Over the past decade, immunotherapy has transformed cancer treatment, markedly enhancing the prognosis of patients with various types of cancer​ (1, 2). In particular, immune checkpoint inhibitors (ICIs) including cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), programmed cell death protein 1 (PD-1), and programmed death-ligand 1 (PD-L1) have shown promise in treating gynecological malignancies (3, 4). Several therapeutic strategies have emerged that integrate ICIs with conventional treatment to enhance response rates and prolong survival. For example, integrating pembrolizumab with chemotherapy has been shown to improve outcomes in patients with advanced endometrial cancer, particularly for tumors with deficient mismatch repair (dMMR) (5). Similarly, combining pembrolizumab with chemoradiotherapy has shown encouraging results in the treatment of locally advanced cervical cancer (6). However, the increased use of new immunotherapy combinations has also raised safety concerns, particularly regarding immune-related adverse events.

Immune-related adverse events are caused by immune activation induced by ICIs, and they most commonly affect the skin, gastrointestinal tract, and endocrine system (7). In our clinical experience, gynecologic cancer patients who develop immune-related adverse events often have improved responses to immunotherapy. However, while growing evidence from other cancer types suggests a positive association between immune-related adverse events and treatment outcomes, data on gynecologic malignancies remain limited (8, 9). Given the increasing use of immunotherapy, further investigations are warranted to determine whether immune-related adverse events may serve as predictive markers of treatment efficacy.

Beyond immune-related adverse events, the identification of reliable biomarkers to predict the response to immunotherapy remains a critical issue due to the heterogeneity in clinical outcomes. Established biomarkers such as PD-L1 expression, MMR, microsatellite instability, and tumor mutational burden can offer valuable insights into patient selection (10, 11). However, a substantial proportion of patients fail to achieve a durable response, underscoring the need for additional biomarkers to optimize treatment outcomes. Recent efforts have focused on simple blood-based indices as potential predictive markers, and particularly the neutrophil-to-lymphocyte ratio (12). The neutrophil-to-lymphocyte ratio reflects the systemic inflammatory response and can serve as a surrogate marker of the balance between pro-tumor and anti-tumor immune activity. An elevated baseline neutrophil-to-lymphocyte ratio has been associated with a poor prognosis in various solid tumors, and recent studies have suggested that it may also be correlated with the outcomes of immunotherapy, including response and survival (13). Despite its potential, the role of neutrophil-to-lymphocyte ratio remains underexplored in gynecologic malignancies.

As the interplay between neutrophil-to-lymphocyte ratio, immune-related adverse events, and treatment efficacy remains unclear, this study aimed to evaluate whether the development of immune-related adverse events and pretreatment neutrophil-to-lymphocyte ratio could serve as predictive markers of treatment response and survival in gynecologic cancer patients receiving pembrolizumab.

This study retrospectively analyzed a consecutive cohort of patients with gynecologic malignancies who received pembrolizumab at Kaohsiung Chang Gung Memorial Hospital between May 2017 and July 2024. Patients were eligible for inclusion if they had a pathologically confirmed diagnosis of gynecological malignancy and had received at least one cycle of pembrolizumab. Patients were excluded if they had synchronous cancers, previously received immunotherapy, not been treated or followed up at the hospital, or incomplete records. The primary objective of this study was to evaluate differences in objective response rate (ORR), progression-free survival and overall survival based on the occurrence of immune-related adverse events. This study was conducted in compliance with relevant ethical guidelines and approved by the Institutional Review Board of Chang Gung Memorial Hospital (No. 202402017B0).

Demographic and clinical variables included age at diagnosis, body mass index (BMI), Eastern Cooperative Oncology Group Performance Status (ECOG-PS), cancer type, International Federation of Gynecology and Obstetrics (FIGO) stage at diagnosis, and number of prior lines of chemotherapy. The neutrophil-to-lymphocyte ratio was recorded immediately preceding the first pembrolizumab administration. Tumor MMR status was assessed using immunohistochemical analysis of MLH1, MSH2, MSH6, and PMS2 expressions. dMMR was defined as the loss of expression of at least one of these proteins. Immunotherapy-related data included pembrolizumab dose per cycle, total number of cycles, combination treatment regimen, and the interval to immune-related adverse event onset. Immune-related adverse events (irAEs) were defined as adverse events with suspected immune-mediated etiology occurring after pembrolizumab therapy onset, for which alternative etiologies (e.g., infection, disease progression, or toxicity from concomitant treatments) were not considered more likely based on clinical evaluation and available diagnostic work-up. Potential irAEs were identified via systematic review of electronic medical records, including oncology progress notes, subspecialty consultation notes, laboratory results, and imaging reports. Events were reviewed and confirmed as irAEs when documented as immune-mediated and managed accordingly by the treating gynecologic oncologist and/or relevant subspecialists. All irAEs were graded using the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 and categorized by organ system (endocrine, gastrointestinal, hepatic, pulmonary, dermatologic, renal, neurologic, and others). If multiple irAEs occurred in a patient, the earliest event was used to define time-to-onset, and the highest CTCAE grade was used to summarize severity at the patient level. Clinical data were retrieved from electronic medical records.

Progression-free survival was defined as the interval from the start of pembrolizumab to disease progression or recurrence. Overall survival was defined as the time to death from any cause or last follow-up. Objective tumor response was evaluated using the Response Evaluation Criteria in Solid Tumors (RECIST 1.1), and the ORR was defined as the percentage of patients with a complete response or partial response. Pretreatment NLR was available for all patients because complete blood counts were routinely obtained prior to pembrolizumab administration as part of standard care. Response assessment was performed according to RECIST v1.1; patients without adequate post-baseline imaging or clinical evidence to evaluate the response were considered not evaluable for response and were excluded from response-rate calculations, with denominators reported for each analysis, and categorized as “Could not evaluate”.

Continuous variables were reported as means and ranges, and categorical variables as numbers and percentages. Survival was analyzed using the Kaplan-Meier method, with group differences assessed by the log-rank test. The optimal pretreatment neutrophil-to-lymphocyte ratio cut-off was determined via receiver operating characteristic curve analysis. The threshold was selected by maximizing the Youden index (sensitivity + specificity − 1) to identify the cut-off that best balanced sensitivity and specificity. Univariate and multivariate Cox proportional hazards models identified significant and independent prognostic predictors of treatment response and survival, respectively. Statistical analyses were conducted using SPSS 25.0 (IBM, Armonk, NY, USA), with a two-sided p-value <0.05 considered significant.

In accordance with the journal’s guidelines, we will provide our data for independent analysis by a selected team by the Editorial Team for the purposes of additional data analysis or for the reproducibility of this study in other centers if such is requested.

This study enrolled 94 gynecologic cancer patients with a median age of 55.5 years (range, 32–83), 11.7% classified as being older (≥70 years). Most (66.0%) had an ECOG-PS score of 0, with diagnoses of corpus uteri (46.8%), cervical (33.0%), ovarian (13.8%), or vaginal cancer (6.4%). FIGO stage III (29.8%) or IV (45.7%) was prevalent at diagnosis. Pembrolizumab was used as first-line (7.5%), second line (34.0%), or third-line/later therapy (58.5%), with a mean pretreatment neutrophil-to-lymphocyte ratio of 4.2 (range, 1.4–92.0). Of 71 patients with MMR testing, 33.8% were MMR-deficient; among 20 cervical cancer patients with PD-L1 testing, 55.0% had a combined positive score (CPS) ≥1. Combination therapy was used in 61.7% of patients, with 39.4% (n=37) receiving pembrolizumab plus chemotherapy and 34.0% (n=32) receiving pembrolizumab with anti-angiogenic agents. The median number of pembrolizumab cycles was 8.7 (range, 1–64). Most patients (88.3%, n=83) received a 100 mg fixed dose per cycle (Table 1).

Overall, immune-related adverse events of any grade were observed in 52 patients (55.3%) with median time to onset was 89.5 days (range, 5 to 546) (Supplementary Table 1). Among them, 46 had low-grade immune-related adverse events, including 28 with grade I and 18 with grade II, while 6 had grade III immune-related adverse events. No grade IV immune-related adverse events or fatal adverse events were reported. The most frequently observed immune-related adverse event was skin related (46.1%), followed by diarrhea (26.9%), hypothyroidism (23.1%) and hyperthyroidism (11.5%). The occurrence of immune-related adverse events was significantly associated with treatment response (Table 2). Patients who developed immune-related adverse events achieved a markedly higher ORR compared to those without immune-related adverse events (84.6% vs. 15.4%, p < 0.001). However, when stratified by immune-related adverse event type, neither skin-related nor thyroid-related immune-related adverse events were significantly associated with treatment response (Supplementary Table 2). The development of immune-related adverse events was not influenced by pembrolizumab dose (p = 0.290) or MMR status (p = 0.153). The pretreatment NLR cut-off was selected based on ROC curve analysis by identifying the threshold that maximized discrimination. The patients with low pretreatment neutrophil-to-lymphocyte ratio (<4.62) demonstrated a significantly higher incidence of irAE than those with high neutrophil-to-lymphocyte ratio (68.6% vs. 39.5%, p = 0.005). Additionally, patients with low pretreatment neutrophil-to-lymphocyte ratio (<4.54) demonstrated a significantly higher ORR than those with high neutrophil-to-lymphocyte ratio (56.0% vs. 27.5%, p = 0.007).

At the time of data cut-off, the median follow-up duration was 9.6 months (range, 0.4 to 89.8). A total of 39 patients had an objective response (ORR = 41.5%), including 15 with a complete response (38.5%) and 24 with a partial response (61.5%). In addition, 44 patients had progressive disease, while the treatment response could not be evaluated in 4 patients. ECOG-PS and combination therapy were significantly correlated with an objective response (p < 0.05). Moreover, dMMR status was significantly associated with a higher response rate (60.9% vs. 28.9%, p = 0.011). In contrast, age, BMI, FIGO stage at diagnosis, pembrolizumab dose per cycle, and immune-related adverse event grade were not significantly correlated with ORR (Table 2).

In the univariate analysis, a lower neutrophil-to-lymphocyte ratio (<4.07) and ECOG-PS score of 0 were significantly associated with longer progression-free survival (p = 0.003 and p = 0.001, respectively). The occurrence of an immune-related adverse event (p < 0.001) and dMMR status (p = 0.030) were also associated with prolonged progression-free survival (Figure 1). In contrast, age and pembrolizumab dose were not significantly associated with progression-free survival. Regarding overall survival, neutrophil-to-lymphocyte ratio, ECOG-PS, occurrence of an immune-related adverse event, and MMR status acted as significant prognostic factors in univariate analysis (Figure 2), while age and treatment dose were not. In the multivariate analysis, both low neutrophil-to-lymphocyte ratio and the occurrence of an immune-related adverse event remained independent predictors of longer progression-free survival, with hazard ratios (HRs) of 0.54 (95% CI, 0.294–0.982; p = 0.043) and 0.23 (95% CI, 0.120–0.443; p < 0.001), respectively. In contrast, age (≥70 years), BMI (≥24 kg/m²), pembrolizumab dose, ECOG-PS, and MMR status were not independently associated with progression-free survival in the multivariate analysis. For overall survival, both low neutrophil-to-lymphocyte ratio (<4.07) and the occurrence of an immune-related adverse event remained independently associated with improved overall survival (HR, 0.33; 95% CI, 0.166–0.650; p = 0.001; and HR, 0.15; 95% CI, 0.070–0.330; p < 0.001, respectively). Similarly, an ECOG-PS score of 0 was also independently associated with longer overall survival (HR, 0.403; 95% CI: 0.201–0.792; p = 0.008), while age, BMI, pembrolizumab dose, and MMR status were not significantly associated with overall survival (Figure 3).

This study is the largest real-world analysis of pembrolizumab outcomes and immune-related adverse events in gynecologic cancer patients. It highlights MMR status, neutrophil-to-lymphocyte ratio, and immune-related adverse events as key predictors of treatment response and survival. Integrating these biomarkers provides a prognostic framework for risk stratification and personalized monitoring during PD-1 blockade therapy. Easily monitored via routine toxicity assessments (immune-related adverse event) and blood tests (neutrophil-to-lymphocyte ratio), these factors guide tailored management strategies. This approach optimizes immunotherapy delivery without restricting access.