| Objective To explore the diagnostic performance of serum lnc RNAs by comparing the differentiation between ovary cancers and controls;to establish the clinical base model,lnc RNAs model and base model plus lnc RNAs model using logistic regression analysis to further explore the early diagnosis power of serum lnc RNAs.MethodsThe specimens were collected from January 2013 to October 2014.Serum and clinical information were collected from 93 women before ovarian tumor resection.The univariable logistic regression models were used to explore the risk factors of ovary cancer from the clinical characteristics of age,family history,bearing history,age of menarche,serum CA125 and imaging which include type-B ultrasonic,computer tomography(CT),positron emission tomography and computer tomography(PET-CT)and magnetic resonance imaging(MRI).The diagnostic performance of these risk factors was assessed by the area under the receiver operating characteristic curve(ROC-AUC).The expression levels of tissue and serum lnc RNAs m RNA were detected by quantitative real time polymerase chain reaction(q RT-PCR).The lnc RNA copies were calculated by using standard curves.The univariable logistic regression models were used to explore the correlation between each lnc RNA and ovary cancer.The diagnostic power of the lnc RNAs to predict the tumor outcome was evaluated with AUC-ROC analysis.Z tests were used to compare the diagnostic performance between serum lnc RNA and the serum CA125 levels.The clinical base model which includes clinical risk factors,lnc RNAs model which includes multiple lnc RNAs and base model plus lnc RNAs model were established by multivariable logistic regression analysis.The diagnostic power of models to predict the tumor outcome was evaluated with AUC-ROC analysis.ResultsUnivariable logistic regression analysis found that age,serum CA125 and imaging were the independent risk factors to predict ovary cancer.The odds ratios(OR)(95% confidence interval(CI))of age,serum CA125 and imaging were 1.0977(1.0489-1.1488),1.0026(1.0012-1.0039)and 22.8667(7.5091-69.6335),respectively.The ROC – AUC of age,serum CA125 and imaging were 0.784,0.784 and 0.814,respectively.Based on this available clinical information of this cohort of patients,there is no enough evidence to demonstrate family history,bearing history and age of menarche as the independent risk factors to predict ovary cancer.The odds ratios(OR)(95% confidence interval(CI))of age,serum CA125 and imaging were 2.4490(0.8379-7.1574),2.7077(0.8959-8.1837)and 1.1702(0.8756-1.5640),respectively.The clinical base model which was established by multivariable logistic regression analysis got the AUC of 0.889(0.806-0.946)and predictive accuracy of 82.22%.The expression levels of lnc RNA H19,HOTAIR and PVT1 in ovary cancers are much higher than those of control samples.Univariable logistic regression analysis found that lnc RNA H19,HOTAIR and PVT1 were the independent risk factors to predict ovary cancer.The odds ratios(OR)(95% confidence interval(CI))of lnc RNA H19,HOTAIR and PVT1 were 1.0027(1.0000-1.0055),1.0117(1.0060-1.0175)and 1.0050(1.0020-1.0081),respectively.The ROC – AUC of H19,HOTAIR and PVT1 were 0.609,0.789 and 0.814,respectively.The lnc RNAs model which was established by multivariable logistic regression analysis got the AUC of 0.927(0.854-0.971)and predictive accuracy of 84.95%.Combined model with base model and lnc RNA model got a high diagnostic power with AUC-ROC of 0.970(0.912-0.994)and predictive accuracy of 89.13%.Further analysis revealed that lnc RNA model is much better than CA125 to diagnosing ovarian cancer(p=0.013).Conclusion Age,serum CA125 and imaging were the independent risk factors to predict ovary cancer.Serum lnc RNA H19,HOTAIR and PVT1 sufficed to differentiate ovary cancers from controls.Lnc RNA model can got a better diagnostic power that CA125 and could serve a noninvasive biomarkers for detecting ovarian cancer. |
PDF Full Text Request