Predict
TITLE: Machine Learning Prediction of Perinatal Sentinel Events
Sites:
Principal Investigators:
|
Summary: Perinatal sentinel events (PSEs), including placental abruption, cord prolapse, uterine rupture and shoulder dystocia, are rare but serious complications of pregnancy that often result in severe morbidity and mortality. Outcomes range from perinatal death to early brain injury to chronic conditions, such as cerebral palsy (CP). Prediction of these events remains challenging. Using data from the Maternal and Newborn Clinical Management System (MN-CMS), this study will identify and compare pregnancies complicated by PSEs with those without such events. These data will be used to explore maternal, fetal, and intrapartum risk factors, and with machine learning techniques to develop prediction models capable of identifying pregnancies at increased risk for PSEs. By identifying those at risk, this project aims to reduce preventable perinatal morbidity and mortality and improve long-term neonatal outcomes. |
AI4LIFE II
TITLE: Artificial Intelligence for Fetal Wellbeing II
Sites:
Principal Investigators:
|
Summary: Continuous cardiotocography (CTG) is the current standard for monitoring of the fetus in labour. It is recognised that there are significant differences in intra- and inter-observer rates of CTG interpretation, resulting in failure to diagnose the at-risk fetus. There is also a high false positive rate resulting is potentially unnecessary and often dangerous interventions. Artificial intelligence (AI) provides an exciting opportunity to analyse/ interpret CTG in an unbiased manner while reducing human error. To develop an AI-assisted decision support system, data from the maternal and neonatal electronic healthcare records will be extracted and paired with CTG data. This data will then be pseudo-anonymised and used for the development of an AI-assisted monitoring system. |
AI4NICU
TITLE: Intelligent physiological monitoring of preterm infants to predict brain injury associated with CP and with confirmed CP in the first days after birth
Sites:
Principal Investigators:
|
Summary: Cerebral palsy (CP) remains prevalent among very (28-31 weeks) and extremely (<28 weeks) preterm infants. Prematurity-related brain injuries (high-grade intraventricular haemorrhage, IVH or periventricular leukomalacia, PVL) develop usually during the first few days after birth and are a strong predictor of CP. Development of IVH or PVL has been linked to hemodynamic instability. Artificial intelligence (AI) provides an opportunity to detect patterns from continuous physiological monitoring data that are associated with prematurity-related brain injury. A dataset of very and extremely preterm infants will be collated from previous INFANT centre cohorts. An separate but aligned dataset will be obtained from the University of Graz. Continuous physiological monitoring data will be analysed with AI algorithms to identify the infants with the highest risk of brain injury. |
DETECT
Earlier Detection of Developmental Delays in At-Risk Babies
AIM-HIGH
TITLE: Assessing Intellectual and Motor outcomes in High-Risk infants
Sites:
Principal Investigators:
|
Summary: Cerebral palsy (CP) is the most common physical disability in childhood, affecting 1.6 per 1,000 children in high-income countries. It results from an early injury or abnormality of the motor pathways of the brain, causing permanent changes in motor development. Risk factors are multifactorial and span from the periconception period to 2 years of age. Many children also have co-existing disabilities, including learning difficulties (50%), feeding problems, epilepsy (25%), visual loss (10%) or hearing loss.Early screening currently relies on neonatal expertise, using general movements assessment (GMA) and standardised neurological examination. More information is needed on the pathways to motor dysfunction to develop quantifiable measures of early motor function. Objective measures such as EEG sleep maturation or biochemical profiles could help detect perturbations in normal motor trajectories. However, outcomes are often grouped together as “disability” (CP, intellectual disability or death), which is unhelpful for clinicians and families. Current pathways also focus on developmental milestones and lack detailed assessment of cognitive function.This project will examine motor, cognitive and executive outcomes separately, using targeted cohorts. It aims to provide robust diagnostic and prognostic information by 4 months after birth (corrected age for preterm infants), and ideally by NICU discharge. |