Placenta-specific maternal plasma proteomic biomarkers of fetal death - Tarca, A.L., 2022 The fetal mortality rate in the USA is 5.73 per 1,000 births, and there are significant disparities among racial and ethnic groups. While factors such as complications of the placenta, umbilical cord, congenital anomalies, and maternal conditions may be involved, one-third of fetal deaths are classified as unexplained. Our preliminary data suggest that novel placenta-specific proteins offer more discriminating power for fetal death cases compared to current maternal plasma biomarkers. However, the question of whether detection of such protein dysregulation in samples collected earlier in gestation from asymptomatic patients will improve prediction of fetal death compared to known biomarkers remains to be answered. The studies proposed in this R21 application are important because reliable identification of women at risk for fetal death is needed to enable use of preventive strategies such as pravastatin, a drug that prevents recurrent fetal death and placenta and maternal vascular diseases. Therefore, we will profile most promising maternal plasma protein candidates in samples collected weeks before fetal death diagnosis and in gestational age- matched controls. Samples were identified in the Wayne State University Perinatal Biobank. Using machine learning and predictive analytics pipelines as we have previously described, we will generate a placenta-specific plasma protein signature that is predictive of fetal death in asymptomatic patients. Multi-protein machine learning models, such as elastic net and random forest, will be developed and tested using cross-validation to ensure generalizability. Furthermore, we will assess the specificity of the fetal death proteomic model in an independent existing longitudinal dataset that includes data from normal pregnancies and cases of preeclampsia and small-for-gestational-age neonates. Finally, we will explore the molecular subclassification of fetal death cases using clustering algorithms based on plasma protein dysregulation prior to fetal death. The goal of such analysis is to generate additional hypotheses and tackle the heterogeneity of the fetal death phenotype, which has historically hindered the identification and validation of biomarkers for this syndrome.
