s functional status distribution and metastatic disease proportion (85.7%) align closely with Von Hoff et al. (2013) (2) and Conroy et al. (2011) (3). This minimizes bias from baseline differences, strengthening cross-study comparability. Potential heterogeneities (e.g., race, comorbidities) warrant confirmation in larger head-to-head trials.The high fibrous stroma of PDAC may delay the morphological response of tumors to treatment, while DCR can more comprehensively reflect the clinical benefits of “disease stability + partial response.” Notably, the modest improvement in mPFS (7.87 vs. 6.9 months with AG monotherapy) carries clinical significance in advanced PDAC, where rapid disease progression often leads to severe symptoms (e.g., pain, obstruction). Even short-term disease control can alleviate symptoms and improve quality of life. The high DCR (85.7%)—markedly higher than the 65%-70% with AG alone (2)—further supports the clinical value of this regimen, as stable disease in this setting translates to meaningful symptom management and reduced need for emergency interventions.

In PDAC, dense fibrotic stroma and poor perfusion can blunt early tumor shrinkage on conventional imaging (34, 35), while immunotherapy may produce atypical or delayed responses (36). Under these conditions, DCR pragmatically captures stable disease plus responses, providing an estimation-focused signal in a small, retrospective cohort with limited follow-up; ORR, PFS and OS are retained as key secondary outcomes for prospective confirmation.

This study found that the baseline PLR ≤165.62 may be a potential predictive factor worthy of further validation.Patients with PLR ≤ 165.6 had a significantly prolonged median overall survival (mOS) of 12.4 months, which is a 41.7% improvement compared to the high PLR group (8.7 months) (P = 0.025), and this is highly consistent with the conclusions of a meta-analysis that included 17 cohorts (HR = 1.28, P<0.0001) (44). This phenomenon may stem from a dual mechanism: low PLR indicates that lymphocyte-mediated anti-tumor immunity is dominant, while the secretion of platelet-derived pro-tumor factors (such as VEGF, TGF-β) is reduced, collectively enhancing the immune activation effect of the dual antibodies (45–47). Additionally, patients with NLR ≤ 3.03 and LDH ≤ 257 U/L had a significantly increased disease control rate (DCR) (P<0.05) (40), indicating that a reduced systemic inflammatory burden can improve treatment response. It is noteworthy that although the association between LMR and objective response rate (ORR) did not reach statistical significance (P = 0.066), the trend still suggests that the monocyte/lymphocyte balance affects tumor-killing potential. These serological markers constitute a low-cost, dynamically monitored prognostic assessment system, especially the strong predictive value of PLR provides a practical tool for clinical screening of beneficiary populations. Compared to previous PD-1 monoclonal antibody combined chemotherapy regimens (such as NLR for patient stratification) (9, 10), this study establishes the core role of PLR in PD-1/CTLA-4 bispecific antibody treatment for the first time, which may stem from the higher dependence of dual antibodies on the functional status of lymphocytes for synergistic blockade (18).

Patients with reduced CA19–9 levels after treatment showed significant survival benefits, with a median overall survival (mOS) of 12.92 months, extending 4.03 months compared to the non-target group (8.89 months) (HR = 4.77, 95% CI 1.20-19.01, P = 0.027, Figures 3A, B ), suggesting that dynamic changes in CA19–9 may serve as a potential predictor of OS, requiring further validation. Notably, there was no statistically significant difference in progression-free survival (PFS) between the two groups (P = 0.142), indicating that CA19–9 is more suitable for assessing overall prognosis rather than short-term disease progression, which may be related to the high fibrous stroma in PDAC leading to delayed morphological responses of the tumor. This result is consistent with previous studies linking the decrease in CA19–9 levels to survival benefits (29, 48), further validating its clinical value in efficacy monitoring. Changes in CA125 levels were independently associated with prolonged PFS (P = 0.018) and may serve as a predictive factor for efficacy in specific treatment scenarios. Combined detection of biomarkers such as MUC5AC and DUPAN-2 can improve the limitations of false negatives of CA19–9 in Lewis-negative individuals (49), while the combined detection of CEA and CA125 aids in the differential diagnosis of pancreatic cancer and other gastrointestinal tumors (50, 51). Multi-biomarker dynamic monitoring provides precise assessment tools for personalized treatment.Exploratory analyses of other tumor markers (CEA, CA125, AFP) were conducted based on two considerations: first, their potential complementary value to CA19-9, as CA19–9 is unreliable in Lewis-negative patients (49), and CEA/CA125 have been reported to aid in differentiating pancreatic cancer from other gastrointestinal tumors (50, 51); second, preliminary evidence suggesting CA125 may correlate with PFS in some gastrointestinal malignancies (29, 51). However, it is important to clarify that these markers are not currently used in routine clinical practice for PDAC, and their predictive value in this setting requires further validation in larger cohorts.

In terms of safety, all patients experienced treatment-related adverse events (TRAE) of grade 1 or higher, with a grade 3 TRAE incidence of 57.1%, primarily consisting of nausea (21.4%) and appetite loss (14.3%), with no grade 4 or higher toxicity reported ( Table 5 ). No immune-related adverse events (irAE) were reported in the entire group, which is significantly lower compared to previous PD-1/CTLA-4 dual-agent regimens (≥60% incidence of grade 3 or higher irAEs) (17), indicating that the Fc-silent design of Cadonilimab significantly reduces immune-related toxicity (18).

This study has the following limitations: First, as a single-center retrospective study, the sample size is small (n=14), and there is no control group, which may lead to selection bias, and the generalizability of the results needs to be validated; second, not all patients’ tumor tissues were obtained for molecular biological analysis, preventing exploration of the drug’s mechanisms of action and resistance at the genetic level;third, the prognostic analysis was conducted solely through serological markers and peripheral blood inflammatory indicators, lacking in-depth mechanism studies such as immune cell typing in the tumor microenvironment;additionally, the PLR cutoff (165.62) derived from limited data may carry overfitting risk; lack of tumor tissue molecular analysis and tumor microenvironment studies limits mechanistic insights; finally, the study did not explore the optimal dosage and treatment duration for combination therapy, which may affect the precision of efficacy assessment. We fully recognize the limitations of a single-center retrospective study with a small cohort (n=14). As shown in Supplementary Table 1 , the study has limited statistical power: only large survival effects (HR ≈0.4–0.45) could be detected with 80% power (37, 38), whereas our observed HR ≈0.74 falls outside this range, indicating a non-trivial risk of type II error. The wide confidence intervals around DCR (≈60–96%) further highlight imprecision (39). These findings should therefore be interpreted as hypothesis-generating rather than confirmatory. In this context, our pilot study primarily aimed to observe safety signals and feasibility, while preliminary efficacy observations serve to inform the design of ongoing larger-scale, prospective studies (ChiCTR2400093744) that can provide more definitive evidence.These limitations suggest that the results of this study need to be further validated in larger sample size multi-center prospective studies.Our institution treated patients with advanced PDAC using AG monotherapy during the same period, but due to unobtained informed consent, these data were not included. A formal comparative analysis will be incorporated into our prospective trial (ChiCTR2400093744).

Baseline PLR ≤165.6 and on-treatment CA19–9 decline were prespecified for validation in our ongoing phase II trial (ChiCTR2400093744) with blinded independent analyses, and will be externally tested in a multicenter registry. Cut-off evaluation (time-dependent ROC with bootstrap validation), non-linear modeling (restricted cubic splines), and prespecified sensitivity analyses (handling of missing data, multiplicity, and alternative thresholds) were planned to ensure robustness (52–55).

Notably, to address the limitations of small sample size and single-arm design, we have initiated a single-arm, open-label, single-center phase II clinical trial (Chinese Clinical Trial Registry: ChiCTR2400093744) in 2024, which aims to further evaluate the efficacy and safety of Cadonilimab combined with AG chemotherapy in unresectable locally advanced or metastatic PDAC. This trial has enrolled some patients and is actively recruiting, with results expected to validate the findings of the current retrospective study and lay the groundwork for subsequent multi-center prospective trials.

In summary,this study suggests potential clinical value of Cadonilimab combined with AG chemotherapy: the DCR reached 85.7%, mOS was 11.45 months, with efficacy close to the FOLFIRINOX regimen but better safety; a baseline PLR ≤ 165.62 and dynamic decline of CA19–9 may suggest potential survival benefits, warranting further validation; grade 3 TRAE was mainly gastrointestinal reactions, with no severe immune-related toxicities. This regimen provides a new direction for immunotherapy in PDAC. Subsequent multicenter phase III trials are needed to further clarify the molecular characteristics of the benefiting population.