BMI-Stratification-Driven Optimization of NIPT Timing and Fetal Abnormality Determination

Noninvasive prenatal testing (NIPT) analyzes cell-free fetal DNA fragments in maternal blood and faces two key challenges: precisely selecting the testing time point and efficiently determining fetal chromosomal abnormalities. Integrating generalized linear modeling and clustering, this study focuses on BMI-driven optimization of NIPT timing and the determination of fetal anomalies. First, we employ a generalized linear regression model with Pearson correlation analysis to quantitatively characterize the associations between fetal Y-chromosome concentration and indicators such as gestational age and BMI. The results indicate that the quadratic BMI term exhibits a pronounced nonlinear inhibitory effect;evaluation based on GLM metrics shows that the model stably captures the positive effect of gestational age and the overall suppressive effect of BMI. Next, this article combines K-means clustering with decision-tree regression to stratify maternal BMI among pregnancies with male fetuses, yielding a grouping scheme with within-cluster homogeneity of 92%, and construct a composite risk function R by integrating window-period risk and accuracy risk to determine the optimal time point. Finally, this article adopts the pipeline “fit a Gaussian mixture model and use the midpoint c of the two component means as a data-driven threshold→ multi-model cross-validation for model selection→ gestational-age grid computation→ use τ to obtain the earliest attainment time T*→ Pareto frontier trade-off, ”and, with K-means–based BMI stratification and τ sensitivity analysis, shows that the recommended time point exhibits notable robustness.