Relationship between white matter integrity and neurological function in preterm infants at 30 weeks postmenstrual age

J. George1, J. Fripp2, K. Shen, K. Pannek2, A. Chan2, R. Ware3, S. Rose2, P. Colditz4, R.N. Boyd1 1Queensland Cerebral Palsy and Rehabilitation Research Centre, School of Medicine, The University of Queensland, Brisbane, Australia. 2The Australia E-Health Research Centre, Brisbane, Australia.

3Children’s Health Research Centre and School of Population Health, The University of Queensland, Brisbane, Australia. 4Centre for Clinical Research, The University of Queensland, Brisbane, Australia

Objectives:

Results: Structural

Results: DTI

Our 30 week DTI measures were consistent with previous studies acquired on older cohorts (Rose 2014 for 34-38 weeks and Oishi 2011 for 37-53 weeks). In particular:

• Posterior aspects had higher FA as compared with anterior aspects (eg. See Posterior to Anterior Corona Radiata and internal capsule in Figure 3).

• Centrally located fibres had higher FA compared to peripheral tracts (eg. See PLIC to RLIC to PTR to SS in Figure 3).

• FA values lower and MD values higher than reported in previous near term data. However, this relates to PMA and the PMA adjusted trend is remarkably similar to Rose et al (Figure 4).

Participants: At 30 weeks PMA 71 infants born <31weeks were assessed: 44 male, mean gestational age (GA) at birth 28.2 wks (SD=1.7), mean birthweight (BW) 1106g (SD=312.1), mean PMA at MRI scan 32.17wks (SD=1.4).

This is a sub cohort of the full study (n=110), with inclusion criteria

• MRI acquired and passed quality control.

• MRI abnormality classification of normal or mild (Kidokoro 2013).

• General Movements classification of normal or poor repetoire.

Conclusions: In this preliminary study of very early brain development in infants born preterm with no evidence of brain injury: a) Our Structural and DTI measures were consistent with MRI studies acquired at near term equivalent age. b) A weak association was observed between bed-side neurological assessment and structural or diffusion MRI at 30 weeks PMA. Infants can be scanned safely early and advanced imaging with corresponding clinical tools has an important role to play in earlier detection of motor delay and infants at risk of a later diagnosis of Cerebral Palsy.

Acknowledgments: The University of Queensland

Cerebral Palsy Alliance Research Foundation (IRG1413)

Queensland Government (Smart State; Health Practitioner Stimulus Grant)

Financial Markets Foundation for Children (2014-074),

National Health and Medical Research council (CDF-2 103887 )

Methods: Infants underwent 3T MRI using an MRI compatible incubator without sedation at 30-32 weeks PMA with structural and diffusion MR imaging. The structural MRI was segmented in gray matter (GM), white matter (WM) and corticospinal fluid (CSF), and labelled using the Albert Atlas (Gousias 2014). A population specific Fractional Anisotropy (FA) atlas was created and aligned to the John Hopkins Unit (JHU) neonatal Atlas. Clinical assessment was performed within a week of MRI. It consisted of General Movements Assessment (GMA), Hammersmith Neonatal Neurological Examination (HNNE) and Premie-Neuro (PN). The relationship between regional volumetric (n=8) or diffusion measures (n=18) and clinical assessment were investigated using general linear models with adjustments for GA at birth, sex, BW, brain size and PMA at MR scanning with statistical significance accounting for multiple comparisons set at 0.002.

The aim of this prospective cohort study was to examine the relationship between very early brain structure (structural MRI), white matter integrity (diffusion MRI) and neurological function at 30 weeks post menstrual age (PMA) in infants born very preterm. In this study, we investigate this on a sub cohort of infants with no evidence of brain injury.

Results: Associations Response: MRI measures and 30 week Clinical Assessment

Figure 1: Example slice from the (top) structural and (bottom) FA Atlas created from population of 30 week PMA MRIs. Colourmap overlay and surface rendering illustrate consensus (top) GM segmentation and Albert atlas labelling (bottom) WM labels used from the JHU atlas labelling.

Figure 3: Average Fractional Anisotropy and Mean Diffusivity in each of the 18 JHU regions.

Figure 4: Scatter graph of early brain DTI values (mean PMA-at-scan=32.2 weeks) for (top) Fractional Anisotropy and (bottom) Mean Diffusivity in various tracts.

DTI Region Acronyms: anterior (ACR), superior (SCR), and posterior (PCR) corona radiata, (CC) corpus callosum, anterior (ALIC) and posterior (PLIC) internal capsule, Cerebral peduncle (CP), retrolenticular part of the internal capsule (RLIC), posterior thalamic radiation (PTR), sagittal striatum (SS), external capsule (EC), superior longitudinal fasciculus (SLF), fornix (Fx), stria terminalis (ST), and subcortical nuclei (Caudate, Thalamus, Putamen, Globus pallidus).

Figure 2: Scatter graph of brain volumetry (mean PMA-at-scan=32.2 weeks) for various regions, with confidence intervals.

Our 30 week structural MR measures appear consistent with previous studies (Zacharia 2006). Although motion and the resolution of the scans generally limited reduced the accuracy of the quantification.

• Total intracranial volume (TIV) was significantly associated with PMA-at-birth (p < 0.01), PMA-at-scan (p < 0.001), gender (p < 0.05).

• GA-at-birth was only weakly associated with cortical volumetry after adjusting for TIV, PMA-at-scan and gender.

Figure 5: Box plot of raw regional FA measures in each GMA group.

In our cohort, no association was found between thalamus or other volumetry and prematurity, or with corpus collosum DTI measures. A box plot showing FA measures in GMA groups is illustrated in Figure 5. No significant association between DTI regional measures and GMA was found. A weak association (which doesn’t survive multiple comparison correction) was observed between total HNNE and Cerebral Peduncle.

Figure 6: Scatter graph relating regional MD measures with 30 week total HNNE assessment.

Response: 40 week Clinical assessment Regional assessment did not find a significant association between 40 week clinical assessments and 30 week MRI measures.

Response: 30 week Clinical assessment Although a weak association can be seen in many regions between HNNE and DT measures (Figure 6). No significant associations remain after controlling for GA, PMA-at-scan, BW and gender and correcting for multiple comparisons. Weak non-significant (0.002 < p < 0.05) associations require further investigation (PCR, internal capsule, CP, PTR, SS and SLF) and MD (SCR).

References: 1. Gousias IS, et al. Magnetic Resonance Imaging of the Newborn Brain: Automatic Segmentation of Brain Images into 50 Anatomical Regions. PLoS ONE. 2013;8(4). 2. Kidokoro H, et al. New MR Imaging Assessment Tool to Define Brain Abnormalities in Very Preterm Infants at Term. AJNR Am J Neuroradiol. 201;34(11):2208-14. 3. Rose J, et al. Brain Microstructural Development at Near-Term Age in Very-Low-Birth-Weight Preterm Infants: An Atlas-Based Diffusion Imaging Study. NeuroImage 2014;86:244–56. 4. Zacharia A, et al. Early Assessment of Brain Maturation by MR Imaging Segmentation in Neonates and Premature Infants. AJNR Am J Neuroradiol 2006;27(5):972–77.

Contact: Joanne George: E: j.george2@uq.edu.au