Automatic semantic segmentation of EHG recordings by deep learning: An approach to a screening tool for use in clinical practice.

BACKGROUND AND OBJECTIVE: Preterm delivery is an important factor in the disease burden of the newborn and infants worldwide. Electrohysterography (EHG) has become a promising technique for predicting this condition, thanks to its high degree of sensitivity. Despite the technological progress made in predicting preterm labor, its use in clinical practice is still limited, one of the main barriers being the lack of tools for automatic signal processing without expert supervision, i.e. automatic screening of motion and respiratory artifacts in EHG records. Our main objective was thus to design and validate an automatic system of segmenting and screening the physiological segments of uterine origin in EHG records for robust characterization of uterine myoelectric activity, predicting preterm labor and help to promote the transferability of the EHG technique to clinical practice. METHODS: For this, we combined 300 EHG recordings from the TPEHG DS database and 69 EHG recordings from our own database (Ci2B-La Fe) of women with singleton gestations. This dataset was used to train and evaluate U-Net, U-Net++, and U-Net 3+ for semantic segmentation of the physiological and artifacted segments of EHG signals. The model's predictions were then fine-tuned by post-processing. RESULTS: U-Net 3+ outperformed the other models, achieving an area under the ROC curve of 91.4 % and an average precision of 96.4 % in detecting physiological activity. Thresholds from 0.6 to 0.8 achieved precision from 93.7 % to 97.4 % and specificity from 81.7 % to 94.5 %, detecting high-quality physiological segments while maintaining a trade-off between recall and specificity. Post-processing improved the model's adaptability by fine-tuning both the physiological and corrupted segments, ensuring accurate artifact detection while maintaining physiological segment integrity in EHG signals. CONCLUSIONS: As automatic segmentation proved to be as effective as double-blind manual segmentation in predicting preterm labor, this automatic segmentation tool fills a crucial gap in the existing preterm delivery prediction system workflow by eliminating the need for double-blind segmentation by experts and facilitates the practical clinical use of EHG. This work potentially contributes to the early detection of authentic preterm labor women and will allow clinicians to design individual patient strategies for maternal health surveillance systems and predict adverse pregnancy outcomes.

Prediction of Cesarean Section for Intrapartum Fetal Compromise: A Multivariable Model from a Prospective Observational Approach.

OBJECTIVE: A cesarean section for intrapartum fetal compromise (IFC) is performed to avoid potential damage to the newborn. It is, therefore, crucial to develop an accurate prediction model that can anticipate, prior to labor, which fetus may be at risk of presenting this condition. MATERIAL AND METHODS: To calculate a prediction model for IFC, the clinical, epidemiological, and ultrasonographic variables of 538 patients admitted to the maternity of La Fe Hospital were studied and evaluated using univariable and multivariable logistic regression analysis, using the area under the curve (AUC) and the Akaike Information Criteria (AIC). RESULTS: In the univariable analysis, CPR MoM was the best single parameter for the prediction of CS for IFC (OR 0.043, p < 0.0001; AUC 0.72, p < 0.0001). Concerning the multivariable analysis, for the general population, the best prediction model (lower AIC) included the CPR multiples of the median (MoM), the maternal age, height, and parity, the smoking habits, and the type of labor onset (spontaneous or induction) (AUC 0.80, p < 0.0001). In contrast, for the pregnancies undergoing labor induction, the best prediction model included the CPR MoM, the maternal height and parity, and the smoking habits (AUC 0.80, p < 0.0001). None of the models included estimated fetal weight (EFW). CONCLUSIONS: CS for IFC can be moderately predicted prior to labor using maternal characteristics and CPR MoM. A validation study is pending to apply these models in daily clinical practice.

Prediction of Failure to Progress after Labor Induction: A Multivariable Model Using Pelvic Ultrasound and Clinical Data.

Objective: Labor induction is one of the leading causes of obstetric admission. This study aimed to create a simple model for predicting failure to progress after labor induction using pelvic ultrasound and clinical data. Material and Methods: A group of 387 singleton pregnant women at term with unruptured amniotic membranes admitted for labor induction were included in an observational prospective study. Clinical and ultrasonographic variables were collected at admission prior to the onset of contractions, and labor data were collected after delivery. Multivariable logistic regression analysis was applied to create several models to predict cesarean section due to failure to progress. Afterward, the most accurate and reproducible model was selected according to the lowest Akaike Information Criteria (AIC) with a high area under the curve (AUC). Results: Plausible parameters for explaining failure to progress were initially obtained from univariable analysis. With them, several multivariable analyses were evaluated. Those parameters with the highest reproducibility included maternal age (p < 0.05), parity (p < 0.0001), fetal gender (p < 0.05), EFW centile (p < 0.01), cervical length (p < 0.01), and posterior occiput position (p < 0.001), but the angle of descent was disregarded. This model obtained an AIC of 318.3 and an AUC of 0.81 (95% CI 0.76-0.86, p < 0.0001) with detection rates of 24% and 37% for FPRs of 5% and 10%. Conclusions: A simplified clinical and sonographic model may guide the management of pregnancies undergoing labor induction, favoring individualized patient management.

Cerebroplacental Ratio as a Predictive Factor of Emergency Cesarean Sections for Intrapartum Fetal Compromise: A Systematic Review.

Background: This systematic review aimed to clarify the association between the cerebroplacental ratio (CPR) and emergency cesarean sections (CSs) due to intrapartum fetal compromise (IFC). Methods: Datasets of PubMed, ScienceDirect, CENTRAL, Embase, and Google Scholar were searched for studies published up to January 2024 regarding the relationship between the CPR and the rate of CS for IFC, as well as the predictive value of the CPR. Results: The search identified 582 articles, of which 16 observational studies were finally included, most of them with a prospective design. A total of 14,823 patients were involved. A low CPR was associated with a higher risk of CS for IFC. The predictive value of the CPR was very different among the studies due to substantial heterogeneity regarding the group of patients included and the time interval from CPR evaluation to delivery. Conclusions: A low CPR is associated with a higher risk of CS for IFC, although with a poor predictive value. The CPR could be calculated prior to labor in all patients to stratify the risk of CS due to IFC.