Research on Individualized Decision-Making for Timing of Non-Invasive Prenatal Testing (NIPT)Based on Chance-Constrained Programming and Multifactorial Probabilistic Modeling

DOI: https://doi.org/10.62517/jes.202502413

Author(s)

Jiayi Wu1, Jiaqi Zhang1, Lu Li1, Jiangang Zhang2, Mingming Gong2

Affiliation(s)

1School of Artificial Intelligence and Big Data, Henan University of Technology, Zhengzhou, China 2iFlytek Co., Ltd., Hefei, Anhui, China

Abstract

To address the issues of test failure and elevated risk caused by insufficient fetal cell-free DNA proportions in non-invasive prenatal testing(NIPT)for pregnant women with high body mass index(BMI),this study designed a decision-making framework that integrates multifactorial probabilistic modeling and stochastic optimization to generate individualized precision testing protocols. First, multiple linear regression analysis of male fetus Y chromosome concentrations identified X chromosome concentration as the strongest predictive factor. Subsequently, a gestational age-dependent probability model for meeting target fetal DNA fraction thresholds was established, incorporating chance-constrained programming: optimizing optimal testing time points for high-BMI pregnant women by minimizing comprehensive risks under the constraint that achievement probability exceeds predefined confidence levels. Pregnant women were categorized into six risk groups, with recommended testing time windows ranging progressively from 16.0 weeks for low-risk groups to 20.0 weeks for high-risk groups. Finally, extending to abnormality detection in female fetuses, a two-stage methodology combining rule-based prioritization with logistic regression assistance was established through sample balancing and classifier comparison, improving diagnostic accuracy to 90.11%.This systematic approach provides a complete clinical solution for NIPT implementation, while its methodological framework demonstrates generalizability to similar decision-making scenarios involving uncertainty management.

Keywords

Non-Invasive Prenatal Testing; Chance-Constrained Programming; Probabilistic Modeling; Personalized Decision Making; Optimization of Testing Timing; Abnormality Determination

References

[1]Xue, Y., Zhao, G. D., Qiao, L. W., et al. Research progress on fetal cell-free DNA enrichment technology in maternal peripheral blood for NIPT. Chinese Journal of Birth Health&Heredity,2023,31(05):1087-1090.DOI:10.13404/j.cnki.cjbhh.2023.05.029. [2]Ju,J., Li,J., Liu,S., et al. Estimation of cell-free fetal DNA fraction from maternal plasma based on linkage disequilibrium information. npj Genomic Medicine,2021,6:85. [3]Guo,Z., Zhang,B., Yang,D., Wang,L., et al. Multidimensional roles of cfDNA fragmentomics in preeclampsia: from placental hypoxia and TLR9 inflammation to clinical risk stratification. Frontiers in Medicine,2025,12:1539651. [4]Liscovitch-Brauer, N., Mesika, R., Rabinowitz, T., et al. Machine learning-enhanced noninvasive prenatal testing of monogenic disorders. Prenatal Diagnosis, 2024,44(9):1024-1032.DOI:10.1002/pd.6570. [5]Zhang, J., Wu, Y., Chen, S., et al. Prospective prenatal cell-free DNA screening for genetic conditions of heterogenous etiologies. Nature Medicine, 2024, 30(2): 470-479. DOI:10.1038/s41591-023-02774-x. [6]Liu, W. Q., Yang, J. X., Zhang, J., et al. Technical standards consensus for high-throughput sequencing of fetal cell-free DNA in maternal peripheral blood plasma for pathogenic copy number variation screening. Chinese Journal of Medical Genetics, 2021, 38(7): 613-619. [7]Xu, C., Li, J., Chen, S., et al. Genetic deconvolution of fetal and maternal cell-free DNA in maternal plasma enables next-generation non-invasive prenatal screening. Cell Discovery, 2022,8(1): 109. DOI:10.1038/s41421-022-00457-4. [8]Pender, J., Ko, Y. M., & Xu, J. The number of overlapping customers in Erlang-A queues: an asymptotic approach. Probability in the Engineering and Informational Sciences,2025:1-26. [9]Lu, Y., Chen, Y., Ding, S., et al. Performance analysis of non-invasive prenatal testing for trisomy 13,18,and 21:A large-scale retrospective study(2018–2021).Heliyon,2024,10(13). [10]Chen, T., & Guestrin, C. XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016: 785-794.