CellSearch-based isolation of circulating tumor cells from ascitic fluid enables molecular characterization of ovarian cancer Authors: Francesca Gorini, Thais Maloberti, Emma Coschina, Francesco Mezzapesa, Biljana Petrovic, Giulia Gallerani, Manuela Ferracin, Elisa Scarpelli, Antonio De Leo, Ivana Kurelac, Pierandrea De Iaco, Sabrina Angelini, Anna Myriam Perrone, Dario de Biase, Gloria Ravegnini Published: 2026 ABSTRACT BackgroundHigh-grade serous tubo-ovarian cancer (HGSOC) frequently presents with malignant ascites, a biologically relevant compartment rich in tumor cells. While circulating tumor cells (CTCs) in peripheral blood (PB) are often scarce, ascitic fluid (AF) may represent a more accessible and abundant source of tumor-derived material. This study evaluated the feasibility of isolating AF-derived CTCs using the CellSearch platform for subsequent molecular characterization.MethodsPaired PB and AF samples from 12 HGSOC patients were collected at surgery and analyzed using the CellSearch system. After CTC enumeration, DNA was extracted from CellSearch cartridges containing AF-derived CTC-enriched fractions and subjected to next-generation sequencing (NGS) using a panel clinically validated for solid tumors. Associations between AF-derived CTC burden and clinical variables were explored.ResultsCTCs were detected in 100% of AF samples and in 50% of PB samples. AF contained markedly higher numbers of CTCs than PB, whereas the latter showed low or undetectable levels. AF-derived CTCs frequently formed multicellular aggregates, potentially leading to underestimation by automated counting. An optimized protocol for DNA extraction from post-processing CellSearch cartridges enabled successful downstream NGS analysis from CTC-enriched AF fractions. Tumor-specific TP53 mutations identified in AF-derived CTC-enriched fractions showed complete concordance with matched tumor tissue in all analyzed cases, supporting the feasibility of using these fractions for molecular characterization of the tumor. No significant associations were observed between AF-derived CTC burden and CA125, PCI, progression-free survival (PFS), or overall survival (OS), although higher CTC levels were associated with a trend toward shorter PFS.ConclusionsCompared to PB, AF is a richer source of CTCs in HGSOC. Importantly, AF-derived CellSearch fractions enabled reliable molecular profiling through downstream NGS analysis, suggesting that this approach may be exploited to enrich for tumor-poor peritoneal washings usually considered for staging IC HGSOC. The complete mutational concordance between AF-derived CTCs and matched tumor tissue supports the biological validity of this approach and highlights its potential as a minimally invasive liquid biopsy for tumor characterization and longitudinal disease monitoring in advanced HGSOC. Introduction HighGrade Serous tubo-Ovarian Cancer (HGSOC), the most common and aggressive histological subtype of epithelial ovarian cancer, remains one of the most lethal gynecologic malignancies worldwide, largely due to late-stage diagnosis and the development of treatment resistance. Most patients present with advanced disease characterized by extensive peritoneal dissemination and the accumulation of malignant ascites [[1], [2], [3]]. Ascitic fluid (AF) in HGSOC is a complex and dynamic microenvironment comprising tumor cells, stromal elements, immune cells, cytokines, and extracellular vesicles. Among these components, circulating tumor cells (CTCs) and their aggregates are considered important mediator of metastatic spread within the peritoneal cavity [4,5]. The unique tumor microenvironment in AF plays a crucial role in disease progression, immune modulation, and therapeutic response. CTCs isolated from AF often display enhanced survival, increased resistance to anoikis, stem-like properties, and a marked propensity to form multicellular aggregates. Increasing evidence indicates that these clusters may contribute substantially to peritoneal dissemination, immune evasion, and chemoresistance, as cell-cell interactions promote tumor cell survival, protect against cytotoxic agents, and enhance metastatic potential [6,7]. In fact, unlike hematogenous dissemination observed in other solid tumors, OC predominantly spreads via a transcoelomic routes, where tumor cells detach from the primary lesion and survive in suspension within AF before implanting on peritoneal surfaces [8]. The analysis of CTCs in peripheral blood (PB) has been widely investigated as a valuable approach in several malignancies, including breast, prostate, and colorectal cancer. However, their clinical utility is often limited by their low abundance in circulation, which poses significant technical challenges for their reliable detection and characterization [9,10]. On the other hand, AF presents a richer source of liquid biopsy-derived CTCs [11] but, despite its biological significance and potential clinical utility, ascites remains largely underutilized as a biospecimen in current clinical practice and standard-of-care workflows. Specifically, the FDA-approved CellSearch device (Menarini Silicon Biosystems) for CTC enumeration in PB has never been applied to isolate cell from AF. Demonstrating the feasibility of this approach for CTC enrichment, and downstream molecular profiling would be especially valuable for detecting malignant cells in tumor-poor peritoneal washings. This could improve identification of occult peritoneal disease, enabling upstaging of early-stage HGSOC in patients with positive peritoneal cytology, and guiding decisions regarding adjuvant chemotherapy [12,13]. Thus, in this study, we aimed to isolate and characterize CTCs from AF samples from patients with HGSOC and to compare the resulting recovery rate with CTCs obtained from PB. Furthermore, we investigated the feasibility of extracting tumor-derived DNA from CellSearch-enriched AF-derived CTC fractions for mutational profiling and to assess the mutational profile and evaluate its concordance with the matched primary tumor. Methods Study design and patients’ enrolment This was an exploratory pilot study involving twelve patients diagnosed with HGSOC and treated at the Division of Gynecologic Oncology of the IRCCS Sant’Orsola-Malpighi Hospital in Bologna, Italy. This study was conducted in accordance with the principles of the Helsinki Declaration. Informed consent was obtained from all participants; the study received approval from the Institutional Review Board 788/2021/Sper/AOUBo and was registered to ClinicalTrials.gov (Identifier: NCT05146505). Inclusion criteria were: i) confirmed diagnosis of HGSOC; ii) presence of AF; iii) patients aged 18 years or older. Exclusion criteria included i) a history of other malignancies; ii) the presence of significant concomitant diseases; or iii) prior receipt of chemotherapy. Sample collection and processing PB and AF samples were collected at the time of surgery using CellSave Preservative Tubes (Menarini Silicon Biosystems), which contain a specialized preservative designed to stabilize CTCs for up to 96 h at room temperature. Immediately after collection, tubes were gently inverted three times to ensure adequate mixing with the preservative and to maintain sample integrity for subsequent analyses. For the analysis of ascitic samples, we applied a previously published method originally developed for the detection of tumor cells in cerebrospinal fluid (CSF) [14]. CTC identification and enumeration Samples were analyzed using the CellSearch CTC Kit (Menarini Silicon Biosystems), a clinically validated and FDA-cleared system for the identification, isolation, and enumeration of CTCs of epithelial origin (CD45−, EpCAM+, cytokeratins 8, 18+, and/or 19+) from PB [15,16]. CTCs were defined as intact cells with a diameter >4 μm, a DAPI-positive nucleus located within the cytoplasmic boundaries, positive staining for cytokeratin (CK-PE), negative staining for CD45 (CD45-APC), and exhibiting cellular morphology. Before patient sample analysis, system performance was verified by running the CellSearch CTC Control Kit. CTC analysis was then performed according to the standardized three-step workflow of sample processing, immunomagnetic enrichment, and fluorescence-based imaging using the Celltracks Autoprep and Analyzer II systems. CTCs were manually enumerated using the CellSearch system (Menarini Silicon Biosystems), based on the presence of characteristic signals, including DAPI-positive nuclei and cytokeratin-positive (PE) epithelial markers, together with the absence of a CD45 (APC) leukocyte signal, allowing tumor cells to be distinguished from hematopoietic cells. DNA extraction After CTC enumeration, the CellSearch cartridges were stored at 4 °C protected from light until further processing. Subsequently, the fluid within each cartridge was collected, and the cartridge was washed three times with Dulbecco's phosphate-buffered saline (DPBS), with all wash fractions pooled together. The collected fluids were centrifuged at 1500 x g for 5 min to obtain a pellet, which was washed once with DPBS and centrifuged again at 1500 x g for 5 min before DNA extraction. DNA was subsequently extracted from the resulting cell pellet for downstream molecular analyses. Initially, genomic DNA was extracted using a semi-automated workflow consisting of manual cell lysis followed by nucleic acid purification on the Maxwell platform (Promega) using the Maxwell CSC DNA FFPE Kit (Promega), according to the manufacturer's instructions. Subsequently, because of the low DNA yield, an alternative manual extraction protocol was adopted using the MasterPure Complete DNA and RNA Purification Kit (Epicentre – Lucigen – Bioresearch Technologies). Briefly, 10 μL of proteinase K and 300 μL of tissue lysis buffer were added to each sample, followed by incubation at 56 °C overnight. Then, 175 μL of Tissue and Cell Lysis Solution was added, and the sample was vortexed vigorously for 10 s. The debris was pelleted by centrifugation at 10,000 × g for 10 min at 4 °C in a microcentrifuge. The supernatant was transferred to a clean 1.5 mL tube. 500 μL of isopropanol was added to the recovered supernatant, and the sample was inverted 30–40 times. Total nucleic acids were pelleted by centrifugation at 10,000 × g for 10 min at 4 °C. Isopropanol was poured off without dislodging the pellet, and then the pellet was washed twice with 500 μL of 70% ethanol. The total nucleic acid was resuspended in 12 μL of TE buffer. Mutational profile through NGS solid panel DNA concentration was quantified using a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). Sequencing was performed using multi-gene NGS panels developed in the Solid Tumor Molecular Pathology Laboratory, IRCCS Azienda Ospedaliero Universitaria di Bologna, Bologna, Italy, allowing the analysis of the following genomic regions: BRAF (NM_004333.6) (exon 11, 15), CTNNB1 (NM_001904.4) (exon 3, 7, 8), DICER1 (NM_030621.4) (exon 10, 21, 26, 27, 29), DPYD (NM_000110.4) (DPYD*2A, DPYD*13, DPYD D949V, HapB3), EGFR (NM_005228.5) (exon 18, 19, 20, 21), EIF1AX (NM_001412.4) (exon 1, 2), GNA11 (NM_002067.5) (exon 4, 5), GNAQ (NM_002072.5) (exon 4, 5), GNAS (NM_000516.7) (exon 8, 9), H3F3A (NM_001379043.1) (exon 1), HRAS (NM_001130442.2) (exon 2, 3, 4), IDH1 (NM_005896.4) (exon 4), IDH2 (NM_002168.4) (exon 4), KIT (NM_000222.3) (exon 8, 9, 11, 13, 17), KRAS (NM_033360.4) (exon 2, 3, 4), MED12 (NM_005120.3) (exon 1, 2), MET (NM_001127500.3) (exon 2, 14), NRAS (NM_002524.5) (exon 2, 3, 4), PDGFRa (NM_006206.6) (exon 12, 14, 18), PIK3CA (NM_006218.4) (exon 8, 10, 21), PTEN (NM_000314.8) (exon 5), RET (NM_020975.6) (exon 5, 8, 10, 11, 13, 14, 15, 16), RNF43 (NM_017763.6) (exon 2-9), SMAD4 (NM_005359.6) (exon 2-12), TERT (NM_198253.3) (promotor), TP53 (NM_000546.6) (exon 2-11), TSHR (NM_000369.5) (exon 1-10), VHL (NM_000551.4) (exon 1-3) [17]. The amplicon library preparation was performed with the AmpliSeq Plus Library Kit. Templates were prepared with an Ion Chef Machine, and sequencing was performed using the Gene Studio S5 Prime Sequencer (Thermo Fisher Scientific). Sequencing data were analyzed with the Ion Reporter tool (v. 5.20 – Thermo Fisher Scientific). The pathogenicity of each mutation was assessed using the Varsome tool (https://varsome.com/). Statistical analysis Associations between ascitic CTC levels and clinicopathological variables were assessed using nonparametric statistical tests because of the small sample size and the non-normal distribution of the data. Comparisons between two independent groups were performed using the Mann–Whitney U test, whereas comparisons among more than two groups were carried out using the Kruskal–Wallis test. Correlations between continuous variables were evaluated using Spearman's rank correlation coefficient. All tests were two-sided, and a p-value < 0.05 was considered statistically significant. Results Higher number of CTCs in ascitic fluid compared to peripheral blood samples The characteristics of the 12 patients included in the study are summarized in Table 1.Table 1Demographic and clinical characteristics of OC patients enrolled in the study.Table 1VariablePatients (n = 12)Age at diagnosis, mean (±sd)62 (±9)Age group, N (%)≤50 yo1 (8.33)>50 yo11 (91.67)BMI, mean (±sd) (kg/m2)27.16 (±6.94)FIGO stage (2014), N (%)I0II0IIIA1 (8.33)IIIB2 (16.67)IIIC6 (50)IV3 (25)BRCA status, N (%)BRCA1/2 WT10 (83.33)BRCA1/2 MUT2 (16.67)yo: years old; wt: wild-type; mut: mutant Table 1Demographic and clinical characteristics of OC patients enrolled in the study.Table 1VariablePatients (n = 12)Age at diagnosis, mean (±sd)62 (±9)Age group, N (%)≤50 yo1 (8.33)>50 yo11 (91.67)BMI, mean (±sd) (kg/m2)27.16 (±6.94)FIGO stage (2014), N (%)I0II0IIIA1 (8.33)IIIB2 (16.67)IIIC6 (50)IV3 (25)BRCA status, N (%)BRCA1/2 WT10 (83.33)BRCA1/2 MUT2 (16.67)yo: years old; wt: wild-type; mut: mutant Demographic and clinical characteristics of OC patients enrolled in the study. yo: years old; wt: wild-type; mut: mutant We analyzed paired PB and AF samples from each patient. Because AF samples required multiple dilutions for accurate quantification, a total of 36 CellSearch cartridges were analyzed; Table 2 summarizes CTC enumeration results in PB and AF.Table 2CTC enumeration in blood and ascites samples.Table 2Patient codeSampleDilutionCTC enumeration in the cartridgeOVC-271PBUndiluted2AFUndilutedSaturated Cartridge1:26460OVC-305PBUndiluted45AFUndilutedSaturated Cartridge1:2Saturated Cartridge1:20Saturated Cartridge1:200Saturated Cartridge1:20005562OVC-308PBUndiluted1AF1:20005OVC-310PBUndiluted0AF1:200022OVC-312PBUndiluted0AF1:200017OVC-314PBUndiluted0AF1:200551:2000Not enough objectsOVC-315PBUndiluted0AF1:2003391:2000Not enough objectsOVC-318PBUndiluted0AF1:20002OVC-319PBUndiluted2AFUndilutedSaturated Cartridge1:200451:2000Not enough objectsOVC-327PBUndilutedAbortedAFUndilutedSaturated Cartridge1:200072OVC-333PBUndiluted1AFUndilutedSaturated Cartridge1:20009728OVC-335PBUndiluted1AFUndilutedSaturated Cartridge1:20005890AF: ascitic fluid; PB: Peripheral blood; OVC: ovarian cancer; Table 2CTC enumeration in blood and ascites samples.Table 2Patient codeSampleDilutionCTC enumeration in the cartridgeOVC-271PBUndiluted2AFUndilutedSaturated Cartridge1:26460OVC-305PBUndiluted45AFUndilutedSaturated Cartridge1:2Saturated Cartridge1:20Saturated Cartridge1:200Saturated Cartridge1:20005562OVC-308PBUndiluted1AF1:20005OVC-310PBUndiluted0AF1:200022OVC-312PBUndiluted0AF1:200017OVC-314PBUndiluted0AF1:200551:2000Not enough objectsOVC-315PBUndiluted0AF1:2003391:2000Not enough objectsOVC-318PBUndiluted0AF1:20002OVC-319PBUndiluted2AFUndilutedSaturated Cartridge1:200451:2000Not enough objectsOVC-327PBUndilutedAbortedAFUndilutedSaturated Cartridge1:200072OVC-333PBUndiluted1AFUndilutedSaturated Cartridge1:20009728OVC-335PBUndiluted1AFUndilutedSaturated Cartridge1:20005890AF: ascitic fluid; PB: Peripheral blood; OVC: ovarian cancer; CTC enumeration in blood and ascites samples. AF: ascitic fluid; PB: Peripheral blood; OVC: ovarian cancer; CellSearch identified CTCs in 6 out of 12 PB liquid biopsies (50%), with a median CTC number of 1.5 (±7.5 SD) among positive cases (Table 2), corroborating the rare occurrence of these entities in circulation. Whereas CTCs in PB samples were consistently low or undetectable, AF samples exhibited significantly higher tumor cell burdens. Of note, because the validated linear range of the CellSearch CTC Kit extends up to 1238 cells per cartridge, AF samples were diluted prior to analysis to ensure accurate CTC enumeration and avoid saturation effects. A 1:2000 dilution was selected as the standard dilution for quantification. When this dilution did not yield countable values within the readable range, samples were reanalyzed using a 1:200 dilution. Quantifications at 1:200 and 1:2000 dilutions revealed CTC counts ranging from 2 to 9728, with dilution normalized values ranging from 500 to 1,472,500/mL, underscoring the markedly higher CTC presence in ascitic fluid compared to PB. Manual review of CellSearch images revealed that CTC clusters, which are characteristic of AF samples, were frequently counted as single events by the automated CellSearch algorithm, resulting in an underestimation of the actual number of tumor cells. (Fig. 1).Fig. 1Representative examples of CTC aggregates detected in AF. Although CTCs are present as multicellular aggregates, the automated counting system identifies each aggregate as a single event, leading to an underestimation of the actual number of CTCs. Fluorescence channels are displayed as pseudocolors: cytokeratin (CK-PE) is shown in green and DAPI-stained nuclei in magenta. Representative examples of CTC aggregates detected in AF. Although CTCs are present as multicellular aggregates, the automated counting system identifies each aggregate as a single event, leading to an underestimation of the actual number of CTCs. Fluorescence channels are displayed as pseudocolors: cytokeratin (CK-PE) is shown in green and DAPI-stained nuclei in magenta. Optimization of DNA extraction from CellSearch fractions improves DNA yield and NGS performance In our previous work, we showed that ascites may be exploited for TP53 sequencing by NGS [11], a common mutational event in HGSOC. Here, we hypothesized that CellSearch-enriched AF-derived CTC fractions could also serve as a source of DNA for downstream analysis using an expanded NGS panel routinely employed in our clinical practice for solid tumor samples [18]. This approach could be particularly advantageous in cases where patients cannot undergo mutational analysis due to limited solid biopsy material, and/or when CTC number within the peritoneum is particularly low, such as in peritoneal washing performed for staging of early-stage HGSOC [12,19]. DNA extraction was initially performed using a semi-automated approach, consisting of manual lysis followed by purification with the Maxwell system (Promega). However, this method yielded very low DNA concentrations (<2 ng/μL), most likely due to interference between the ferrofluid used during CTC enrichment and the magnetic beads employed in the extraction process, which was particularly evident in samples with saturated or non-diluted cartridges. Consequently, sequencing performance was suboptimal, with mean coverage below 50–80× (Supplementary Table 1), which is insufficient for robust detection of somatic variants, as the minimum desired coverage is usually 500× and preferably >1000×. To improve DNA yield and quality, an alternative fully manual extraction protocol was implemented. This approach included manual processing and elution in a reduced volume (12 μL), allowing for increased DNA concentration. NGS analyses were then performed using a panel designed for solid tumor biopsies. The manual extraction protocol substantially improved DNA recovery from both saturated, undiluted ascites cartridges and cartridges diluted 1:2000 that yielded readable CellSearch results. (Supplementary Table 1). Detection of tumor-specific TP53 variants in ascites-derived CTCs Next, qualitative analysis of NGS data was performed on DNA isolated from ascites-derived CTC-enriched fractions of five HGSOC patients. TP53 was selected as the primary molecular endpoint because its alterations represent a defining molecular hallmark of HGSOC. Thus, the assessment of TP53 mutations provided a robust proof-of-concept to determine whether DNA recovered from CellSearch-enriched ascites fractions could recapitulate tumor-specific genomic alterations. Somatic TP53 mutations were detected in all analyzed samples, with TP53 Variant Allele Frequencies (VAF) ranging from 3.6% to 92% (Table 3). Given the high VAF observed in samples OVC-319 and OVC-335 (82% and 92%, respectively), we wanted to rule out the possibility that these two variants were germline mutations. Therefore, for these two samples, germline DNA was obtained from PB and sequenced with the same NGS panel. Both samples were found to be TP53 wild-type, confirming that the two TP53 variants were attributable to the somatic DNA isolated from CTCs. Notably, complete concordance was observed between TP53 alterations detected in the ascites-derived CTC-enriched fractions and matched tumor tissue biopsies in all five patients analyzed by NGS. In the two cases with high VAFs (OVC-319 and OVC-335; 82% and 92%, respectively), germline DNA obtained from PB was additionally analyzed using the same NGS panel and was TP53 wild-type, confirming the somatic origin of the variants detected in the CTC-enriched fractions. These findings demonstrate the feasibility of the proposed workflow for recovering tumor-derived DNA from post-CellSearch ascites fractions and support its ability to reproduce tumor-specific TP53 alterations.Table 3TP53 Mutations in ascites-derived CTCs.Table 3Patient CodeCTC enumeration in the cartridgep.c.VAF (%)Mean Coverage (reads)Mutation in Tissue BiopsyGerminalOVC-31455p.Arg342Terc.1024C > T14700p.Arg342Ter/OVC-315339p.Cys176Tyrc.527G > A482000p.Cys176Tyr/OVC-3182/c.559+1G > A3.6468c.559+1G > A/OVC-319Saturated Cartridge/c.560-1G > A82900c.560-1G > AWTOVC-3355890p.Arg273Hisc.818G > A921310p.Arg273HisWTSaturated Cartridgep.Arg273Hisc.818G > A96744CTC: circulating tumor cell; OVC: ovarian cancer; VAF: variant allele frequency; WT: wild-type Table 3TP53 Mutations in ascites-derived CTCs.Table 3Patient CodeCTC enumeration in the cartridgep.c.VAF (%)Mean Coverage (reads)Mutation in Tissue BiopsyGerminalOVC-31455p.Arg342Terc.1024C > T14700p.Arg342Ter/OVC-315339p.Cys176Tyrc.527G > A482000p.Cys176Tyr/OVC-3182/c.559+1G > A3.6468c.559+1G > A/OVC-319Saturated Cartridge/c.560-1G > A82900c.560-1G > AWTOVC-3355890p.Arg273Hisc.818G > A921310p.Arg273HisWTSaturated Cartridgep.Arg273Hisc.818G > A96744CTC: circulating tumor cell; OVC: ovarian cancer; VAF: variant allele frequency; WT: wild-type TP53 Mutations in ascites-derived CTCs. CTC: circulating tumor cell; OVC: ovarian cancer; VAF: variant allele frequency; WT: wild-type We further explored whether CTC abundance was associated with the analytical performance of downstream molecular profiling. CTC abundance showed a positive association with TP53 VAF (Spearman's ρ = 0.70) and with mean sequencing coverage (Spearman's ρ = 0.80; Fig. S1). The sample with the lowest CTC counts successfully analyzed by NGS contained two CTCs and showed a TP53 VAF of 3.6% and a mean coverage of 468X. These findings suggest that CTC burden may influence both the detectability of somatic variants and the overall quality of molecular profiling; however, given the very small sample size (n = 5), they should be considered exploratory and do not allow the establishment of a validated minimum CTC threshold for reliable genomic profiling. No alterations other than TP53 were observed in the other genes analyzed by the NGS panel. Ascites-derived CTC burden shows no significant association with clinicopathological features The estimated concentration of ascites-derived CTCs showed marked inter-patient variability. Exploratory analyses were performed to investigate potential associations between CTC burden and clinical parameters, including serum CA125 levels, peritoneal cancer index (PCI), overall survival (OS), and progression-free survival (PFS). No significant correlation was observed between CTC number per mL and CA125 levels (p = 0.70) or PCI (p = 0.87). Similarly, no significant association was found between CTC burden and OS (p = 0.65), or PFS (p = 0.97). To explore the potential prognostic impact of CTCs, patients were stratified into low- and high-CTC groups using the median CTC count (4875 CTC/mL) as the cutoff. Kaplan–Meier analysis revealed a trend toward shorter PFS in patients with high CTC burden compared with those with low CTC counts, with a median PFS of 4.7 and 8.2 months, respectively (log-rank p = 0.342) (Fig. S2). With regard to OS, no statistically significant difference was detected (log-rank p = 0.138), all observed death occurred in the high-CTC group, whereas no deaths were recorded among patients with low CTC counts during the available follow-up period. Overall, these exploratory findings suggest that elevated CTC abundance may be associated with a less favorable clinical outcome and earlier disease progression. However, given the limited sample size and number of events, these results should be considered exploratory and warrant validation in larger prospective cohorts. Overall, these findings suggest that ascites-derived CTC burden does not directly reflect tumour load as measured by CA125 or PCI. Although no statistically significant association with survival outcomes was observed, the shorter median PFS among patients with higher CTC concentrations may indicate a potential relationship between ascites-derived CTC burden and disease aggressiveness but still warrants further investigation in larger cohorts. Discussion CTCs have emerged as a valuable biomarker in several solid tumors; however, their clinical utility in HGSOC remains limited due to their low abundance in PB. In this context, AF represents a biologically relevant and potentially more informative compartment, as it directly reflects the peritoneal tumor microenvironment [20]. In the present exploratory pilot study, we demonstrated that AF from patients with HGSOC contains substantially higher numbers of CTCs compared with PB and may represent a valuable and underexploited source of tumor-derived material for molecular characterization. Our findings support the feasibility of isolating CTCs from AF using the CellSearch platform and, importantly, recovering DNA suitable for downstream NGS analysis from post-enumeration cartridges. The markedly higher abundance of CTCs observed in AF compared with PB is biologically consistent with the unique metastatic behavior of HGSOC and aligns with the findings reported by Tu and colleagues in esophageal cancer [21]. Unlike many epithelial malignancies that disseminate predominantly through hematogenous spread, HGSOC metastasizes mainly through a transcoelomic route, whereby tumor cells exfoliate from the primary lesion and survive in suspension within the peritoneal cavity before colonizing distant peritoneal surfaces. In this context, AF acts not only as a passive reservoir but also as a dynamic tumor-supportive microenvironment enriched in tumor cells, inflammatory mediators, stromal components, and extracellular vesicles. Therefore, rather than simply representing a more abundant source of tumor cells, AF provides a unique window into the peritoneal metastatic compartment, offering direct access to the biological processes underlying transcoelomic dissemination in HGSOC. The relatively low detection rate of PB CTCs in patients from our cohort (50%) is consistent with previous reports on CTC-based PB liquid biopsy in HGSOC [22,23] and the low PB CTC burden in positive patients further supports the challenges associated with blood-derived CTC, highlighting the biological relevance of ascites as an alternative liquid biopsy source. To the best of our knowledge, this is the first study directly comparing CTC enumeration between PB and AF in HGSOC using the CellSearch system, an FDA-cleared technology currently validated for blood-based CTC detection. Although originally designed for PB samples, the platform proved technically feasible also in ascites, enabling enrichment and visualization of tumor cells despite the markedly different cellular composition of this biological matrix. Our findings also highlight important methodological challenges, particularly the need for extensive sample dilution to avoid cartridge saturation and the limitations of automated counting in highly cellular specimens. However, we believe CellSearch enrichment may actually be even more valuable for detecting cancer cells within peritoneal washings performed as a standard staging procedure in the context of apparent early-stage HGSOC [12,19]. Although CellSearch has previously been coupled with downstream molecular analyses, including targeted NGS following additional enrichment strategies such as DEPArray-based single-cell isolation, its application to the molecular characterization of ascites-derived CTCs remains largely unexploredA major strength of this study is the demonstration that CellSearch-enriched ascites-derived CTC fractions can be integrated into a workflow combining CTC enumeration with downstream targeted NGS analysis. Importantly, molecular analyses revealed complete concordance between TP53 variants identified in ascites-derived CTC DNA and matched tumor tissue biopsies. Because TP53 mutations are a defining molecular hallmark of HGSOC, present in more than 95% of cases, this finding corroborates our previous data [11] and strongly supports the biological reliability of ascites-derived CTCs as a surrogate source of tumor genomic information. It should also be acknowledged that DNA recovered from post-CellSearch cartridges does not exclusively derive from CTCs. Despite EpCAM-based enrichment, residual leukocytes remain within the cartridge and are co-isolated during sample recovery. Therefore, downstream NGS was performed on a mixed cell population comprising both CTCs and background hematopoietic cells. Despite this non-tumor background, tumor-specific TP53 mutations were consistently detected and showed complete concordance with matched tissue samples, further supporting the robustness of the approach. Another aspect worth considering is the complementary role of ascites-derived CTCs and circulating tumor DNA (ctDNA) within the evolving landscape of liquid biopsy. While ctDNA provides a sensitive and minimally invasive approach for detecting tumor-specific genomic alterations, it is limited to cell-free nucleic acids and does not preserve the cellular context. In contrast, CTCs offer access to intact viable tumor cells, enabling not only genomic profiling but also potential downstream transcriptomic, proteomic, phenotypic, and functional analyses. This unique characteristic may provide additional insights into tumor heterogeneity, mechanisms of treatment resistance, and metastatic potential. Therefore, rather than representing competing approaches, ascites-derived CTCs and ctDNA should be regarded as complementary tools that may provide integrated molecular information for precision oncology in advanced ovarian cancer. In patients with malignant ascites, the abundance of tumor-derived material within the peritoneal cavity may further enhance the value of CTC-based analyses, making ascites a particularly attractive source for longitudinal molecular characterization and disease monitoring. From a translational perspective, these findings may have relevant clinical implications. Tissue availability remains a major challenge in advanced HGSOC, particularly in patients undergoing neoadjuvant treatment not suitable for diagnostic laparoscopy, or recurrent disease where tumor material may be limited or exhausted. In this scenario, ascites-derived CTCs, obtained for example from paracentesis, could represent a complementary and minimally invasive source of somatic DNA to support molecular profiling and potentially guide precision oncology approaches. Although our NGS panel focused primarily on hotspot mutations routinely used in clinical diagnostics for solid tumors, future studies should explore whether broader genomic analyses-including homologous recombination deficiency (HRD)-related alterations, resistance-associated mechanisms, or longitudinal clonal evolution-can also be reliably assessed using this approach. From a clinical perspective, no statistically significant associations were observed between ascites-derived CTC burden and clinicopathological parameters, including CA125, PCI, PFS, or OS. However, patients with higher CTC counts showed a trend toward earlier disease progression than those with lower CTC levels. Although this association did not reach statistical significance, likely due to the limited sample size and exploratory nature of this pilot study, the observed trend is consistent with previous reports linking increased blood-derived CTC burden to poorer clinical outcomes in ovarian cancer [24,25]. From a biological point of view, elevated CTC levels may reflect a greater propensity for tumor dissemination and disease aggressiveness, supporting their potential role as a prognostic biomarker [26]. Therefore, while these findings should be interpreted with caution, they reinforce the rationale for further investigating CTC enumeration in larger prospective cohorts. Several limitations should be acknowledged. First, this study included a relatively small number of patients from a single institution, limiting statistical power and generalizability. Second, molecular analyses were feasible only in a subset of samples, mainly because of technical constraints related to DNA recovery and sequencing quality. Although no false-negative TP53 results were observed among the five paired CTC/tissue samples successfully analyzed, the limited sample size does not allow a reliable estimation of the false-negative rate or the analytical sensitivity of the proposed approach. Larger studies, including a greater number of paired samples, will be required to formally assess diagnostic performance and determine the risk of false-negative molecular results. Third, the CellSearch platform was originally optimized for epithelial CTC detection in PB and may fail to identify cancer cells undergoing epithelial-to-mesenchymal transition, potentially underestimating tumor heterogeneity typical of HGSOC [27]. In addition, the analytical linearity range of the CellSearch system extends up to 1238 cells; samples exceeding this threshold may yield less reliable cell counts, potentially affecting the accuracy of CTC quantification in highly cellular ascites specimens CellSearch Circulating Endothelial Cell Kit [28]. Finally, the exploratory nature of this pilot feasibility study precludes definitive conclusions regarding the prognostic value of ascites-derived CTCs. Conclusion In conclusion, our study demonstrates that AF contains substantially higher numbers of CTCs than PB in HGSOC patients and represents a rich and biologically meaningful source of tumor-derived material. Furthermore, we demonstrate the feasibility of combining CellSearch-based enrichment of ascites-derived CTCs with downstream NGS profiling, with complete concordance of TP53 alterations between CTC-enriched fractions and matched tumor tissue in the five patients successfully analyzed. These findings support the potential role of ascites-derived CTCs as an alternative liquid biopsy platform for molecular characterization and possibly disease monitoring in advanced HGSOC, warranting validation in larger prospective studies. Authors' contributions Conceptualization: GR, DDB; Methodology: FG, TM, EC, BP, GG; Software: FG, TM, EC, BP; Validation: FG, TM, EC; Formal analysis: FG, GR Investigation: FM, ES, ADL; Resources: AMP, PDI, DDB, GR; Data Curation: FG, TM, FM; Writing - Original Draft: FG, TM, IK; Writing - Review & Editing: DDB, GR, IK, SA; Visualization: FG, GR Supervision: MF, DDB; GR; Project administration: DDB, GR; Funding acquisition: DDB, GR. Ethics declaration Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed. This study included organ or tissue donors. This study includes human biological material and consent was obtained by donors, or their next of kin or legal representatives, for use in this study and for publication of the article. The samples used in this research were not sourced from executed prisoners or prisoners of conscience. This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. This study was approved by the Area Vasta Emilia Centrale (AVEC). (Approval No. Institutional Review Board 788/2021/Sper/AOUBo) Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the authors used ChatGPT to support the creation of the graphical abstract and to assist with linguistic editing and sentence correction. After using this tool, the author(s) reviewed and edited the content as needed and take full responsibility for the content of the published article. Funding The work reported in this publication was supported by Fondazione Cassa di Risparmio di Bologna (CARISBO), 2025.0339 to GR. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.