Graphical abstract

INTRODUCTION

Ovarian cancer ranks as the eighth most common cancer in women worldwide, with an age-standardized incidence and mortality rates of 6.7 and 4.0 per 100,000 women, respectively [1]. High-grade serous ovarian carcinoma (HGSOC) accounts for approximately 85%–90% of cases of ovarian carcinoma and exhibits aggressive clinical behavior [2]. Despite advancements in surgical and chemotherapeutic interventions, survival remains poor due to late-stage diagnoses and the emergence of chemoresistance. The standard treatment for ovarian cancer is debulking surgery followed by adjuvant chemotherapy. In a proportion of patients where this is not feasible, neoadjuvant chemotherapy (NACT) followed by cytoreductive or interval debulking surgery and adjuvant chemotherapy is administered, which is non-inferior to the standard approach [3]. In such cases, it is mandatory to obtain a tissue diagnosis by core biopsy of the mass lesion. In patients with ovarian carcinoma presenting with ascites, routine cytological evaluation is performed, and immunohistochemistry on effusion cell blocks can confirm HGSOC histotype, an acceptable alternative to core biopsy [4]. In these cases, the pathologist can assess the response to chemotherapy objectively using the chemotherapy response score (CRS) on the interval debulking surgery specimen [5].

The tumor microenvironment in ovarian carcinoma plays a pivotal role in tumor progression and therapeutic response. It involves a complex interplay among cancer cells, stromal components, and infiltrating immune cells. The immune cells comprise T cells and their major subsets, including CD4⁺ (T-helper) and CD8⁺ (T-cytotoxic) cells, B cells, and the macrophages (M1 and M2 subsets). Tissue-based studies have evaluated the prognostic implications of the tumor microenvironment [6-8], but only a few have examined the immune microenvironment in effusions in general or specifically in HGSOC [9-11]. Hence, this was the major objective of the current study. We applied flow cytometric immunophenotyping to effusion samples from ovarian carcinoma patients and compared them with effusions from other cancers and non-malignant effusions. Genetic alterations, notably mutations in the BRCA1 and BRCA2 genes, which are involved in DNA repair, define a subset of HGSOC tumors. These mutations influence tumor immunogenicity by increasing mutational burden and neoantigen presentation, potentially enhancing immune recognition [12]. Despite this, the interactions between BRCA mutation status, immune cell infiltration, and treatment response remain poorly understood, and hence, this was the secondary objective. As these patients receive NACT, we further explored the associations between immune and genetic parameters and CRS and clinical outcomes, including progression-free survival (PFS) and overall survival (OS). Such insights are essential for identifying prognostic biomarkers and therapeutic targets for personalized treatment of ovarian carcinoma patients presenting with ascites.

While regulatory T cell (Treg) enrichment within ovarian cancer ascites is described, limited data exist regarding the BRCA1/2-stratified immune composition of treatment-naive HGSOC ascites. Accordingly, this study employed flow-cytometric immune profiling to characterize treatment-naive HGSOC ascites, evaluating its association with tumor-detected BRCA1/2 alterations and clinical outcomes.

MATERIALS AND METHODS

Patient selection and sample collection

All patients with significantly high-volume ascites at presentation and who were treatment-naive were included in the study. Ascitic fluid samples were collected at initial presentation, before the initiation of NACT. A total of 133 patients presenting with ascites were included in the study between April 2022 and October 2024. The cohort comprised three groups: (1) 69 patients with histologically confirmed HGSOC, who were managed according to standard institutional protocols with NACT followed by interval cytoreductive surgery and adjuvant chemotherapy, where clinically feasible; (2) 31 patients with ascites secondary to non-ovarian cancers (NOC) arising from the colon, stomach, pancreas, or gallbladder; and (3) 33 patients with non-malignant ascites (NM) associated with diverse etiologies, including chronic liver disease, tuberculosis, or uncertain etiology.

The distribution of cases across the clinical outcome, CRS, and molecular analysis subsets is summarized in Supplementary Fig. S1. In all cases, the presence or absence of malignancy was confirmed by cytological evaluation. Any case of ovarian malignancy other than HGSOC was excluded from the study.

Ascitic fluid (30 mL–1 L) was centrifuged (2,500 rpm, 10 minutes), and conventional cytology smears and one liquid-based cytology preparation (BD SurePath, BD Diagnostics, Burlington, NC, USA) were prepared, stained with Papanicolaou and May-Grünwald-Giemsa stains and evaluated for the presence of malignant cells. Cell blocks from malignant effusions were prepared using the sodium-alginate method [13] for immunohistochemistry to confirm the origin and histotype in HGSOC samples. The immunohistochemical panel included paired box 8 (PAX8), Wilms tumor 1 (WT1), p53, cytokeratin (CK) 7, CK20, GATA3, thyroid transcription factor 1, special AT-rich sequence-binding protein 2, and caudal-type homeobox 2. The non-ovarian malignant effusions, represented in Supplementary Fig. S2B, were not further subtyped; however, all were classified as adenocarcinomas on cytological evaluation. HGSOC was confirmed by nuclear expression of PAX8 and WT1 and aberrant p53 expression (Supplementary Fig. S2D–F). In the ovarian carcinoma cases, clinical details including age, International Federation of Gynecology and Obstetrics (FIGO) stage, carbohydrate antigen 125 (CA-125) levels, treatment response, and clinical outcome were retrieved from medical records and are provided in Supplementary Table S1.

Flow cytometric analysis of ascitic fluid

Ascitic fluid samples (50–100 mL) were centrifuged at 3,000 ×g for 10 minutes. The cell sediment was collected, checked for adequacy, washed with phosphate-buffered saline, and treated with BD lysing solution (BD Biosciences, San Jose, CA, USA), wherever required, to remove red blood cells. The cell pellet was stained with fluorochrome-conjugated monoclonal antibodies against epithelial cell adhesion molecule (EpCAM), CD45, CD3, CD4, CD8, CD19, CD20, CD22, CD68, CD163, CD7, CD16, CD56, CD25, CD127, FOXP3, programmed cell death 1 (PD-1), and programmed cell death ligand 1 (PD-L1), as detailed in Supplementary Table S2. Intracellular FOXP3 staining was performed after fixation and permeabilization according to the manufacturer’s protocol.

Flow-cytometric acquisition was performed on a 10-color BD FACSCanto flow cytometer using BD FACSDiva software (BD Biosciences). Approximately 1 × 10⁶ total events were acquired per sample, depending on sample cellularity. Data were analyzed primarily using BD FACSDiva software, with additional review in FlowJo where required.

A sequential gating strategy was used. After exclusion of debris and doublets, intact single cells were selected. CD45⁺ leukocytes and EpCAM⁺ tumor cells were quantified as proportions of intact single cells. CD3⁺ T cells, CD19⁺CD20⁺ B cells, CD68⁺ macrophages, and CD163⁺ macrophages were quantified within the CD45⁺ leukocyte population. CD22⁺ B cells were quantified within the CD19⁺ B-cell population. CD4⁺ and CD8⁺ T cells were quantified within the CD3⁺ T-cell population. Tregs were identified sequentially as CD3⁺ lymphocytes → CD4⁺ T cells → CD4⁺CD25⁺ T cells → CD25^highCD127^low cells, with FOXP3 expression used for confirmation. Thus, FOXP3-confirmed CD25^highCD127^low Tregs were quantified within the CD4⁺ T-cell population. For natural killer (NK) cell analysis, CD3⁻CD7⁺ cells were gated within the CD45⁺ lymphocyte population, followed by quantification of CD16⁺CD56⁺ cells within this population. PD-1 expression was assessed within the corresponding CD4⁺ and CD8⁺ T-cell subsets, whereas PD-L1 expression was assessed within EpCAM⁺ tumor cells, CD68⁺ macrophages, and CD163⁺ macrophages.

Targeted next-generation sequencing

Targeted sequencing was performed in 39 HGSOC cases. Genomic DNA was extracted from tumor-containing ascitic cell sediments using the QIAGEN DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany). The cell sediments contained approximately 10⁷–10⁸ cells, with tumor-cell proportions ranging from 2% to 88%. DNA concentrations ranged from 21 to 900 ng/µL. Libraries were prepared using the Ion AmpliSeq Library Kit 2.0 (Thermo Fisher Scientific, Carlsbad, CA, USA) and sequenced on the Ion S5 System. A six-gene panel comprising TP53, BRCA1, BRCA2, PIK3CA, ERBB2, and NF1 was selected based on The Cancer Genome Atlas and cBioPortal datasets. The target regions and flanking sequences specified in Supplementary Table S3 were sequenced. Because matched-normal and germline testing were not performed, BRCA1/2 variants were designated tumor-detected alterations and could not be classified as germline or somatic.

Libraries were quantified by Agilent Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and analyzed with Ion Reporter v5.10. Variants were annotated using OncoKB, ClinVar, and VarSome and classified per American College of Medical Genetics and Genomics guidelines as pathogenic, likely pathogenic, variant of uncertain significance, likely benign, or benign. The sequencing data generated in this study have been deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under BioProject accession PRJNA1283496. Data availability details are provided in the Availability of Data and Material section.

Follow-up and clinical outcomes

All 69 recruited patients with HGSOC were followed up in the Department of Radiation Therapy and Oncology, with nine cases lost to follow-up midway through treatment, leaving 60 cases with complete data. The treatment response was assessed according to Response Evaluation Criteria in Solid Tumors (RECIST) ver. 1.1 criteria, along with serum CA-125 levels according to Gynecological Cancer Intergroup criteria. PFS was defined as the time from treatment initiation to documented disease progression or death from any cause. OS was defined as the time from diagnosis to death from any cause or last follow-up. Patients were grouped as follows: complete responders (n = 15), partial responders (n = 19), and progressive disease (n = 26). For statistical analysis, the responders were clubbed and compared with non-responders (progressive disease). The CRS was evaluated on the omental histological sections of the interval debulking specimen and was graded as 1 (minimal response, n = 19), 2 (partial response, n = 18), or 3 (complete response, n = 13) [5].

Statistical analysis

Continuous variables were summarized as median with interquartile range or mean with range, as appropriate. Intergroup comparisons of immune-cell and immune-checkpoint markers across the three study groups were performed using the Kruskal-Wallis test, followed by Dunn’s post hoc test for multiple comparisons. Two-group comparisons were performed using the Mann-Whitney U test, where applicable. Inter-parameter correlations were assessed using Spearman’s rank correlation test.

Survival curves were generated using the Kaplan-Meier method and compared using the log-rank (Mantel-Cox) test. For survival analysis, immune markers were dichotomized into high- and low-expression groups using the median value of the corresponding marker in the survival-analysis cohort. OS time in months was used as the time-to-event variable, and death was coded as the event; patients alive at the last follow-up were censored.

For exploratory integrated survival analysis, Cox proportional hazards regression was performed in 33 HGSOC cases with complete data on BRCA1/2 status, the evaluated immune-cell variables, and OS. BRCA1/2 status was first evaluated in a univariable Cox proportional hazards model. Separate exploratory multivariable Cox proportional hazards regression models were then constructed, with each model including BRCA1/2 status together with one immune-cell variable: CD3⁺, CD4⁺, CD8⁺, CD19⁺CD20⁺ B-cell, or CD22⁺ B-cell group. For the median-dichotomized immune-cell variables, the high-expression group served as the reference category, whereas the BRCA1/2-mutated group served as the reference category for BRCA1/2 status.

Statistical analyses were performed using R ver. 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria) with the dplyr, rstatix, survival, survminer, ggplot2, and ggpubr packages. Exploratory Cox proportional hazards regression was performed using JASP software (JASP Team, Amsterdam, The Netherlands) with right censoring and the Efron method for handling ties. GraphPad Prism 10 (GraphPad Software, Boston, MA, USA) was used for data visualization. A p-value <.05 was considered statistically significant.

RESULTS

A total of 133 ascitic fluid samples from three groups of patients: HGSOC (n = 69), NOC (n = 31), and NM controls (n = 33) were evaluated. The ascitic fluid cytology of representative cases is illustrated in Supplementary Fig. S2.

Ascitic fluid immune cell composition in HGSOC compared to NOC and NM groups

Multicolor multiparametric flow cytometry was performed to evaluate immune cell subsets and immune-checkpoint markers (Fig. 1). The intergroup comparison is presented in Table 1, and significant differences are illustrated in Fig. 2.

HGSOC ascitic fluid showed a pronounced reduction in CD19⁺ CD20⁺ B cells compared to NOC and non-malignant effusion samples, with marked elevation of CD3⁺ and CD8⁺ cytotoxic T lymphocytes. In contrast, CD4⁺ T cells were highest in NOC. Treg (CD25^highCD127^low) cell frequencies differed significantly among the three study groups (p < .001), with higher levels in malignant effusions than in non-malignant controls and relatively higher levels in NOC than in HGSOC.

The CD163⁺ macrophage subtype was reduced in the HGSOC group. CD8⁺PD-1⁺ T-cell levels differed significantly among the three study groups and were highest in the NOC group (p = .032), whereas CD4⁺PD-1⁺ T-cell levels did not differ significantly among the groups. PD-L1 on HGSOC tumor cells (EpCAM⁺ cells) was negative in seven cases, <1% in 22 cases, 1%–10% in 23 cases, and >10% in 17 (24.6%) cases. PD-L1 expression on EpCAM⁺ tumor cells did not differ between the HGSOC and NOC groups. On the other hand, the levels of PD-L1⁺ macrophages (CD68 and CD163) were significantly lower in HGSOC than in the other groups.

The association among immune-cell expression levels were assessed using Spearman's correlation analysis. HGSOC revealed a correlation between CD4⁺PD-1⁺ and CD8⁺PD-1⁺ (r = 0.59, p < .001), CD68⁺PD-L1⁺ with CD163⁺PD-L1⁺ (r = 0.59, p < .001), PD-L1 EpCAM expression with CD8⁺PD-1⁺ cells (r = 0.36, p = .004) and CD4⁺CD25⁺ T cells with CD8⁺PD-1⁺ expression (r = 0.43, p < .001) (Supplementary Table S4, Supplementary Fig. S3).

We then focused on the HGSOC cohort whose clinical, pathological, and outcome data are detailed in Table 2.

Targeted genetic sequencing

Targeted next-generation sequencing was performed in 39 ascitic fluid samples (Supplementary Tables S5, S6) on cell sediments containing HGSOC cells. TP53 mutations were observed in 38 cases (97.4%). These included a range of hotspot missense variants (Y220C, I195T, R175H, and R273H) and truncating/frameshift mutations (Q192*, D228Efs11*, F328Sfs17*, S241Ifs23*, and T150Hfs20*). The BRCA1/2 variants represented tumor-detected mutations; because matched-normal and germline testing were not performed, their germline or somatic origin could not be determined. BRCA1 mutations were identified in 11 samples (28.2%), comprising truncating mutations (Q780*, L598*, and L1089*) and insertion/frameshift variants (S1480Cfs*26, E23Rfs*18, and E111_N112insK*). BRCA2 mutations (Q175*, E1646*) were found in three samples (7.7%). The PIK3CA-E545Q hotspot mutation was observed in four samples (10.3%). The mutation mapper and Integrative Genomics Viewer plots for mutations in representative cases for the TP53, BRCA1, BRCA2, and PIK3CA genes are shown in Supplementary Fig. S4. A detailed analysis file from cBioPortal is included in Supplementary Table S7.

BRCA1/2 mutation status, survival, and ascitic immune-cell composition

BRCA1/2-stratified analysis for immune cell parameters was performed in 38 cases after excluding one case that showed no detectable alteration in TP53, BRCA1/2, or PIK3CA. BRCA1/2-mutated (BRCA^mut/+) cases (n = 14) were compared with BRCA-wild-type (BRCA^WT) cases (n = 24). BRCA^WT cases showed significantly lower levels of B cells (p = .038) and CD4⁺ T cells (p = .037), and higher levels of Tregs (p = .029) and CD22⁺ B cells (p = .029). Further, survival data were available in 33 of these 38 cases. BRCA^mut/+ cases (n = 14) as compared to BRCA^WT cases showed higher OS (33 vs. 15 months, p = .004) and PFS (22 vs. 13 months, p < .001) (Fig. 3).

Immune cells in chemotherapy responders versus non-responders

CRS was available in 50 HGSOC cases, comprising CRS 1 (n = 19), CRS 2 (n = 18), and CRS 3 (n = 13) cases. Patients with CRS 3 showed numerically longer median OS than those with CRS 1/2 (33 vs. 20 months); however, the difference was not statistically significant by the log-rank (Mantel-Cox) test (p = .097). PFS also did not differ significantly between the groups (p = .107). CD22⁺ B-cell levels were higher in the CRS 1/2 group than in the CRS 3 group (p = .016) (Supplementary Fig. S5).

Patients were classified as complete responders, partial responders, and progressive disease groups according to RECIST ver. 1.1 criteria and CA-125 levels. The complete and partial responders were considered responders, while those with progressive disease were deemed non-responders, showing a significant difference in OS and PFS (Supplementary Fig. S6A). CD3⁺ and CD8⁺ T cells were higher in non-responders than in responders (p < .05) (Supplementary Fig. S6B).

Immune cells and survival

The median OS of the HGSOC cohort with available clinical follow-up data (n = 60) was 15 months (range, 1 to 33 months), and the median PFS was 13 months (range, 1 to 30 months). An exploratory univariate survival analysis was performed to identify immune parameters associated with survival using the log-rank (Mantel-Cox) test; the results are shown in Supplementary Table S8. High levels of CD3⁺ and CD4⁺ T cells were significantly associated with poorer OS (p = .002 and p = .009, respectively), and the corresponding Kaplan-Meier survival curves are shown in Fig. 4. Other immune-cell subsets, including CD8+ T cells, Tregs and the immune-checkpoint inhibitors, PD-1 and PD-L1 marker expression, were not significantly associated with survival.

To further evaluate the association between BRCA1/2 status and OS, exploratory Cox proportional hazards regression was performed in 33 HGSOC cases with complete molecular, immune-cell, and survival data, including 22 deaths. BRCA^WT status was significantly associated with poorer OS (hazard ratio [HR], 3.95; 95% confidence interval, 1.43 to 10.88; p = .008). This association remained significant in separate exploratory multivariable models that included CD3⁺, CD4⁺, CD8⁺, CD19⁺CD20⁺ B-cell, or CD22⁺ B-cell groups individually (adjusted HR range, 3.83 to 4.28; all p < .05), whereas none of the evaluated immune-cell groups was independently associated with OS.

DISCUSSION

Malignant ascites in ovarian carcinoma portends a poor prognosis [14]. In view of the advanced stage of disease, often these patients are treated by NACT as they are unfit for upfront surgery. NACT has been shown to be non-inferior to upfront surgery in this setting. All patients of ovarian carcinoma in this cohort were in advanced FIGO stage III or above, and underwent NACT followed by debulking surgery and adjuvant completion chemotherapy [3,15]. According to Ford et al. [16], evaluating ascitic fluid, including its cellular and non-cellular components, offers an opportunity for translational research that remains relatively unexplored. In the first part of the study, we compared the cellular components of ascitic fluid from ovarian carcinoma with those from NOC and NM. In malignant ascites, the proportion of tumor cells (EpCAM⁺) ranged from 2% to 88% of all cells, with no difference between the ovarian carcinoma and the NOC groups. All cases of ovarian carcinoma were high-grade serous carcinoma histotype, confirmed by immunohistochemistry to be positive for PAX8 and WT1, and to exhibit aberrant p53 expression in the effusion cell blocks. The utility of effusion cell blocks for confirming the specific histological type has been reported previously [4].

Immune profiling of effusion specimens from ovarian cancer patients using flow cytometry has been previously reported in only a few studies [8,10,11,17]. The non-malignant effusions represented a heterogeneous population of varied etiologies; similarly, the NOC were from different primary sites and may influence the immune cell composition. Nevertheless, when compared with the HGSOC ascites, there were some significant differences. The striking reduction in B-cells (CD19⁺CD20⁺) subsets in HGSOC suggests suppression of humoral immunity with impaired antigen presentation and antibody production [18]. Besides this, B cells in the tumor environment also provide support to T-cells and innate mechanisms involving macrophages and NK cells [18].

Ascitic fluid in HGSOC patients showed high levels of CD3⁺ T cells, consistent with a previous report [11]. CD4⁺ T-cell levels were highest in NOC patients, implying helper T-cell dominance in non-ovarian tumors. CD4⁺ T cells are now recognized as playing a role in the priming and effector phases of antitumor immune responses. They can also have a cytotoxic effect directly on tumor cells expressing MHC class II molecules and indirectly through co-stimulation and cytokine production [18]. On the other hand, CD8⁺ cytotoxic T cells were increased in HGSOC but not in the other groups. Comparison of ovarian cancer ascites and peripheral blood has revealed enrichment of CD8⁺ T cells in ascites [10].

The most salient observation in our study was that Tregs (CD4⁺ T cells gated further as CD25^highCD127^low), confirmed by FOXP3 staining, were enriched in malignant ascites, in both NOC and in HGSOC groups, compared with non-malignant effusions. This is consistent with previous reports showing enrichment of Tregs in ascites as compared to peripheral blood [8,10] and NM [8]. The active recruitment of Tregs indicates a highly immunosuppressive microenvironment that hinders cytotoxic T-cell activity [8,10,19]. Tregs may be naive or effector in phenotype, and it is the effector Tregs that are enriched in ovarian carcinoma ascites through active recruitment [8,10]. Curiel et al. [8] have demonstrated high levels of macrophage CCL22, which results in active Treg trafficking into the tumor, which effectively inhibits T cells and suppress the production of interferon-γ and IL-2. Effective cancer immunotherapy, therefore, requires a combination strategy that suppresses Tregs while simultaneously activating cytotoxic T cells, which are abundant in the ascitic fluid. A relative depletion of macrophage subsets, particularly the CD163⁺ (M2) subset, was observed in HGSOC effusions. These cells release anti-inflammatory cytokines that promote disease progression [20]. This suggests that in HGSOC ascites, macrophage-driven mechanisms of tumor progression are replaced or compensated by Treg-mediated pathways [8]. Supporting this interpretation, Gottlieb et al. [21] reported that immune suppression can be sustained even in the context of low CD163⁺ macrophage levels, highlighting the adaptability of immunoregulatory circuits. Thus, ascitic fluid evaluation from patients with HGSOC, compared with non-malignant controls, demonstrated significantly increased Tregs and reduced B-cells as the two key mechanisms that can override the increase in CD8⁺ cytotoxic T-cell infiltration and decreased CD163⁺ (M2 polarized) macrophages, indicating a complex immunosuppressive state that facilitates the tumor’s immune escape [10].

Targeted sequencing was performed in ascitic fluid sediments of HGSOC patients. TP53 mutations were identified in 97.4% of cases (38/39), which is comparable to the near-universal TP53 alteration frequency reported in large HGSOC genomic studies [22,23]. The single TP53 mutation-negative case was reviewed by an expert gynecological pathologist and based on its morphology and immunohistochemistry, the diagnosis of HGSOC was retained. The broad spectrum of TP53 variants, ranging from canonical hotspot missense variants (Y220C, I195T, R175H, and R273H) to frameshift and truncating variants, was consistent with patterns reported in large HGSOC genomic studies. Approximately 76.7% of the variants were located in exons 5–8, which encode the DNA-binding domain of p53 [22,23]. Additionally, two cases showed mutations (R273H, R175H) considered to have a gain-of-function effect [23]. Notably, the Y220C and Q192* mutations were seen in four and five cases, respectively.

Tumor-detected BRCA1 mutations were identified in 28.2% (11/39) of cases with ovarian carcinoma ascites, slightly exceeding the reported rates of 20%–25%, which may reflect cohort-specific or detection-sensitivity factors [24]. Tumor-detected BRCA2 mutations were identified in three samples (7.7%), aligning well with reported frequencies [24]. These truncating and frameshift alterations are indicative of homologous recombination deficiency and suitability for poly(ADP-ribose) polymerase (PARP) inhibitor therapy [24]. Germline BRCA mutation testing was not performed as permission could not be obtained from the study participants due to social concerns. BRCA^mut/+ patients showed superior PFS and OS, consistent with previous reports linking BRCA mutations to improved chemotherapy response and suitability as candidates for synthetic lethality with PARP inhibitors [25]. All four PIK3CA-mutated samples harbored the E545Q hotspot mutation (10.3%), recognized as a constitutive alteration in the AKT pathway [26]. In these patients, phosphoinositide 3-kinase–targeted therapies may be considered. The feasibility and reliability of sequencing ascitic fluid cell sediments to capture these prevalent, clinically actionable mutations support their use for genomic profiling, longitudinal disease monitoring, and informed therapeutic decision-making [16,27].

In this group of patients with HGSOC, responses to chemotherapy (paclitaxel and carboplatin) are variable. The CRS on the interval debulking histopathology specimen is well validated as prognostically significant [5], and is performed routinely; our data are consistent with this report [5]. The overall response is also determined by RECIST criteria and serial CA-125 measurements. As the numbers of responders and non-responders were small, our observations on the correlation between ascitic fluid immune parameters and chemotherapy response are trends that require larger numbers of patients to achieve sufficient statistical power. CD22⁺ B cells were increased in the CRS 1/2 group (non-responders); however, the functional significance of this finding remains uncertain because additional B-cell activation, regulatory, and functional markers were not assessed [18]. We also observed increased CD8⁺ T-cell levels in chemotherapy non-responders. There are no similar studies to compare our observations in the effusion samples.

The prognostic effect of the tumoral immune microenvironment was assessed to generate proof of principle, given the small numbers. In HGSOC, higher levels of CD3⁺ and CD4⁺ T cells were significantly associated with poorer OS, while CD8⁺ T cells did not show any effect. In a previous report by Miceska et al. [11], high CD4⁺ T cells and low CD8⁺ T cells correlated with poor survival. Although T cells are generally considered anti-tumoral, these findings suggest that immune cell abundance alone does not reflect functional anti-tumor activity. Indeed, in advanced ovarian cancer, particularly in malignant ascites, T cells frequently exhibit an exhausted phenotype, limiting their cytotoxic potential [28-30]. In fact, as reported by Lieber et al. [9], a favorable prognosis in ovarian cancer correlated more strongly with the presence of activated, effector-memory CD8⁺ T cells and elevated CXCL9 levels in ascitic fluid than with the sheer number of CD8⁺ T cells. Overall, these findings indicate that the ascitic tumor microenvironment is highly complex, and the balance between effector and regulatory immune subsets determines clinical outcome.

The PD-L1 expression was detected in a significant proportion of HGSOC cases, with nearly 24% showing more than 10% of cells positive in the ascitic fluid. This has made PD-1/PD-L1 signaling a therapeutic target across various malignancies, including ovarian cancer [31]. Although CD8⁺PD-1⁺ T-cell levels differed significantly among the study groups and were highest in the NOC group, CD4⁺PD-1⁺ T-cell levels did not differ significantly. Neither PD-1–expressing T-cell population was significantly associated with chemotherapy response or survival. The lower frequency of PD-L1⁺ CD68⁺ and CD163⁺ macrophages in HGSOC likely reflects the mixed-polarization and functional heterogeneity of ovarian cancer-associated macrophages, rather than a lack of immunosuppression [20]. Moreover, since PD-L1 expression is reportedly higher on monocytic myeloid-derived suppressor cells than on monocytes and macrophages in ovarian cancer ascites, PD-L1–mediated suppression may extend to these alternative populations [32]. Consequently, the reduced frequency of PD-L1⁺ macrophages in our cohort does not preclude an immunosuppressive environment; rather, it points to the potential contribution of PD-L1–independent or non-macrophage-mediated immunoregulatory mechanisms, underscoring the necessity for comprehensive myeloid phenotyping and functional validation.

Analysis of BRCA^mut/+ subgroups revealed higher levels of CD4⁺ T cells and B cells and a reduction in Tregs in effusion, suggesting enhanced immune activation. There are no comparable studies in the literature on ascitic fluid samples. However, comparative studies on tissue samples from ovarian cancer patients have revealed higher tumor-infiltrating lymphocytes and CD8⁺ T cells, as well as higher levels of PD-1 and PD-L1 [12].

In the present study, BRCA^WT cases showed relatively higher Treg levels and poorer OS than BRCA1/2-mutated cases. In separate exploratory multivariable Cox models, BRCA^WT status remained significantly associated with poorer OS after adjustment for CD3⁺, CD4⁺, CD8⁺, CD19⁺CD20⁺ B cell, and CD22⁺ B-cell groups, whereas none of these immune-cell subsets was independently associated with OS. Accordingly, the observed Treg enrichment and BRCA-related survival differences should be interpreted as parallel exploratory associations; the present findings do not establish that Treg enrichment mediates the adverse outcome observed in BRCA^WT cases.

The strength of this study lies in the inclusion of only histologically confirmed HGSOC cases, thereby eliminating inter-histotype heterogeneity. Furthermore, ascitic fluid flow cytometry was performed using rigorous gating strategies, particularly for the enumeration of Tregs, which included confirmation by FOXP3 expression. Paired tumor tissue analysis could not be performed to assess concordance between immune biomarkers and BRCA/TP53 alterations in ascites and solid tumor specimens, as biopsy is rarely performed in patients with ascites and the required pre-chemotherapy diagnosis, including tumor histotype, is provided by cell block immunohistochemistry. Ovarian cancer ascites, rather than acting as a surrogate for tumor tissue, reflects the unique peritoneal microenvironment in advanced disease [33]. The integrated Cox regression analyses based on a limited cohort of 33 cases should therefore be considered exploratory. To reduce overfitting, immune-cell variables were added one at a time rather than together in a single multivariable model. Another limitation was the lack of functional assays, such as cytokine measurements, to confirm immune-cell activity. In addition, germline BRCA mutation status was not assessed because participant consent could not be obtained. Therefore, these findings require validation in larger independent cohorts with complete molecular, immune, and survival data.

In conclusion, ascitic fluid from patients with HGSOC demonstrated a Treg-enriched and reduced B-cell immune microenvironment. Targeted sequencing of ascitic fluid cell sediments was feasible and identified TP53 and BRCA1/2 mutations. BRCA^WT cases, in comparison to BRCA1/2-mutated cases, were associated with higher levels of Tregs and poorer OS. Ascitic fluid may therefore serve as a useful source for integrated immune and molecular profiling in HGSOC, with prognostic implications that require validation in larger independent cohorts.

Supplementary Information

Fig. 1.

Representative flow-cytometric immunophenotyping of ascitic fluid immune-cell subsets. Representative flow-cytometric plots illustrating B cells, T-cell subsets, and regulatory T cells (Tregs) in ascitic fluid samples. The upper panel represents high-grade serous ovarian carcinoma, the middle panel represents non-ovarian carcinoma, and the lower panel represents non-malignant ascites. CD4⁺ T cells were further gated for the CD25^highCD127^low population, with FOXP3 expression used for confirmation of Tregs. The representative plots illustrate the gating strategy used for assessment of major immune-cell subsets and Treg-associated populations across the study groups. All values are expressed as % of the parent population, as defined in the Materials and Methods and Table 1 footnote.

Fig. 2.

Intergroup comparison of immune-cell subsets and immune-checkpoint marker expression in ascitic fluid. Selected immune-cell subsets and immune-checkpoint marker expression were compared among HGSOC, NOC, and NM groups. (A) CD3⁺ T cells, CD8⁺ T cells, and CD19⁺CD20⁺ B cells. (B) CD4⁺ T cells, Tregs (CD25^highCD127^low), and CD163⁺ macrophages. (C) PD-1 expression on CD8⁺ T cells, PD-L1 expression on CD68⁺ macrophages, and PD-L1 expression on EpCAM⁺ tumor cells. All values are expressed as % of the parent population, as defined in the Materials and Methods and Table 1 footnote. HGSOC, high-grade serous ovarian carcinoma; NOC, non-ovarian carcinoma; NM, non-malignant ascites; Tregs, T-regulatory cells; PD-1, programmed cell death 1; PD-L1, programmed cell death ligand 1; EpCAM, epithelial cell adhesion molecule. ^*p < .05, ^**p < .01, ^***p < .001; ns, not significant.

Fig. 3.

BRCA mutation status, survival, and immune-cell profile in high-grade serous ovarian carcinoma. (A) Comparison of overall survival and progression-free survival between BRCA-mutated (BRCA^mut/+) and BRCA-wild-type (BRCA^WT) cases. The survival analysis included 33 cases: BRCA^mut/+ (n = 14) and BRCA^WT (n = 19). (B) Comparison of selected immune-cell subsets between BRCA^mut/+ and BRCA^WT cases. The immune-cell analysis included 38 cases: BRCA^mut/+ (n = 14) and BRCA^WT (n = 24). ^*p < .05.

Fig. 4.

Ascitic fluid immune cells and overall survival in high-grade serous ovarian carcinoma (HGSOC) patients. Higher CD3⁺ (A) and CD4⁺ (B) T-cell levels were associated with poorer overall survival.

REFERENCES

• 1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024; 74: 229-63. ArticlePubMedPMC
• 2. Torre LA, Trabert B, DeSantis CE, et al. Ovarian cancer statistics, 2018. CA Cancer J Clin 2018; 68: 284-96. ArticlePubMedPMCPDF
• 3. Kehoe S, Hook J, Nankivell M, et al. Primary chemotherapy versus primary surgery for newly diagnosed advanced ovarian cancer (CHORUS): an open-label, randomised, controlled, non-inferiority trial. Lancet 2015; 386: 249-57. ArticlePubMed
• 4. Bansal A, Srinivasan R, Rohilla M, et al. Morphologic and immunocytochemical features of high-grade serous carcinoma of ovary in ascitic fluid effusion and fine-needle aspiration cytology. Am J Clin Pathol 2020; 154: 103-14. ArticlePubMedPDF
• 5. Bohm S, Faruqi A, Said I, et al. Chemotherapy response score: development and validation of a system to quantify histopathologic response to neoadjuvant chemotherapy in tubo-ovarian high-grade serous carcinoma. J Clin Oncol 2015; 33: 2457-63. ArticlePubMed
• 6. Zhang L, Conejo-Garcia JR, Katsaros D, et al. Intratumoral T cells, recurrence, and survival in epithelial ovarian cancer. N Engl J Med 2003; 348: 203-13. ArticlePubMed
• 7. Hwang WT, Adams SF, Tahirovic E, Hagemann IS, Coukos G. Prognostic significance of tumor-infiltrating T cells in ovarian cancer: a meta-analysis. Gynecol Oncol 2012; 124: 192-8. ArticlePubMedPMC
• 8. Curiel TJ, Coukos G, Zou L, et al. Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival. Nat Med 2004; 10: 942-9. ArticlePubMed
• 9. Lieber S, Reinartz S, Raifer H, et al. Prognosis of ovarian cancer is associated with effector memory CD8(+) T cell accumulation in ascites, CXCL9 levels and activation-triggered signal transduction in T cells. Oncoimmunology 2018; 7: e1424672. ArticlePubMedPMC
• 10. Landskron J, Helland O, Torgersen KM, et al. Activated regulatory and memory T-cells accumulate in malignant ascites from ovarian carcinoma patients. Cancer Immunol Immunother 2015; 64: 337-47. ArticlePubMedPMCPDF
• 11. Miceska S, Skof E, Bucek S, et al. The prognostic significance of tumor-immune microenvironment in ascites of patients with high-grade serous carcinoma. Radiol Oncol 2023; 57: 493-506. ArticlePubMedPMC
• 12. Strickland KC, Howitt BE, Shukla SA, et al. Association and prognostic significance of BRCA1/2-mutation status with neoantigen load, number of tumor-infiltrating lymphocytes and expression of PD-1/PD-L1 in high grade serous ovarian cancer. Oncotarget 2016; 7: 13587-98. ArticlePubMedPMC
• 13. Gupta S, Gautam U, Susheilia S, Bansal B, Uppal R, Srinivasan R. Sodium alginate versus plasma thrombin cell blocks in diagnostic cytopathology: a comparative analysis. Acta Cytol 2022; 66: 72-8. ArticlePubMedPDF
• 14. Matulonis UA, Sood AK, Fallowfield L, Howitt BE, Sehouli J, Karlan BY. Ovarian cancer. Nat Rev Dis Primers 2016; 2: 16061.ArticlePubMedPMCPDF
• 15. Vergote I, Trope CG, Amant F, et al. Neoadjuvant chemotherapy or primary surgery in stage IIIC or IV ovarian cancer. N Engl J Med 2010; 363: 943-53. ArticlePubMed
• 16. Ford CE, Werner B, Hacker NF, Warton K. The untapped potential of ascites in ovarian cancer research and treatment. Br J Cancer 2020; 123: 9-16. ArticlePubMedPMCPDF
• 17. Davidson B, Dong HP, Berner A, et al. Detection of malignant epithelial cells in effusions using flow cytometric immunophenotyping: an analysis of 92 cases. Am J Clin Pathol 2002; 118: 85-92. ArticlePubMed
• 18. Laumont CM, Nelson BH. B cells in the tumor microenvironment: multi-faceted organizers, regulators, and effectors of anti-tumor immunity. Cancer Cell 2023; 41: 466-89. ArticlePubMed
• 19. Sato S, Matsushita H, Shintani D, et al. Association between effector-type regulatory T cells and immune checkpoint expression on CD8(+) T cells in malignant ascites from epithelial ovarian cancer. BMC Cancer 2022; 22: 437.ArticlePubMedPMCPDF
• 20. Reinartz S, Schumann T, Finkernagel F, et al. Mixed-polarization phenotype of ascites-associated macrophages in human ovarian carcinoma: correlation of CD163 expression, cytokine levels and early relapse. Int J Cancer 2014; 134: 32-42. ArticlePubMedPMCPDF
• 21. Gottlieb CE, Mills AM, Cross JV, Ring KL. Tumor-associated macrophage expression of PD-L1 in implants of high grade serous ovarian carcinoma: a comparison of matched primary and metastatic tumors. Gynecol Oncol 2017; 144: 607-12. ArticlePubMed
• 22. Cancer Genome Atlas Research Network. Integrated genomic analyses of ovarian carcinoma. Nature 2011; 474: 609-15. ArticlePubMedPMCPDF
• 23. Zhang Y, Cao L, Nguyen D, Lu H. TP53 mutations in epithelial ovarian cancer. Transl Cancer Res 2016; 5: 650-63. ArticlePubMedPMC
• 24. Ha HI, Park EY, Eoh KJ, et al. Clinical outcomes of BRCA1/2 pathogenic variants in ovarian cancer cluster region in patients with primary peritoneal, epithelial ovarian, and fallopian tube cancer. Gynecol Oncol 2022; 164: 415-20. ArticlePubMed
• 25. Pennington KP, Walsh T, Harrell MI, et al. Germline and somatic mutations in homologous recombination genes predict platinum response and survival in ovarian, fallopian tube, and peritoneal carcinomas. Clin Cancer Res 2014; 20: 764-75. ArticlePubMed
• 26. Campbell IG, Russell SE, Choong DY, et al. Mutation of the PIK3CA gene in ovarian and breast cancer. Cancer Res 2004; 64: 7678-81. ArticlePubMed
• 27. Nozaki T, Sakamoto I, Kagami K, et al. Molecular analysis of ascitic fluid cytology reflects genetic changes of malignancies of the ovary equivalent to surgically resected specimens. Cancer Cytopathol 2022; 130: 640-9. ArticlePubMedPDF
• 28. Webb JR, Milne K, Nelson BH. PD-1 and CD103 are widely coexpressed on prognostically favorable intraepithelial CD8 T cells in human ovarian cancer. Cancer Immunol Res 2015; 3: 926-35. ArticlePubMedPDF
• 29. Fucikova J, Rakova J, Hensler M, et al. TIM-3 dictates functional orientation of the immune infiltrate in ovarian cancer. Clin Cancer Res 2019; 25: 4820-31. ArticlePubMedPDF
• 30. Woo SR, Turnis ME, Goldberg MV, et al. Immune inhibitory molecules LAG-3 and PD-1 synergistically regulate T-cell function to promote tumoral immune escape. Cancer Res 2012; 72: 917-27. ArticlePubMedPMCPDF
• 31. Zamarin D, Burger RA, Sill MW, et al. Randomized phase II trial of nivolumab versus nivolumab and ipilimumab for recurrent or persistent ovarian cancer: an NRG Oncology Study. J Clin Oncol 2020; 38: 1814-23. ArticlePubMedPMC
• 32. Okla K, Rajtak A, Czerwonka A, et al. Accumulation of blood-circulating PD-L1-expressing M-MDSCs and monocytes/macrophages in pretreatment ovarian cancer patients is associated with soluble PD-L1. J Transl Med 2020; 18: 220.ArticlePubMedPMC
• 33. Dogra S, Adhikari L, Benbrook DM, et al. Harnessing ovarian cancer ascites for translational science: models, biomarkers, and therapeutics. Mol Cancer 2025; 24: 257.ArticlePubMedPMCPDF

Figure & Data

References

Citations

Citations to this article as recorded by