Diffusion tensor MRI of the preterm brain

Despite many improvements in care, infants born prematurely remain at high risk of death and neurodevelopmental impairment. The aim of this project is to use diffusion tensor magnetic resonance imaging data in conjunction with computational analytical tools to investigate the development of the preterm brain.

 

Background

The immature developing brain is vulnerable to perinatal injury and preterm-born infants often suffer long-term morbidity that is more severe with prolonged premature exposure to the extrauterine environment (Bhutta et al., 2002). Despite well-documented clinical outcomes, the neuro-imaging correlates for developmental impairments in preterm infants are incompletely defined.
 
Diffusion tensor imaging (DTI) provides quantitative measures of diffusion in tissue. Within the brain, cell membranes, white matter tracts and macromolecules can restrict molecular motion, and result in reduced and/or anisotropic diffusion. Values of the apparent diffusion coefficient (ADC) and diffusion anisotropy can be determined from DTI and, by calculating the eigenvalues of the diffusion tensor, diffusion parallel and perpendicular to the white matter tracts can be measured. These non-subjective measurements can be used to assess micro-structural abnormalities in the preterm brain that are not evident on conventional MRI, and could provide information that ultimately leads to improvements in clinical outcome.

 

 

 

Study 1: Diffusion Tensor Imaging with Tract-Based Spatial Statistics Reveals Local White Matter Abnormalities in Preterm Infants

Tract-based spatial statistics (TBSS) is an automated observer independent approach for assessing groupwise microstructural differences in the major white matter pathways of the brain (Smith et al., 2006). The aim of this study was to determine if TBSS could be implemented in the preterm population, and to test the hypothesis that preterm infants have microstructural differences in cerebral white matter compared to term born control infants in the absence of focal abnormalities such as cystic periventricular leukomalacia (cPVL) or haemorrhagic parenchymal infarction (HPI) on conventional MR imaging.

 

Results

We found that regions within the centrum semiovale, frontal white matter and the genu of the corpus callosum had a significantly lower FA in preterm infants imaged at term equivalent age compared to term-born controls (voxelwise thresholding uncorrected for multiple comparisons, t > 3, p < 0.05) (Figure 1a-d). Those infants born at less than or equal to 28 weeks gestational age (n = 11) displayed additional reductions in FA in the posterior aspect of the posterior limb of the internal capsule, the external capsule and the isthmus and middle portion of the body of the corpus callosum, and had larger regions of reduced anisotropy within the centrum semiovale, frontal white matter and genu of the corpus callosum (Figure 1e-h).

 

Figure 1: The effect of preterm birth on FA at term equivalent age
Mean FA skeleton overlaid on the mean FA map. Regions of the mean FA skeleton in green represent areas where there were no significant differences in FA values in the preterm infants imaged at term compared to the term-born controls. Areas in blue are regions where the FA was significantly lower in the preterm group (a-d), and can be observed in the centrum semiovale (a), frontal white matter (b) and genu of the corpus callosum (c). Those infants born ≤ 28 weeks gestational age (e-h) had greater regions of reduced anisotropy within the centrum semiovale (e), frontal white matter (f) and genu of the corpus callosum (g), and displayed additional reductions in FA in the posterior aspect of the posterior limb of the internal capsule (g) and the external capsule (g).

 

In order to explore this reduction in FA, we analysed the three eigenvectors of the diffusion tensor. Regions which exhibited decreased FA showed elevated intermediate (λ2) and/or minor (λ3) eigenvalues in both preterm groups (voxelwise thresholding uncorrected for multiple comparisons, t > 3, p < 0.05) (Figures 2 and 3). The preterm groups also displayed an increase in the principal eigenvalue (λ1) in the regions of the frontal white matter and genu of the corpus callosum. A small number of isolated voxels showed higher FA and/or decreased λ2 or λ3 in the preterm infants (seen in red on Figure 1 and blue in Figures 2-3). However, these did not correspond to any well-defined brain region.

 

Figure 3: The effect of preterm birth on λ2 at term equivalent age
Mean FA skeleton overlaid on the mean λ2 map. Regions in green represent areas where there was no significant difference in λ2 values in the preterm infants imaged at term (a-d) and in the subset of infants born ≤ 28 weeks gestational age (e-h) compared to the term-born controls. Areas in red/orange represent regions where the λ2 was significantly higher in the preterm group, and can be observed in the centrum semiovale (a), frontal white matter (b) and genu of the corpus callosum (c). Those infants born ≤ 28 weeks gestational age (e-h) had greater regions of increased λ2 within the centrum semiovale (e), frontal white matter (f) and genu of the corpus callosum (h, i), and displayed additional increases in λ2 in the posterior aspect of the posterior limb of the internal capsule (g), external capsule (g) and the isthmus of the corpus callosum (h). 

 

Figure 4: The effect of preterm birth on λ3 at term equivalent age
Mean FA skeleton overlaid on the mean λ3 map. Regions in green represent areas where there was no significant difference in λ3 values in the preterm infants imaged at term (a-d) and in the subset of infants born ≤ 28 weeks gestational age (e-h) compared to the term-born controls. Areas in red/orange represent regions where the λ3 was significantly higher in the preterm group, and can be observed in the centrum semiovale (a), frontal white matter (b), and genu of the corpus callosum (c). Infants born ≤ 28 weeks gestational age (e-h) had greater regions of increased λ3 within the centrum semiovale (e), frontal white matter (f) and genu of the corpus callosum (g, h), and displayed additional increases in λ3 in the posterior aspect of the posterior limb of the internal capsule (g), external capsule (g) and the middle body and isthmus of the corpus callosum (h). 

 

Discussion and future work

In this study we used DTI and automated tract-based analysis to investigate brain microstructure in preterm infants imaged at term equivalent age. TBSS overcomes some of the limitations of region-of-interest and voxel-based morphometric approaches of analysing neonatal DTI data and offers the possibility of detecting differences non-subjectively in small numbers of subjects. Further studies with larger numbers of infants will allow assessment of the influence of clinical variables such as gender, nutrition and infection on white matter development in the preterm brain.

 

For more details, please refer to:
Anjari et al, Diffusion tensor imaging with tract-based spatial statistics reveals local white matter abnormalities in preterm infants, NeuroImage (2007).
Anjari DTI TBSS 2007  

 

 

 

Study 2: Thalamo-cortical connectivity in children born preterm mapped using probabilistic magnetic resonance tractography

Diffusion tensor imaging (DTI) tractography provides a potentially valuable tool to assess connectivity in vivo.  The aim of this study was to assess the feasibility of using a probabilistic tractography approach to study the major white matter tracts and to define the cortical connectivity map in the thalami of 2 year old infants who had been born preterm but had normal neurodevelopment and no evidence of focal white matter injury on MR imaging. Then, in a proof-of-concept experiment, we examined the detailed connectivity in one 2 year old ex preterm infant with severe cerebral white matter damage, predicting significant differences in the patterns of connectivity compared to the main study group.

 

Thalamo-cortical projections

Cortical and thalamic masks were generated on the segmented T2 weighted image and propagated onto the infant’s native DTI data (Figure 5). Thalamo-cortical connections were assessed for every voxel in the thalamic masks using connectivity-based seed classification, and the volume of connected seeds were determined.
 

Figure 5
Cortical masks displayed on the segmented T2 weighted image (key red = frontal/ temporal cortex, yellow = occipital/ parietal cortex, green = motor region and blue = somatosensory region).

 

Results

DTI

Figures 6 - 9 show connectivity distributions generated in white matter tracts in an infant with normal imaging.

 

Figure 6
(a) Cortico-spinal tracts and (b) optic radiations in an infant with normal imaging.

 

 

Figure 7
Superior longitudinal fasciculus in an infant with normal imaging.

 

Figure 8
(a) genu and (b) splenium of the corpus callosum in an infant with normal imaging.

 

 

Figure 9
Inferior fronto-occipital fasciculus in an infant with normal imaging.

 

Figure 10 shows thalamo-cortical projections in an infant with normal imaging.

 

Figure 10
Thalamo-cortical projections in an infant with normal imaging (key as in Figure 5).

 

Figures 11 show connectivity distributions generated in white matter tracts in an infant with a large unilateral porencephalic cyst. Diminished tract volume and FA was seen on the side ipsilateral to the cyst and we were not able to demonstrate the inferior fronto-occipital fasciculus in the right hemisphere in this infant.

 

Figure 11
(a) Cortico-spinal tracts and (b) optic radiations in an infant with a porencephalic cyst.

 

Diminished volumes of thalamo-cortical projections were demonstrated in the infant with lesions (figure 12) and we were not able to visualise connections between the thalamus and the somatosensory cortex on the side ipsilateral to the large porencephalic cyst.

 

Figure 12
Thalamo-cortical projections in an infant with a large porencephalic cyst (key as in Figure 5).

 

Discussion

Using probabilistic tractography, we were able to visualise and quantify connectivity distributions in a number of white matter tracts. This is the first time that the thalamic corticotopic map has been observed in children in vivo. The ability to visualise and quantify connections between thalamus and cortex offers the exciting possibility of studying normal development and aberrant neural connectivity in infants, and may further our understanding of the aetiology of the cognitive deficits associated with preterm birth.

 

For more details, please see:
Counsell et al, Thalamo-cortical connectivity in children born preterm mapped using probabilistic magnetic resonance tractography, NeuroImage (2007).
Counsell ThalamoCortical Prob Tract 2007