Search

Primary platinum-based chemoresistance occurs in ~30% of patients with serous ovarian cancer. Chemoresistance is the main cause of disease recurrence, and accurate predictors to identify these patients with chemoresistance are required. Alternative splicing (AS) is a post-transcriptional modification process that is altered in cancer. A possible association between AS and chemoresistance is unclear and needs to be studied comprehensively in ovarian cancer. In the present study, RNA-sequencing data and clinical information for 320 patients with ovarian serous cystadenocarcinoma (OV) were downloaded from The Cancer Genome Atlas (TCGA) database. Splicing events were determined using the TCGA SpliceSeq tool. Seven types of AS events were identified. Univariate and multivariate logistic analyses were performed, and predictive models for OV chemoresistance were established, as well as a splicing network. A total of 22,036 AS events were identified in 7,404 genes, with 915 AS events detected in 677 genes that were significantly associated with chemoresistance in patients with OV. A receiver operating characteristic (ROC) curve was constructed for resistance predictive models composed of the most significant AS events. The area under the ROC curve was 0.931, indicating strong and efficient prediction of chemoresistance. Additionally, the high-risk score was associated with shorter overall survival. The splicing correlation network suggested a potential role of splicing factors in chemoresistance. In summary, the present study created a powerful predictor for primary platinum-based chemoresistance in patients with OV, identified splicing networks that could be involved in potential mechanisms of chemoresistance and provided potential targets to overcome chemoresistance. [ABSTRACT FROM AUTHOR]