e18075 Background: Ovarian cancer (OC) represents the most lethal gynecologic cancer. Late-stage diagnosis associated with a lack of effective markers to detect chemotherapy resistance contributes to high mortality and poor prognosis of epithelial ovarian cancer (EOC). OvarianTag Biomarkers Panel was developed as biopsy-based gene expression assay to discriminate prognosis in EOC patients, to understand the aggressiveness and relapse causes of these patients. Methods: Patients were admitted into this study in two phases. First, 45 women were included prospectively, and RNA was obtained from fresh normal ovarian (NO) tissues (n = 18), ovarian serous cystadenoma tumors (n = 11) and EOC (n = 16). The total RNA content was quantified and quantitative PCR (qRT-PCR) was performed to determine relative gene expression, starting set genes related to apoptosis and necroptosis regulated cell death pathway. Machine Learning algorithms were used to classify patient’s data associated and to characterize molecular targets as prognostic markers, named OvarianTag panel. Later, others 55 EOC patients, attended between 2008 and 2016, were included. Within this cohort, tissues were obtained from 106 FFPE samples and shared into primary and metastatic EOC, and from the same NO tissues (as control). RNA was extracted and qRT-PCR determined the OvarianTag panel expression level from each sample. Then, hierarchical clustering allowed us to classify the samples using the cut off obtained by classifier algorithm. Results: Significant differences in gene expression profile were observed between patients’ groups. Molecular classifier algorithm proved the prognostic value of genes evaluated. Decisions trees and biomarker clusters were constructed. In the full training, were obtained 90% of correct classifications and in cross-validation, 89%. In the second phase, the classifier reported specificity of 100% and sensibility of 79% to predict tumor recurrence and platinum resistance. Conclusions: OvarianTag panel is a putative prognosis biomarker of EOC subgroups and could be considered to drive the clinical practice of EOC treatment options, improving outcomes because providing an opportunity for guide clinical intervention for increasing survival rates. Besides of that, it will be allowed include patients in clinical trials of new target drugs or for repositioning of target drugs that have already been tested in humans.
Read full abstract