Serial cerebral magnetic resonance imaging before and after birth in patients with complex congenital heart disease -...

Our objective was to establish a multicentre foetal-to-neonatal magnetic resonance (MR) neuroimaging programme for neonates undergoing surgery for complex congenital heart disease (CHD). We assessed structural and volumetric MRI findings at different timepoints in various types of CHD and evaluated neurodevelopmental outcomes at one year of age. Additionally, we analysed the feasibility, challenges and limitations of implementing this MR neuroimaging programme.

In this prospective, multicentre observational study, we examined brain development and growth in infants with complex CHD requiring neonatal cardiac surgery. Brain MRI was performed at multiple timepoints: at the 32nd week of gestation, after birth (both before and after stage I surgery), and before stage II surgery in single-ventricle CHD. We analysed and compared cardiac diagnoses, treatment approaches, structural and volumetric brain MRI findings and neurodevelopmental outcomes at one year of age (assessed using the Bayley III scale) with those of healthy controls.

Between April 2020 and September 2023, 81 patients and 15 healthy controls underwent at least one MRI. Cardiac diagnoses were biventricular CHD (66.7%), single-ventricle CHD (25.9%) and borderline left ventricle CHD (7.4%). New structural cerebral lesions were found before stage I or after stage I including white matter injury in 3.8% and 8.7%, respectively, ischaemic cerebral lesions in 11.5 and 11.6%, intraventricular haemorrhages in 7.7% and 7.2%, and subdural haemorrhages in 33.6% and 26.1%. Total brain volume at 32.6 (interquartile range [IQR]: 31.3-33.3) gestational weeks was 228.9 ml (213.1-241.2) in biventricular CHD, 194.4 ml (165.3-223.6) in single-ventricle CHD and 196.4 ml (186.4-235.2) in normal healthy controls. After birth, at 6 days (3-16) of life total brain volume was 337.1 ml (310.3-350.2) in biventricular CHD, 331.6 ml (305.9-350.7) in single-ventricle CHD and 406.8 ml (389.9-438.7) in normal healthy controls. After stage I, at 26.5 days (18.3-40.8) total brain volume was 367.7 ml (341.8-385.5) in biventricular CHD, 353.6 ml (338.2-375.7) in single-ventricle CHD and 514.1 ml (482.9-554.6) at 116 days (94.5-118.5) in patients with single-ventricle CHD. At 12.1 months of age, neurodevelopmental performance determined by the Bayley III scale (mean ± SD [standard deviation]) was lower for patients with single-ventricle CHD (cognitive composite score [CCS]: 92.9 ± 13.1; language composite score [LCS]: 88.5 ± 12.0; motor composite score [MCS]: 85.6 ± 14.5) than in patients with biventricular CHD (CCS: 101.2 ± 11.1; LCS: 95.7 ± 13.1; MCS: 87.6 ± 18.0) or in healthy controls (CCS: 113.3 ± 5.6; LCS: 102.3 ± 7.9; MCS: 100.7 ± 8.2). Feasibility for performing cerebral MRI was limited due to maternal/patient safety reasons and further logistical infrastructural reasons.

Structural cerebral lesions were found at various timepoints in both biventricular and single-ventricle CHD during foetal-to-neonatal serial cerebral MRI. Compared to healthy controls, total brain volume was reduced in patients with complex CHD, and neurodevelopmental outcomes at one year of age were mildly to moderately impaired. Several patient-related and infrastructural challenges limit the feasibility of a routine magnetic resonance neuroimaging programme, necessitating further efforts to optimise its implementation into routine clinical practice in the future.

ClinicalTrials.gov NCT04233775.