Related Experiment Video
Updated: Jul 7, 2026
08:07
Electromyometrial Imaging of Uterine Contractions in Pregnant Women
Published on: May 26, 2023
Cervical Dilation Classification from Electrohysterography and Clinical Features: A Machine-Learning-Derived Digital
Otniel Portillo-Rodríguez1,2, Jorge Escalante-Gaytán2, Oscar Osvaldo Sandoval-González3
1Facultad de Ingeniería, Universidad Autónoma del Estado de México (UAEMéx), Toluca, Mexico.
Digital Biomarkers
|July 6, 2026
Summary
A new digital biomarker using electrohysterography (EHG) accurately tracks cervical dilation during labor. This noninvasive method combines EHG signals with maternal and gestational age for precise labor progression monitoring.
Area of Science:
- Biomedical Engineering
- Obstetrics & Gynecology
- Signal Processing
Background:
- Repeated digital cervical examinations during labor carry risks of discomfort and infection.
- Noninvasive monitoring of labor progression is needed to improve patient outcomes and clinical efficiency.
Purpose of the Study:
- To develop and validate an electrohysterography (EHG)-based digital biomarker for noninvasive cervical dilation tracking.
- To assess the model's accuracy in classifying labor stages using objective physiological signals.
Main Methods:
- Analyzed 72 single-channel EHG recordings from low-risk labor cases.
- Extracted 21 EHG descriptors and combined them with maternal age, gestational age, and contraction counts (LC, HC) as predictors.
- Developed a Genetic Algorithm Ensemble Bagged Tree (GA-EBT) model for feature selection and classification of dilation stages (low, moderate, advanced).
Main Results:
- The GA-EBT model identified a four-feature subset (maternal age, gestational age, LC count, HC count) for optimal classification.
- Achieved perfect performance (F1, recall, precision, specificity, accuracy of 1.000) on an independent test set.
- The best cross-validated bagged tree ensemble model showed a median macro-F1 score of 0.898 with 17 predictors.
Conclusions:
- An EHG-derived digital biomarker accurately classifies cervical dilation stages using minimal clinical and EHG-derived features.
- This noninvasive approach shows potential for real-time intrapartum monitoring, reducing the need for repeated digital examinations.
- The pilot-stage classification demonstrates high performance, supporting its potential clinical application in labor management.
