A team of researchers from the SABIEN group at the ITACA Institute of the Universitat Politècnica de València (UPV) has led an international study that comprehensively analyses the use and impact of artificial intelligence (AI) in Neonatal Intensive Care Units (NICUs). The work, published in Seminars in Fetal and Neonatal Medicine, reviews a decade of research in the field and evaluates 41 studies that apply AI to the neonatal clinical setting.

Developed in collaboration with the Hospital Universitari i Politècnic La Fe, the IIS La Fe and Queen Mary University of London, the research provides a detailed analysis of the development of these technologies, identifies technical, methodological and ethical challenges, and defines priorities for promoting their safe clinical implementation.

The analysis shows that AI is playing a significant role in improving early diagnosis, advanced monitoring and the prediction of complications in premature newborns or those with serious pathologies. Specifically, the most common applications are concentrated in areas such as the cardiovascular, neurological, respiratory and infectious systems, where algorithms can detect subtle clinical signs that may go unnoticed in conventional assessments.

'Among the most relevant advances are the early detection of sepsis, the estimation of brain maturation, the prediction of respiratory episodes and the optimisation of continuous monitoring systems based on physiological signals, video or sound,' explain the ITACA researchers.

In fact, they highlight that NICUs generate massive volumes of real-time data, so interpreting them using AI can 'improve clinical decision-making and anticipate complications, reducing risks and strengthening care for the most vulnerable newborns,' says Vicente Traver, head of the SABIEN group and co-author of the article.

The study identifies several obstacles that still limit the widespread adoption of these technologies, including differences between patients, unrepresentative data sets, biases associated with non-invasive monitoring, and the lack of external validation in much of the reviewed work.

It also highlights the need for models that encourage their use by healthcare professionals, as well as larger, high-quality databases to improve the robustness and generalisation of algorithms.

The research also emphasises the importance of multidisciplinary cooperation—between engineers, clinicians, data specialists, and ethics experts—to drive technological advancement in the neonatal-perinatal field.

'As computational models become more sophisticated and access to health data increases, the potential of AI to transform healthcare is considerable, ranging from predictive diagnosis to the development of personalised treatment plans,' says Antonio Martínez Millana, deputy director of ITACA and participant in the study.

However, key challenges remain, such as ensuring privacy, rigorously validating models and guaranteeing equitable access to emerging technologies: 'key areas for future research to build safe clinical environments and more accessible healthcare systems,' says the deputy director of ITACA.

Finally, the scientific work identifies numerous opportunities to move towards more predictive and efficient NICUs. These include the integration of multimodal data, the development of real-time monitoring tools, and the design of models that are explainable and ethically responsible. 'These technologies represent a decisive opportunity to evolve towards more preventive, accurate care tailored to the needs of each newborn,' says Andrea García Montaner, a researcher at SABIEN.

The research has been funded by the INBIO programme, which promotes innovation projects between the Hospital Universitari i Politècnic La Fe and the Universitat Politècnica de Valencia, and is supported by the RICORS-SAMID network and NextGenerationEU European funds, which support the improvement of biomedical research and innovation applied to neonatal health.

Antonio Martínez Millana, Álvaro Solaz-García, Andrea García Montaner, María Portolés Morales, Longwei Xiaod, Yan Sund, Vicente Traver, Máximo Vento, Pilar Sáenz-González; A systematic review on the use of artificial intelligence in the neonatal intensive care unit: far beyond the potential impact. Seminars in Fetal and Neonatal Medicine. DOI: https://www.sfnmjournal.com/article/S1744-165X(25)00084-8/abstract