Development of a spontaneous preterm birth predictive model using a panel of serum protein biomarkers for early pregnant women: A nested case-control study

Shuang Liang; Yuling Chen; Tingting Jia; Ying Chang; Wen Li; Yongjun Piao; Xu Chen

doi:10.1002/ijgo.15876

Development of a spontaneous preterm birth predictive model using a panel of serum protein biomarkers for early pregnant women: A nested case-control study

Int J Gynaecol Obstet. 2024 Aug 27. doi: 10.1002/ijgo.15876. Online ahead of print.

Authors

Shuang Liang^{1

2}, Yuling Chen³, Tingting Jia¹, Ying Chang^{1

2}, Wen Li^{1

2}, Yongjun Piao^{1

2

4}, Xu Chen^{1

2

3}

Affiliations

¹ Tianjin Central Hospital of Gynecology Obstetrics/Nankai University Affiliated Maternity Hospital, Tianjin, China.
² Tianjin Key Laboratory of Human Development and Reproductive Regulation, Tianjin, China.
³ School of Life Sciences, Tsinghua University, Beijing, China.
⁴ School of Medicine, Nankai University, Tianjin, China.

PMID: 39189090
DOI: 10.1002/ijgo.15876

Abstract

Objective: To develop a model based on maternal serum liquid chromatography tandem mass spectrometry (LC-MS/MS) proteins to predict spontaneous preterm birth (sPTB).

Methods: This nested case-control study used the data from a cohort of 2053 women in China from July 1, 2018, to January 31, 2019. In total, 110 singleton pregnancies at 11-13⁺⁶ weeks of pregnancy were used for model development and internal validation. A total of 72 pregnancies at 20-32 weeks from an additional cohort of 2167 women were used to evaluate the scalability of the model. Maternal serum samples were analyzed by LC-MS/MS, and a predictive model was developed using machine learning algorithms.

Results: A novel predictive panel with four proteins, including soluble fms-like tyrosine kinase-1, matrix metalloproteinase 8, ceruloplasmin, and sex-hormone-binding globulin, was developed. The optimal model of logistic regression had an AUC of 0.934, with additional prediction of sPTB in second and third trimester (AUC = 0.868).

Conclusion: First-trimester modeling based on maternal serum LC-MS/MS identifies pregnant women at risk of sPTB, which may provide utility in identifying women at risk at an early stage of pregnancy before clinical presentation to allow for earlier intervention.

Keywords: biomarker; first trimester; liquid chromatography tandem mass spectrometry; machine learning; prediction model; preterm birth; soluble fms‐like tyrosine kinase‐1.