作者: Dana Dvorská , Dušan Braný , Bálint Nagy , Marián Grendár , Robert Poka
DOI: 10.3390/IJMS20174119
关键词:
摘要: Ovarian cancer is a highly heterogeneous disease and its formation affected by many epidemiological factors. It has typical lack of early signs symptoms, almost 70% ovarian cancers are diagnosed in advanced stages. Robust, non-invasive diagnosis will certainly be beneficial. Herein we analysed the regulatory sequence methylation profiles RASSF1, PTEN, CDH1 PAX1 tumour suppressor genes pyrosequencing healthy, benign malignant tissues, corresponding plasma samples. We recorded statistically significant higher levels (p < 0.05) tissues than controls (39.06 ± 18.78 versus 24.22 6.93; 13.55 10.65 5.73 2.19). Higher values gene were also found samples (22.25 14.13 46.42 20.91). A similar pattern with positive correlation between lesions was noted (r = 0.886, p 0.019) 0.771, 0.001). The random forest algorithm combining indices all four age determined 0.932 AUC (area under receiver operating characteristic (ROC) curve) prediction power model classifying controls. Our study results indicate effects changes development suggest that potential candidate for cancer.