Roc Curve Multivariate Modeling in Detection of Fetal Abnormalities Using Screening Down Syndrome Associated Markers in the First and Second Periods of Pregnancy

Early diagnosis of prenatal illnesses and timely treatment of congenital anomalies has been the mainstay of the health system. In this study, our aim is to provide Roc curve multivariate modeling in detection of fetal abnormalities using associated markers in screening Down syndrome in the first and second periods of pregnancy.
This is a descriptive-analytical study that uses information from two sets of data. In the first set, 152 individuals, who had the results of the first- trimester and second screening tests at risk and in the second group, 75 individuals with normal results. The studied variables included the serum markers in the first- trimester and the second- trimester screening, auxiliary variables (includes demographic information). Statistical analysis was performed by using ROC regression, incremental value analysis and Stata 12 software.
In evaluating the value of each diagnostic test in the presence of auxiliary variables using logistic regression and rock curves, the results generally showed that in screening the first- trimester of PAPP-A and in the screening the second-trimester,Inhibin-A can be used alone as a diagnostic test.
Best diagnostic test in the first- trimester, respectively, PPAP-A, NT, FREE B-HCG and in the second- trimester of screening, respectively, Inhibin-A, HGG, UE3 and AFP were based on the area under the ROC curve. In addition, the most significant effect of the predictor variable on the outcome of the diagnostic test was family history.