Circulating tumor-derived small extracellular vesicles micro RNAs plus CA-125 objectively tests residual tumor prior to debulking surgery in advanced ovarian cancer (525)

Abstract

<b>Objectives:</b> No residual disease after debulking surgery (R0 resection) is the most critical independent prognostic factor for advanced ovarian cancer (AOC). There is an unmet clinical need for selecting primary or interval debulking surgery in AOC patients using existing prediction models, such as CA-125, CT, PET-CT, or laparoscopy. Therefore, it is of paramount importance to preoperatively estimate the likelihood of R0 resection for choosing the best therapeutic strategy. In this study, we sought to develop a novel, non-invasive, and objective detection method. <b>Methods:</b> Differentially expressed microRNAs (miRNA) in circulating small extracellular vesicles (sEVs) between any residual disease (R0) and no residual disease (non-R0) AOC patients were revealed using small RNA sequencing. miRNA expression of plasma samples was measured by TaqMan qRT-PCR. The prediction model of residual disease risk was established via LASSO and logistic regression analysis based on the discovery-validation set. Tissue sEVs were extracted from primary tumor tissue, metastatic tumor tissue, adjacent tissue, or non-tumor tissue by differential ultracentrifugation and size exclusion chromatography (SEC). Cell-type-specific markers (EpCAM, FAP, CD45, CD31) in tumor sEVs pool were evaluated by western blot analysis. Plasma sEVs were captured by the magnetic bead sorting system (MACS) using specific surface proteins as markers (EpCAM, CD45, CD235a). <b>Results:</b> A total of 270 subjects were recruited in this study from four clinical centers. According to the comprehensive plasma sEVs miRNAs profile in AOC patients, we identified and optimized a risk model consisting of plasma sEVs 5-miRNA and CA-125 for predicting residual disease (AUC:0.912; sensitivity:0.964; specificity:0.884; PPV:0.911; NPV:0.905;). Moreover, there were significant differences in model index scores between AOC patients and patients with benign ovarian tumors, early-stage ovarian cancer with FIGO I or II, and advanced colorectal cancer. Our data showed that EpCAM-positive sEVs accounted for the highest in tumor sEVs pool. Besides, these 5-miRNAs were a high concentration horizontal packaging in plasma EpCAM-sEVs. By further comparing the sEVs 5-miRNA expression in paired tumor and non-tumor tissue as well as plasma, we found that they were mainly originated from epithelial ovarian cancer cells and their tumor microenvironment. These results indicated that plasma sEVs 5-miRNA in AOC patients was associated with residual disease. <b>Conclusions:</b> Collectively, our study demonstrated a distinct profile of circulating sEVs miRNAs between R0 and non-R0 patients and obtained a robust model of circulating tumor-derived sEVs 5-miRNA combined with CA-125 to preoperatively anticipate a high risk of residual disease for optimizing clinical treatment and guiding differential diagnosis. This prediction panel, as a tool of non-invasive liquid biopsy, can be part of a standard monitoring strategy for screening high-risk patients who are allowed for neoadjuvant chemotherapy.