Background: Ovarian cancer (OC) remains the leading cause of gynaecological cancer mortality, with about thirty percent of patients presenting with ascites at initial diagnosis. While ascites fluid accumulation is a hallmark of advanced, relapsed or metastatic disease, most ascitic fluid is discarded as clinical waste after routine paracentesis. This represents a significant untapped resource for translational analysis from liquid biopsy. Despite this, the cellular landscape of OC-derived ascites remains poorly characterised at a single-cell level.
Method: We developed an advanced imaging pipeline to enable quantitative, multiplexed biomarker analysis of patient-derived ascites. Cells isolated from OC ascites were analysed using multiplexed immunofluorescence, quantifying 70 protein and phospho-protein markers per cell. Single-cell image quantification was coupled with machine learning to profile cell type composition, signalling diversity, and lineage-associated marker expression.
Results: Isolated ascitic cells were compatible with our multiplexed imaging methodology, enabling quantitative profiling of expression levels and subcellular localisation for 70 markers per cell. Single-cell statistical and machine learning analyses revealed extensive cell type heterogeneity within epithelial, immune, and stromal compartments. This deep phenotyping defined distinct cell types and states, indicating dynamic interactions within the ascites microenvironment.
Conclusion: This world-first approach establishes a scalable platform that unlocks the analytical potential of patient-derived ascites, transforming wasted material into a treasure trove for cancer research. Our work provides a path to characterise the cell populations and molecular mechanisms driving OC progression. This platform is useful for comprehensive studies on tumour heterogeneity, epithelial-mesenchymal transition, and the functional mapping of treatment-responsive cell states.