NEONATAL SOCIETY ABSTRACTS
Abnormal microstructural development of the cortex in congenital heart disease is related to impaired cerebral oxygenation
Presented at the Neonatal Society 2018 Spring Meeting (programme).
Kelly C1, Christiaens D1, Batalle D1, Makropoulos A2, Cordero-Grande L1, Steinweg J1, O'Muircheartaigh J1, Khan H3, Lee G3, Victor S1, Alexander D4, Zhang H4, Simpson J5, Hajnal J1, Edwards D1, Rutherford M1, Counsell S1
1 Centre for the Developing Brain, King's College London, St. Thomas' Hospital
2 Biomedical Image Analysis Group, Imperial College London
3 Neonatal Intensive Care Unit, St Thomas Hospital
4 Centre for Medical Image Computing, University College London
5 Paediatric Cardiology Department, Evelina London Children's Hospital
Background: Neurodevelopmental impairment has become a major remaining challenge in congenital heart disease (CHD) (1). Macroscopic structural abnormalities of the cortex in CHD have been reported previously (2,3). although the biological substrate remains unclear (4). We hypothesised that reduced cerebral oxygen delivery (CDO2) in CHD is associated with impaired cortical microstructure. We predicted that infants with CHD would exhibit higher cortical fractional anisotropy and lower orientation dispersion index vs. healthy matched controls, and that infants with the lowest CDO2 would demonstrate the more severe impairment.
Methods: The analysis included 96 newborn infants: 48 with confirmed complex CHD scanned prior to surgery, and 48 age-matched healthy infants. Imaging was performed at 3T and included T2w, T1w, and diffusion-weighted (DWI). Phase contrast angiography (PCA) was performed in the CHD group. T2w images were motion corrected, segmented, surface reconstructed, and registered to a group template. DWI data were reconstructed and motion-corrected, and neurite orientation dispersion and density imaging (NODDI) metrics were calculated. CDO2 was calculated using PCA measurements. Statistical analysis was performed using a modified version of FSL TBSS (5) for cortical analysis, using family-wise error correction for multiple comparisons with threshold-free cluster enhancement (6). Project approved by NRES West London (CHD: 07/H0707/105, Controls: 14/LO/1169) and informed written parental consent was obtained prior to imaging. Study funded by BHF, MRC, EU, NIHR.
Results: There were no significant differences in gestational age (GA) at birth, GA at scan, and sex between groups. The median GA at scan was 39.1 weeks (IQR 38.6 39.7) for both groups. T1w, T2w and DWI were acquired in all subjects. PCA was acquired with acceptable quality in 81% of those with CHD. Cortical microstructural development was abnormal in infants with CHD, with higher fractional anisotropy (FA) and lower orientation dispersion index (ODI) compared to healthy age-matched controls, correcting for gestational age at birth and scan. There were no differences in mean diffusivity or neurite density index. Secondly, we demonstrated that reduced cortical ODI in CHD is related to impaired cerebral oxygen delivery, supporting the hypothesis that chronic suboptimal cerebral oxygenation is associated with atypical cortical maturation. Thirdly, we showed that macrostructural gyrification of the cortex occurs in tandem with microstructural properties including increasing ODI and reducing FA.
Conclusion: These results support the interpretation that the primary component of cortical dysmaturation in infants with CHD is impaired development of dendritic arborisation, associated with reduced cerebral oxygen delivery, which is responsible, at least in part, for macrostructural findings in this population. Cortical ODI may prove to be a valuable early indicator of the success of future fetal intervention studies that aim to restore the faltering trajectory of cortical development in CHD.
Corresponding author: firstname.lastname@example.org
1. Gaynor et al. Pediatrics 135, 816825 (2015)
2. Clouchoux C et al. Cereb. Cortex 23, 29322943 (2013)
3. Limperopoulos C et al. Circulation 121, 2633 (2010)
4. Morton PD, Circ. Res. 120, 960977 (2017)
5. Smith SM, Neuroimage 31, 1487505 (2006)
6. Ball G et al., PNAS, 110, 95416 (2013).