Optimization of NIPT Testing Timing and Fetal Anomaly Determination Models

Abstract

This study explores optimization strategies for Non-Invasive Prenatal Testing (NIPT) timing and fetal anomaly detection through data-driven modeling. By analyzing over 1,600 NIPT cases, the research identifies maternal BMI and gestational age as major determinants of fetal fraction, a key factor influencing test accuracy. For male fetuses, Y chromosome concentration showed a positive correlation with gestational age (r=0.45) and a negative correlation with BMI (r=–0.32). An XGBoost regression model achieved robust performance (R²>0.7), highlighting weight and height as significant predictors. For female fetuses, a probabilistic model integrating Z-scores, GC content, read proportions, and BMI achieved >95% accuracy and <5% false-positive rate in detecting trisomies 21, 18, and 13. The study further proposes BMI-stratified testing windows—12, 14, and 16 weeks for low-, medium-, and high-BMI groups—to ensure sufficient fetal DNA concentration. These results emphasize the necessity of personalized NIPT timing and model-based optimization to reduce false negatives and enhance detection efficiency. Future integration of AI-driven prediction systems and hybrid cell-based testing may enable real-time, individualized prenatal screening with higher clinical applicability.