Artificial Intelligence in Obstetrics and Prenatal Medicine: Current Evidence, Clinical Validation and Implementation Künstliche Intelligenz in der Geburtshilfe und Pränatalmedizin: aktuelle Evidenz, klinische Validierung und Implementierung
- 1Department of ObstetricsUniversity of Cologne, Faculty of Medicine and University Hospital CologneCologneGermany.
Abstract
Introduction
Artificial intelligence (AI) is increasingly entering obstetrics and prenatal medicine, particularly in prenatal ultrasound, fetal echocardiography, fetal magnetic resonance imaging, genomic screening, risk prediction and clinical decision support. The aim of this review was to summarize current evidence, recent developments and implementation requirements for AI-based applications in prenatal care.Materials and methods
A structured narrative review was performed with emphasis on peer-reviewed and clinically relevant literature published from 2021 to 2026, supplemented by landmark methodological, regulatory and reporting guidance. Priority was given to systematic reviews, diagnostic accuracy studies, external validation studies, prospective evaluations, randomized educational or clinical studies, regulatory documents and reporting standards relevant to AI in medicine.Results
The strongest current evidence supports AI-assisted standard-plane recognition, image-quality assessment and fetal biometry, followed by emerging applications in fetal cardiac screening, blind-sweep ultrasound, fetal MRI segmentation, placental phenotyping, prediction of fetal growth restriction, preterm birth and preeclampsia, and machine learning-enhanced cfDNA analysis. Generative and multimodal AI systems may support report drafting, guideline retrieval and education, but remain insufficiently validated for autonomous clinical decision-making.Discussion
AI has the potential to improve access, standardization, efficiency and quality assurance in prenatal medicine. However, most applications remain limited by retrospective design, enriched datasets, insufficient external validation, unclear calibration, limited fairness analyses and incomplete workflow evaluation. Clinical adoption should therefore follow a human-in-the-loop model with prospective validation, transparent reporting, regulatory oversight, bias monitoring and post-market surveillance. AI should augment rather than replace specialist expertise in obstetrics and fetal medicine.
Related
- Development and Implementation of Quality Standards in Psychosocial Services Entwicklung und Implementierung von Qualitätsstandards im Psychosozialen Dienst
- Potential of Artificial Intelligence to Estimate the Birthweight of Newborns Based on Ultrasound Data – Comparison between AI and Hadlock Formula Das Potenzial künstlicher Intelligenz zur Schätzung des Geburtsgewichts von Neugeborenen anhand von Ultraschalldaten – Vergleich zwischen künstlicher Intelligenz und der Hadlock-Formel
- Artificial intelligence-enabled mental health applications for adolescents: a scoping review of safety, effectiveness and clinical implementation
- Digital, Mobile and Artificial Intelligence‐Supported Interventions for Pressure Injury Prevention and Management: A Systematic Review of Clinical, Workflow and Implementation Outcomes
- Bridging the Artificial Intelligence Translation Gap in Perioperative Pulmonary and Respiratory Complications: A Practical Framework for Institution-Level Implementation
BibTeX
@article{recker2026artificial,
title = {{Artificial Intelligence in Obstetrics and Prenatal Medicine: Current Evidence, Clinical Validation and Implementation Künstliche Intelligenz in der Geburtshilfe und Pränatalmedizin: aktuelle Evidenz, klinische Validierung und Implementierung}},
author = {Florian Recker},
journal = {Geburtshilfe und Frauenheilkunde},
year = {2026},
month = aug,
doi = {10.1055/a-2920-1072},
url = {https://europepmc.org/article/PMC/PMC13754454},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Artificial Intelligence in Obstetrics and Prenatal Medicine: Current Evidence, Clinical Validation and Implementation Künstliche Intelligenz in der Geburtshilfe und Pränatalmedizin: aktuelle Evidenz, klinische Validierung und Implementierung
AU - Florian Recker
PY - 2026
DA - 2026/08/12
JO - Geburtshilfe und Frauenheilkunde
DO - 10.1055/a-2920-1072
UR - https://europepmc.org/article/PMC/PMC13754454
AB - Introduction Artificial intelligence (AI) is increasingly entering obstetrics and prenatal medicine, particularly in prenatal ultrasound, fetal echocardiography, fetal magnetic resonance imaging, genomic screening, risk prediction and clinical decision support. The aim of this review was to summarize current evidence, recent developments and implementation requirements for AI-based applications in prenatal care.Materials and methods A structured narrative review was performed with emphasis on peer-reviewed and clinically relevant literature published from 2021 to 2026, supplemented by landmark methodological, regulatory and reporting guidance. Priority was given to systematic reviews, diagnostic accuracy studies, external validation studies, prospective evaluations, randomized educational or clinical studies, regulatory documents and reporting standards relevant to AI in medicine.Results The strongest current evidence supports AI-assisted standard-plane recognition, image-quality assessment and fetal biometry, followed by emerging applications in fetal cardiac screening, blind-sweep ultrasound, fetal MRI segmentation, placental phenotyping, prediction of fetal growth restriction, preterm birth and preeclampsia, and machine learning-enhanced cfDNA analysis. Generative and multimodal AI systems may support report drafting, guideline retrieval and education, but remain insufficiently validated for autonomous clinical decision-making.Discussion AI has the potential to improve access, standardization, efficiency and quality assurance in prenatal medicine. However, most applications remain limited by retrospective design, enriched datasets, insufficient external validation, unclear calibration, limited fairness analyses and incomplete workflow evaluation. Clinical adoption should therefore follow a human-in-the-loop model with prospective validation, transparent reporting, regulatory oversight, bias monitoring and post-market surveillance. AI should augment rather than replace specialist expertise in obstetrics and fetal medicine.
SN - 0016-5751
SN - 1438-8804
ER -