Abstract
Premature birth occurs during a period of rapid brain growth. In this context, interpreting clinical neuroimaging can be complicated by the typical changes in brain contrast, size and gyrification occurring in the background to any pathology. To model and describe this evolving background in brain shape and contrast, we used a Bayesian regression technique, Gaussian process regression, adapted to multiple correlated outputs. Using MRI, we simultaneously estimated brain tissue intensity on T1- and T2-weighted scans as well as local tissue shape in a large cohort of 408 neonates scanned cross-sectionally across the perinatal period. The resulting model provided a continuous estimate of brain shape and intensity, appropriate to age at scan, degree of prematurity and sex. Next, we investigated the clinical utility of this model to detect focal white matter injury. In individual neonates, we calculated deviations of a neonate's observed MRI from that predicted by the model to detect punctate white matter lesions with very good accuracy (area under the curve > 0.95). To investigate longitudinal consistency of the model, we calculated model deviations in 46 neonates who were scanned on a second occasion. These infants' voxelwise deviations from the model could be used to identify them from the other 408 images in 83% (T2-weighted) and 76% (T1-weighted) of cases, indicating an anatomical fingerprint. Our approach provides accurate estimates of non-linear changes in brain tissue intensity and shape with clear potential for radiological use.
Highlights
Neuroimaging during the perinatal period is both practically and technically challenging (Lodygensky and Thompson, 2017)
From this we derive a family of multimodal 4D growth curves, providing statistical measures of variation across the cohort. From this we show that individuals develop along trajectories defined by these growth curves and have a multimodal brain imaging fingerprint that persists with increasing chronological age; that focal abnormalities such as punctate white matter lesions are reflected by deviations from these typical trajectories at a voxel level in individual infants, allowing accurate automated detection of lesions; and that the global effects such as premature birth can lead to detectable deviations in global and local brain morphology, which can be quantified by datadriven approaches
We did not exclude neonates with radiologically-reported punctate white matter lesions (PWML), small subependymal cysts or small haemorrhages in the caudothalamic notch, as these are a common finding in the preterm population in particular
Summary
Neuroimaging during the perinatal period is both practically and technically challenging (Lodygensky and Thompson, 2017). Over a very short period, the brain changes in size and shape, tissue contrast changes, and transient developmental structures disappear (Kostovicand Jovanov-Milosevic, 2006). When investigating perinatal brain injury this evolving background represents a substantial hurdle, as imaging changes themselves can be both spatially and temporally heterogeneous (Rutherford et al, 2006) Because of this complexity, studies have shown inter- and intra-rater reliability in interpreting neonatal MRI to be moderate to low (Morel et al, 2016). Studies have shown inter- and intra-rater reliability in interpreting neonatal MRI to be moderate to low (Morel et al, 2016) This is especially the case with age-related image intensity or shape changes that may indicate dysmaturation, such as diffuse white matter injury, small punctate white matter lesions or ventricular dilation. Myelination may be disrupted by brain injury and prematurity and the degree of disruption can be dependent on both the actual pathology and the age of insult (Volpe, 2009)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
