Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

Predicting survival in extremely preterm infants: A multicenter machine learning study from Spain.

Paula Sol Ventura¹, Juan Leva-Bueno², Angélica Atehortúa²

¹Barcelona Institute for Global Health (ISGlobal), Barcelona, Spain; Neonatology Unit, Paediatric Department, Hospital Germans Trias i Pujol, Universitat Autónoma de Barcelona, Badalona, Spain; Paediatric Department, Hospital Universitario Arnau de Vilanova Lleida, Lleida, Spain.

Computers in Biology and Medicine

|June 13, 2026

Summary

Related Concept Videos

Survival Tree

Survival Tree

Survival trees are a non-parametric method used in survival analysis to model the relationship between a set of covariates and the time until an event of interest occurs, often referred to as the "time-to-event" or "survival time." This method is particularly useful when dealing with censored data, where the event has not occurred for some individuals by the end of the study period, or when the exact time of the event is unknown.
Building a Survival Tree
Constructing a survival tree begins...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Clinical outcome and safety of amnioreduction in singleton pregnancies with severe polyhydramnios.

Fetal diagnosis and therapy·2026

Same author

Cardiovascular risk and metabolomic profile in adolescents: A comparative study between anorexia nervosa, normal weight, and obesity.

Clinica e investigacion en arteriosclerosis : publicacion oficial de la Sociedad Espanola de Arteriosclerosis·2026

Same author

CineScribe: LLM-based detection and mitigation of ambiguity in cine cardiac magnetic resonance reports.

Computers in biology and medicine·2026

Same author

Rupture of membranes and short-term respiratory outcomes in extremely preterm infants: a multicenter retrospective cohort study.

Pediatric research·2026

Same author

ZeCardioAI: combining zebrafish, AI, and xAI for an in-depth cardiac phenotyping platform.

Briefings in bioinformatics·2026

Same author

Spanish 2026 guidelines for neonatal stabilization and resuscitation: analysis, adaptation, and position statement of the neonatal resuscitation Group of the Spanish Society of Neonatology.

Anales de pediatria·2026

This summary is machine-generated.

Interpretable machine learning models significantly improve survival prediction for extremely preterm infants compared to traditional scores. These advanced models identify key factors and infant subgroups, enabling earlier, targeted interventions for better outcomes.

Area of Science:

Neonatology
Machine Learning in Healthcare
Predictive Analytics

Background:

Predicting outcomes for extremely preterm infants is a critical challenge in neonatology.
Traditional scores like CRIB I and CRIB II have limitations in predictive accuracy.
Machine learning (ML) offers potential for more precise outcome prediction in this vulnerable population.

Purpose of the Study:

To develop and evaluate interpretable ML models for predicting survival to hospital discharge in extremely preterm infants.
To compare the performance of ML models against established scores (CRIB I, CRIB II).
To identify key factors influencing survival and understand infant subgroups.

Main Methods:

Trained multiple ML classifiers on data from 8080 extremely preterm infants (22-26 weeks' gestation).

Keywords:

Clinical decision support Delivery room management Extremely preterm infants Interpretable machine learning Neonatal outcome prediction

Related Experiment Videos

Used area under the ROC curve (ROC AUC) and precision-recall curve (PR AUC) for performance evaluation.

Employed SHAP (Shapley Additive Explanations) for model interpretability and UMAP for clustering survival profiles.

Main Results:

XGBoost model achieved superior performance (PR AUC: 0.81, ROC AUC: 0.77) over CRIB I and CRIB II.
Key predictors included gestational age, birth weight, admission temperature, nCPAP, 5-minute Apgar score, and antenatal corticosteroids.
SHAP-UMAP analysis identified two distinct infant subgroups with differing survival rates (76% and 52%).

Conclusions:

Interpretable ML models enhance survival prediction accuracy for extremely preterm infants beyond traditional methods.
These models offer clinical transparency and enable earlier identification of high- and low-risk infants.
Further validation and integration into clinical workflows are essential for real-world application.