School-aged Outcomes Post Hypoxic Ischaemic Encephalopathy

The study will recruit children with a history of HIE in the newborn period (cases) and healthy children with an unremarkable newborn period (controls) born between January 2013-December 2020. During this period several prospective cohort studies of term infants ≥36 weeks at risk of HIE were conducted in Cork University Maternity Hospital (CUMH) neonatal intensive care unit (NICU).

Infants enrolled in these previous studies had neonatal data collected, in addition to detailed neurological assessments, neurophysiological monitoring, neonatal MRIs, and standardised neuro-developmental follow-up at two years of age using the Bayley Scale of Infant Development (BSID). These children were not followed after the BSID was performed. Approval for these studies was given by the Clinical Research Ethics Committee of the Cork Teaching Hospitals (CREC), and informed consent obtained. Cases for the proposed study will be identified from those who agreed on the original consent form to be contacted for future related research.

For control recruitment, the study will be advertised on social media, the INFANT website and in the Paediatric outpatient area in the adjacent Paediatric department. Healthy children aged between 5 and 12 years of age will be invited to contact the research team to be part of the control group. To be eligible, they must have been born at ≥ 36 weeks gestation, now attending mainstream school, did not have a diagnosis of HIE or have been admitted to the neonatal unit in the first 12 hours of birth. A new ethics submission for the current study was approved by the local CREC.

Sample Size: Taking the mean population IQ as 100 with a standard deviation of 15, we calculated that 98 cases will be needed to detect a 6-point difference in composite IQ with an alpha of 0.05 and a power of 80%. The same number of controls will be recruited, giving a total number of 196 participants. This sample is powered to show a meaningful difference in IQ between populations from which the predictive models are built. Regarding the radiomic analysis proposed there is very limited data in neonatal populations, but the sample size proposed would represent the largest neonatal radiomic study to-date, and would also be comparable to recent adult studies with similar proposed methodology.

Study Procedures:

If the parents/guardians agree to participate, written informed consent and age-appropriate assent will be obtained.

There will be two study visits. The first will involve performing the cognitive assessment, administration of questionnaires, and a short play therapy session to introduce the child to the MRI using a mock MRI scanner and video of the MRI department and scanner. The assessment will take a standardised format, taking approximately 1.5 hours in total and breaks will be provided as necessary.

The MRI will be performed at the second visit. The scanning protocol will be approximately 19 minutes in duration, with four key MRI sequences being performed. No sedation or anesthesia will be provided for the MRI. Qualitative and quantitative analysis of the MRI will be performed.

The goal of this project is to develop a predictive neonatal MRI biomarker algorithm that combines radiomic features associated with evolution of brain injury on School-age MRI scans, and qualitative neonatal MRI assessment, to predict School-age cognitive outcome following Neonatal HIE.

The specific objectives are

1. To develop a Neonatal MRI machine learning model of radiomic features predictive of significant future evolutional changes evident on school-age MRI following HIE

   • Perform School-age MRI in both a large cohort of infants with HIE (with available neonatal MRI) and an age-matched control population (did not have neonatal MRI).
   • School-age MRI to be assessed for qualitative injury, and perform regional volumetric analysis.
   • Psychological assessment of cognitive function, executive function, and full-scale IQ in all grades of HIE (mild, moderate, and severe) and compare to an age-matched control population.
   • Correlation of School-age MRIs (Case and Control) with School age psychological outcome to identify key features associated with outcome.
   • In HIE cohort, following identification of key features predictive of cognitive outcome on School-age MRI (regions of interest[ROI]), to project these ROIs onto the Neonatal MRI to identify radiomic features predictive of evolution of brain imaging findings.
   • Develop machine learning algorithm of radiomic features on neonatal MRI predictive of significant school-age MRI changes
2. To develop MRI Only predictive model for School-age Outcome: To develop predictive algorithm combining Neonatal MRI radiomic model developed (objective 1) and qualitative neonatal MRI findings, to predict School-age Psychological outcome.
3. To develop MRI and Clinical model for School-age Outcome: To develop a predictive machine learning algorithm combining Clinical and sociodemographic factors, with our Neonatal MRI radiomic model, and qualitative neonatal MRI findings to predict School-age Psychological outcome.