A Novel Predictive Machine Learning Model Integrating Cytokines in Cervical-Vaginal Mucus Increases the Prediction
Hector Borboa-Olivares1,2, Maria Jose Rodríguez-Sibaja3, Aurora Espejel-Nuñez4
1Community Interventions Research Branch, Instituto Nacional de Perinatología Isidro Espinosa de los Reyes, Mexico City 11000, Mexico.
Insights
Predicting preterm birth (PB) is improved by including cytokine levels. This study shows a new model incorporating cytokines significantly enhances PB detection rates, aiding early intervention strategies.
Area of Science:
- Reproductive Medicine
- Immunology
- Biomedical Data Science
Background:
- Preterm birth (PB) is a major cause of infant mortality and morbidity.
- Current prediction methods, primarily cervical length measurement, have limited detection rates (~70%).
- Cytokine-mediated inflammation is implicated in PB pathophysiology, but not clinically utilized for prediction.
Purpose of the Study:
- To investigate the role of specific cytokines in cervical-vaginal mucus as predictors of preterm birth.
- To develop and evaluate machine learning models for improved preterm birth prediction incorporating cytokine levels.
Main Methods:
- Quantified cytokines (IL-2, IL-6, IFN-γ, IL-4, IL-10, IL-1ra) in cervical-vaginal mucus from pregnant women (18-23.6 weeks gestation).
- Collected clinical obstetric data to stratify risk for PB based on cervical length.
- Developed Random Forest models for PB prediction using clinical variables and cytokine data.
Main Results:
- Elevated IL-2, IL-6, IFN-γ, IL-4, and IL-10, with lower IL-1ra, were observed in the high-risk PB group.
- An adjusted model using maternal age, IL-2, and cervical length achieved an 87% detection rate, outperforming the gold standard.
- The cytokine-inclusive model demonstrated improved performance metrics: higher detection rate (87% vs. 66%), lower false positive rate (3.33% vs. 12%), and lower false negative rate (6.66% vs. 28%).
Conclusions:
- Cytokine profiling in cervical-vaginal mucus can significantly enhance preterm birth prediction accuracy.
- Incorporating cytokines like IL-2 into predictive models, alongside clinical factors, offers a more effective approach to identifying at-risk pregnancies.
- This improved prediction accuracy may facilitate timely preventive interventions for preterm birth.
Abstract:
Preterm birth (PB) is a leading cause of perinatal morbidity and mortality. PB prediction is performed by measuring cervical length, with a detection rate of around 70%. Although it is known that a cytokine-mediated inflammatory process is involved in the pathophysiology of PB, none screening method implemented in clinical practice includes cytokine levels as a predictor variable. Here, we quantified cytokines in cervical-vaginal mucus of pregnant women (18-23.6 weeks of gestation) with high or low risk for PB determined by cervical length, also collecting relevant obstetric information. IL-2, IL-6, IFN-γ, IL-4, and IL-10 were significantly higher in the high-risk group, while IL-1ra was lower. Two different models for PB prediction were created using the Random Forest machine-learning algorithm: a full model with 12 clinical variables and cytokine values and the adjusted model, including the most relevant variables-maternal age, IL-2, and cervical length- (detection rate 66 vs. 87%, false positive rate 12 vs. 3.33%, false negative rate 28 vs. 6.66%, and area under the curve 0.722 vs. 0.875, respectively). The adjusted model that incorporate cytokines showed a detection rate eight points higher than the gold standard calculator, which may allow us to identify the risk PB risk more accurately and implement strategies for preventive interventions.
