Optimization of the optimal time point for NIPT detection in pregnant women with male fetuses based on FCM fuzzy clustering and TOPSIS

Abstract

In response to the clinical pain point of insufficient accuracy of the unified detection time point due to the difference in body mass index (BMI) of pregnant women in non-invasive prenatal testing (NIPT) for male fetuses, this paper proposes an optimization scheme for the optimal detection time point of NIPT combining FCM fuzzy clustering and TOPSIS methods. Firstly, the original NIPT data of 1082 male pregnant women with fetuses were preprocessed (abnormal samples of BMI and gestational weeks were cleaned, and gestational weeks were converted to continuous values), and valid samples were retained. FCM fuzzy clustering (fuzzy coefficient m=2) is adopted to achieve flexible grouping of BMI, solving the problem of misjudgment of sample boundaries in traditional hard clustering. Then, the "compliance rate - risk - cost" decision matrix is constructed through the TOPSIS method, the optimal detection time points of each group are screened, and the influence of detection errors on the results is analyzed. The results show that FCM can achieve reasonable grouping of BMI, and TOPSIS can effectively determine the optimal detection time point for each group. When the detection error is ≤5%, the stability of the grouping results is good, and the compliance rate of each group's detection meets the clinical needs. There is a correlation between BMI and the optimal detection time point. This scheme can match the NIPT testing needs of pregnant women with different BMI levels, providing reliable support for the timing planning of clinical NIPT testing.