Predicting the Birth Delivery Mode: Uncovering the Link Between EHG Topological Features and BMI

The appropriateness of selected features, contingent upon anthropometric factors, influences the efficacy of algorithm classification in pregnancy monitoring and obstetric outcome forecasting. Obstetricians often lack the technology to determine whether a cesarean delivery is necessary based on antepartum circumstances. The objective of this study is to develop an automated system for an effective detection of mode of birth delivery using EHG and analyze the data points shape structures across different BMI groups. Topological features (Area, Perimeter, Circularity, Variance, Bending Energy) were extracted from the boundaries of convex and concave hulls which applied to the whole signal, including the quiescent periods in this study. Pregnant women who underwent cesarean sections and those who delivered spontaneously were categorized into four BMI groups: unhealthy obese, healthy obese, overweight, and normal weight. EHG signals from the 31st week of gestation were examined. Topological characteristics derived from the border shape of the concave hull with a tolerance of 1.7 exhibits a marked statistically significant difference (p<0.005) between cesarean and spontaneous EHG signals. All of the features reached an accuracy above 90% in the classification using a random forest classifier between cesarean and vaginal deliveries across all BMI groups. On the other hand, it was found that aside from variance, women who delivered spontaneously exhibited higher topological feature values compared to those who underwent cesarean sections. The obtained results assist obstetricians in selecting the optimal birth delivery mode, hence minimizing unnecessary cesarean sections, complications, and emergency medical interventions.