Publicação

Artificial intelligence on prenatal ultrasound: advantages and limitations

Ver documento

Detalhes bibliográficos
Resumo:Abstract The integration of artificial intelligence (AI) into prenatal ultrasound represents one of the most promising innovations in contemporary Fetal Medicine. This opinion article examines the main advantages and limitations of AI application in this context, highlighting advances in the automation of biometric measurements, reduction of clinicians’ cognitive workload, and diagnostic support for fetal anomalies - particularly cardiac and central nervous system malformations. The use of convolutional neural networks has shown high efficacy in the segmentation and detection of fetal structures, enhancing both efficiency and consistency in screening. However, several challenges remain, including the need for large and diverse datasets, technical constraints, ethical considerations, and difficulties in effective implementation within clinical practice. The widespread adoption of these technologies will depend on continued research, appropriate regulatory frameworks, and close collaboration between clinicians and engineers, ensuring safe, effective, and equitable integration across varied healthcare settings.
Autores principais:Cruz,Jader
Outros Autores:Guedes-Martins,Luís
Assunto:Artificial intelligence Prenatal ultrasound Fetal medicine Automated diagnosis
Ano:2025
País:Portugal
Tipo de documento:artigo
Tipo de acesso:acesso aberto
Instituição associada:Fundação para a Ciência e Tecnologia
Idioma:inglês
Origem:SciELO Portugal
Descrição
Resumo:Abstract The integration of artificial intelligence (AI) into prenatal ultrasound represents one of the most promising innovations in contemporary Fetal Medicine. This opinion article examines the main advantages and limitations of AI application in this context, highlighting advances in the automation of biometric measurements, reduction of clinicians’ cognitive workload, and diagnostic support for fetal anomalies - particularly cardiac and central nervous system malformations. The use of convolutional neural networks has shown high efficacy in the segmentation and detection of fetal structures, enhancing both efficiency and consistency in screening. However, several challenges remain, including the need for large and diverse datasets, technical constraints, ethical considerations, and difficulties in effective implementation within clinical practice. The widespread adoption of these technologies will depend on continued research, appropriate regulatory frameworks, and close collaboration between clinicians and engineers, ensuring safe, effective, and equitable integration across varied healthcare settings.