PTerm Classifier promising tool for Accurately Predicting Preterm Birth: Study Finds

To establish the model, Zhiwei Guo, Southern Medical University, Guangzhou, China, and colleagues conducted a comprehensive, multi-center study involving 2,590 pregnant women—518 with spontaneous preterm births and 2,072 with full-term deliveries—recruited from three independent hospitals. Whole-genome sequencing of plasma cfDNA was performed, focusing on promoter regions that regulate gene expression. Advanced machine learning techniques, including support vector machines and feature selection algorithms, were used to build the predictive model.

Key Findings:

• PTerm, the Promoter profiling classifier for preterm prediction, achieved the highest accuracy among all tested models with an AUC of 0.878 based on leave-one-out cross-validation.
• The classifier maintained strong predictive performance across three independent validation cohorts, with a consistent AUC of 0.849, highlighting its reliability across varied populations.
• A major benefit of PTerm is its compatibility with current non-invasive prenatal testing (NIPT) workflows, requiring no changes in procedure or additional cost.
• Its integration into routine prenatal screening could help identify high-risk pregnancies early, enabling timely and targeted medical interventions.

The researchers emphasized that incorporating tools like PTerm into clinical practice could enhance early risk assessment and help lower the global burden of preterm births. They hope future studies will further validate its clinical utility and encourage widespread adoption of cfDNA-based risk prediction models in obstetric care.

"PTerm showed strong predictive accuracy for identifying preterm birth risk. Moreover, it can be applied directly to existing non-invasive prenatal testing data without altering current procedures or incurring additional costs, making it a practical and scalable option for early screening in clinical settings," the authors concluded.

Reference:

Guo Z, Wang K, Huang X, Li K, Ouyang G, Yang X, et al. (2025) Genome-wide nucleosome footprints of plasma cfDNA predict preterm birth: A case-control study. PLoS Med 22(4): e1004571. https://doi.org/10.1371/journal.pmed.1004571