Machine Learning Algorithms Combining Slope Deceleration and Fetal Heart Rate Features to Predict Acidemia

Q: What is the study trying to predict using machine learning?

The study predicts neonatal acidemia during labor using features extracted from the fetal heart signal captured by electronic fetal monitoring (EFM).

Q: How is acidemia defined in the study?

Acidemia is defined as an umbilical arterial blood pH below 7.10.

Q: Which model performed best and how did logistic regression compare?

The random forest model achieved the highest AUC of 0.971, while logistic regression showed the best specificity (0.879) at a sensitivity of 0.95 and demonstrated clinical utility with a 31% cutoff.

Aug 4, 2026

Upload by Anda

Electronic fetal monitoring (EFM) serves as the standard method for intrapartum assessment of fetal well-being, yet visual interpretation and parameter categorization can yield limited accuracy. This study uses machine learning to predict neonatal acidemia from fetal heart signal features extracted in a 30-minute window, emphasizing the last deceleration closest to delivery. A case-control cohort of 502 infants (1:1) defines acidemia as umbilical arterial pH < 7.10 and compares logistic regression, classification trees, random forest, and neural networks using discrimination, calibration, and clinical utility validation.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages22

Reading Time55 min

AccessFree to download

Introduction

Electronic fetal monitoring and its limitations

Machine learning approaches for acidemia prediction

Study objective and rationale

applied sciences Article Machine Learning Algorithms Combining Slope Deceleration and Fetal Heart Rate Features to Predict Acidemia Luis Mariano Esteban 1, *, Berta Cast¡n 2, Javier Esteban-Escaño 3, Gerardo Sanz-Enguita 4, Antonio R. Laliena 5, Ana Cristina Lou-Mercad² 6, Marta Châliz-Ezquerro 7, Sergio Cast¡n 8 and Ricardo Savirân-Cornudella 9 Citation: Esteban, L.M.; Castán, B.; Esteban-Escaño, J.; Sanz-Enguita, G.; Laliena, A.R.; Lou-Mercadé, A.C.; Chóliz-Ezquerro, M.; Castán, S.; Savirón-Cornudella, R. Machine Learning Algorithms Combining Slope Deceleration and Fetal Heart Rate Features to Predict Acidemia. Appl. Sci. 2023, 13, 7478. [https://](https://)[ ](https://)[doi.org/10.3390/app13137478](doi.org/10.3390/app13137478) Academic Editor: Jan Egger Received: 18 May 2023 Revised: 18 June 2023 Accepted: 23 June 2023 Published: 25 June 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 Departament of Applied Mathematics, Escuela Universitaria Polit²cnica de La Almunia, Institute for Biocomputation and Physic of Complex Systems, Universidad de Zaragoza, 50100 La Almunia de Doña Godina, Spain 2 Department of Obstetrics and Gynecology, San Pedro Hospital, 26006 Logroño, Spain; [bcastan@riojasalud.es](bcastan@riojasalud.es) 3 Department of Electronic Engineering and Communications, Escuela Universitaria Polit²cnica de La Almunia, Universidad de Zaragoza, 50100 La Almunia de Doña Godina, Spain; javeste@unizar.es 4 Department of Applied Physics, Escuela Universitaria Polit²cnica de La Almunia, Universidad de Zaragoza, 50100 La Almunia de Doña Godina, Spain; [cherraldin@unizar.es](cherraldin@unizar.es) 5 Departament of Applied Mathematics, Escuela Universitaria Polit²cnica de La Almunia, Universidad de Zaragoza, 50100 La Almunia de Doña Godina, Spain; [arlalibi@unizar.es](arlalibi@unizar.es) 6 Department of Obstetrics and Gynecology, Lozano Blesa University Hospital, 50009 Zaragoza, Spain; [dra.aclou@gmail.com](dra.aclou@gmail.com) 7 Department of Obstetrics, Dexeus University Hospital, 08028 Barcelona, Spain; [martacholiz@gmail.com](martacholiz@gmail.com) 8 Department of Obstetrics and Gynecology, Miguel Servet University Hospital, 50009 Zaragoza, Spain; [scastan@salud.aragon.es](scastan@salud.aragon.es) 9 Department of Obstetrics and Gynecology, Hospital Cl½nico San Carlos and Instituto de Investigaciân Sanitaria San Carlos (IdISSC), Universidad Complutense, Calle del Prof Mart½n Lagos s/n, 28040 Madrid, Spain; [rsaviron@gmail.com](rsaviron@gmail.com) * Correspondence: lmeste@unizar.es Abstract: Electronic fetal monitoring (EFM) is widely used in intrapartum care as the standard method for monitoring fetal well-being. Our objective was to employ machine learning algorithms to predict acidemia by analyzing speciﬁc features extracted from the fetal heart signal within a 30 min window, with a focus on the last deceleration occurring closest to delivery. To achieve this, we conducted a case–control study involving 502 infants born at Miguel Servet University Hospital in Spain, maintaining a 1:1 ratio between cases and controls. Neonatal acidemia was deﬁned as a pH level below 7.10 in the umbilical arterial blood. We constructed logistic regression, classiﬁcation trees, random forest, and neural network models by combining EFM features to predict acidemia. Model validation included assessments of discrimination, calibration, and clinical utility. Our ﬁndings revealed that the random forest model achieved the highest area under the receiver characteristic curve (AUC) of 0.971, but logistic regression had the best speciﬁcity, 0.879, for a sensitivity of 0.95 . In terms of clinical

FAQs

What is the study trying to predict using machine learning?

The study predicts neonatal acidemia during labor using features extracted from the fetal heart signal captured by electronic fetal monitoring (EFM).

How is acidemia defined in the study?

Acidemia is defined as an umbilical arterial blood pH below 7.10.

Which model performed best and how did logistic regression compare?

The random forest model achieved the highest AUC of 0.971, while logistic regression showed the best specificity (0.879) at a sensitivity of 0.95 and demonstrated clinical utility with a 31% cutoff.

Aug 4, 2026

Upload by Anda