[Computerized intrapartum cardiotocography: Current evidence, controversies, and future directions]

Cardiotocography (CTG) is defined as the simultaneous recording of fetal heart rate and uterine contractions. It is used worldwide during labour to assess foetal well-being and to detect early signs of hypoxia. However, its interpretation relies mainly on visual assessment by clinicians based on standardized guidelines. This subjective evaluation is associated with substantial inter-observer variability, which continues to fuel debate regarding the actual effectiveness of cardiotocographic monitoring during the peripartum period. From the 1980s onward, researchers and clinicians began developing computerized systems to assist professionals in CTG interpretation. However, these decision-support tools have not yet convincingly demonstrated their ability to improve diagnostic performance or, more importantly, perinatal outcomes. Over the past decade, the availability of large clinical databases and the rise of artificial intelligence (AI) have enabled the development of models capable of analysing CTG signals and predicting the risk of foetal hypoxia. These systems have shown promising performance in identifying pathological situations when evaluated on retrospective cohorts. Nevertheless, several important challenges must be addressed before considering their integration into clinical practice. It is essential to develop and share large, multicentric, open databases that allow the training of robust, representative, and generalizable models. Furthermore, these systems must be embedded within a structured care protocol to ensure consistent and operational use in clinical settings.