Purpose

The brainstem is often involved in hindbrain anomalies and several different segmentation disorders can be distinguished. Prenatal characterization of brainstem pathologies is a difficult task by means of ultrasound and MRI, since a fundamental knowledge of anatomy, imaging patterns and genetics is necessary. In paediatric and adult MR imaging it is known that the relation of the midbrain, the pons and the medulla oblongata should hold a ratio of about 1:2:1 [1]. However, little is known about the normal proportions of fetal brainstem segments and their relative changes throughout pregnancy. This study aimed to assess midsagittal 2D linear measurements and the area of brainstem substructures using fetal MRI and to provide anatomical reference data.

Methods and Materials

Fetal MRI cases between 15 and 40 GW with normal brain development (as visualized by prenatal MRI and ultrasound) and sufficient quality of a midsagittal T2-weighted image were retrospectively assessed. The midbrain, pons, medulla oblongata and the cerebellar vermis were linearly measured (height and anteroposterior diameter) and two-dimensionally segmented using the software ITK snap. The ratios of the brainstem substructures were calculated and correlated to gestational age.

Results

161 brain-normal fetal MRI with a mean age of 25.7±5.4 gestational weeks (GW) were included. The ratio midbrain:pons decreased significantly (p<0.01) from 1.6 to 0.5 whereas the ratio pons:medulla oblongata increased significantly (p<0.01) from 0.9 to 1.8 between 18 and 40 GW. The ratio midbrain:medulla oblongata decreased significantly (p<0.01) from 1.7 to 1.0. The ratio between Midbrain, pons and medulla oblongata changed from 2:1:1 early in gestation (prior to 22 GW) to the ratio of 1:2:1 (after 22 GW), which is the known normal value in adults. The one-dimensional measurements (especially the height of brainstem substructures) showed the same pattern during gestation.

Conclusion

Two-dimensional segmentation of brainstem substructures revealed a relative increase in the size of pons compared to midbrain and medulla oblongata during the 2nd and 3rd trimester. Brainstem proportions at the early second trimester significantly differ from those postnatally. These physiological changes of fetal brainstem proportions have to be acknowledged in the prenatal assessment of fetal hindbrain malformations and especially in the analysis and detection of brainstem segmentation disorders.

Reference

[1] Doherty, D., Millen, K. J., & Barkovich, A. J. (2013). Midbrain and hindbrain malformations: advances in clinical diagnosis, imaging, and genetics. The Lancet Neurology, 12(4), 381–393.

BACKGROUND AND PURPOSE: Research into memory deficits associated with hypoxic-ischemic encephalopathy has typically focused on the hippocampus, but there is emerging evidence that the medial diencephalon may also be compromised. Recently, we published a single center experience with focus on abnormal signal intensity of the mammillary bodies in perinatal asphyxia. Our new goal is to determine whether this abnormal mammillary signal, which could be a biomarker for subsequent mammillary body atrophy and memory impairment, is a universal finding across centres,

MATERIALS AND METHODS: Brain MRIs were retrospectively reviewed in full-term infants with hypoxic-ischemic encephalopathy (HIE) born in three tertiary NICUs between 2004 and 2019. Those born from 2008 onwards received hypothermia for 72 hours. Two centers used a 1.5 or 3 Tesla Philips with T2-weighted 2mm slice thickness, the third center 2-4 mm slice thickness. In two centers a pediatric neuroradiologist with two dedicated neuro-neonatologists independently assessed the mammillary bodies for abnormal signal on T1 and T2-weighted and DWI sequences. Grading was 0 for normal signal, grade 1 for abnormal signal without swelling and grade 2 for abnormal signal and swelling. In the third tertiary NICU center two pediatric neuroradiologists retrospectively reviewed 104 brain MRIs of neonates with HIE (age range 32-41w) acquired between 2005 and 2019. Patients with cardiac or brain vascular malformations, stroke and metabolic conditions or born extremely preterm were excluded.

RESULTS: MRIs of 426 infants with HIE were reviewed, 235 in center one and 87 in center two and 104 in center three. In center one 70 were not cooled. In center 2 and 3 all received hypothermia. MRIs were performed within 10 days of birth. In center one 31 neonates (13.2%), had abnormal high mammillary body signal on T2-weighted and low signal on T1-weighted sequences, 4 with mild, 25 with moderate, and 2 with severe hypoxic-ischemic encephalopathy. In addition, restricted diffusion was seen in 6 neonates who had MR imaging between days 5 and 7. For these 31 neonates, the most common MR imaging pattern (41.9%) was abnormal signal restricted to the mammillary bodies with the rest of the brain appearing normal. Follow-up MRIs were available for 9 patients: 8 acquired between 3 and 19 months and 1 acquired at 7.5 years. There was mammillary body atrophy in 8 of the 9 follow-up MRIs.

In the second center 87 infants were enrolled and 31 (34%) had abnormal high mammillary body signal on T2-weighted and low signal on T1-weighted sequences. No follow-up scans were performed in this center.

In the third center 34 cases (33%) of the neonates showed abnormal signal changes (4 cases with restricted diffusion). Follow-up scans were performed in 18/34 subjects (range 1mth-6yrs) and showed MB atrophy in 8 cases.

CONCLUSIONS:  abnormal high mammillary body signal on T1- and T2-weighted images is commonly seen during the acute phase in full-term infants with HIE. Given initial abnormal signal can progress to mammillary body atrophy, this could be a significant contributor to memory impairments in survivors of HIE.

Introduction: Magnetic resonance imaging (MRI) of the fetal brain is enabling identification and evaluation of congenital diseases of the central nervous system such as open spinal dysraphism (spina bifida, SB). In order to quantitatively analyse these images, automatic segmentation techniques for pathological brains are required, as manual fetal brain segmentation is challenging and time consuming. Here, we present an automatic segmentation neural network that is able to segment 3 distinct groups of fetal brains: fetal SB brains prior to fetal lesion repair, fetal SB brains after fetal lesion repair, and non-pathological fetal brains.

Methods: Fetal MRI was acquired on 1.5T and 3T whole-body scanners using a T2-weighted single-shot fast spin echo sequence (SSFSE) for 15 patients with SB both before and after fetal spinal lesion closure surgery, as well for 15 controls with normal neurodevelopment. We reconstructed multiple orthogonal SSFSE images of the fetal brain into high resolution 3D images using the total variation super-resolution (SR) reconstruction algorithm. Images were manually segmented into six different brain tissues: cerebrospinal fluid (CSF), grey matter (GM), white matter (WM), ventricles, cerebellum, and brainstem. Three images from each group were reserved as a validation set. The remaining 36 images were used to train the neural network. A 2D U-Net was chosen, with slices fed into the network in each orientation (axial, sagittal, coronal), resulting in three label maps which are then combined through voting. An Adam optimizer with dice coefficient loss was used with a learning rate of 1E-5, 50 epochs, a batch size of 16, and cross-validation.

Results: Examples of resulting segmentations can be seen in Figure 1. The automatic segmentations of the validation images were compared to the manually segmented labels, and the dice coefficient scores for each label were calculated (Table 1). The networks were able to most accurately identify the labels in the post-operative SB group and the controls, as the delineation between structures in the reconstruction was clearer.

Discussion and Conclusion: With a relatively small number of hand annotated samples, a 2D U-Net can be trained to automatically segment multiple types of high resolution fetal brains, in line with results from segmentation networks for normal fetal brains. Implementing data augmentation methods could be applied in the future to improve the dice coefficient scores. By including data sets from three different populations (pre-operative SB, post-operative SB, controls), we have been able to create a neural network able to automatically segment the three populations, which can be used in future studies to quantitatively analyse the fetal brain.

Introduction

When the brain is affected by pathology during early development, it attempts to compensate using its unique neural plasticity, particularly in children with unilateral cerebral palsy (uCP), in whom the largely intact dominant hemisphere can take over the motor control of the impaired hand.

Previous research showed significant relationships between motor outcome of the impaired hand and structural connectivity of the cortical spinal tract (CST), but also the sensory and interhemispheric tracts.

However, the relation between whole brain connectivity and hand function still remains unknown in uCP. Especially considering the high variability in lesion extent and location, each patient’s brain might exhibit a unique compensatory pattern beyond the previously investigated tracts. Here, we explored subject-specific global remodeling of the structural connectome in children with uCP and its relation to clinical outcome.

Methods

We recruited 55 children with spastic unilateral CP, aged 5-16y (Median=10y9m). For this analysis,  30 children were excluded due to: not meeting inclusion criteria (n=9), incomplete data (n=4) or having unconnected nodes in the resulting connectome (n=17).

Following images were acquired on a  3T Philips Ingenia CX MR with a 32-channel phased-array head coil: diffusion-weighted MRI (TR/TE=7800/90ms, Resolution=2.5x2.5x2.5mm, b-values=0/700/100/2800, directions=10/25/40/75), resting-state functional MRI (rs-fMRI) (T2*-weighted EPI, TE/TR=33/1700ms, resolution=3.59x3.59x4mm, 250 volumes),  T1-weighted MRI  (MPRAGE, resolution=.98x.98x1.2mm, TR/TE=9.6/4.6 ms) and T2-FLAIR (turbo-spin echo, resolution=.71x.71x1.2mm, TE/TR/TI=415/4800/1650ms). Additionally, we gathered descriptive and clinical characteristics on age, gender and lesion type, i.e. predominantly white matter (pWM, N=15) or grey matter (pGM, N=10) lesions. Daily life manual ability was assessed using the Manual Ability Classification System (MACS; I=9,II=9,III=7) and bimanual performance was assessed using the Assisting Hand Assessment (AHA). CST wiring patterns, innervating the impaired hand, were identified using transcranial magnetic stimulation, i.e ipsilateral (N=7), bilateral (N=8) or contralateral (N=7), with N=3 not assessed.

All images were analysed using a customized pipeline, detailed in figure 1. Rs-fMRI was utilized for subject-specific parcellations of the cerebrum into 50 regions using the Ncut algorithm, appended with 4 brainstem regions from the JHU atlas and left/right cerebellum. These 56 regions were used as nodes in the structural diffusion-based connectome, which was modelled using multi-shell multi-tissue constrained spherical deconvolution, anatomically constrained tractography and spherical-deconvolution informed filtering of tractograms. Normalized weighted graph measures of characteristic path length, global and local efficiency and clustering coefficient were calculated for each connectome and compared between different lesion types, CST wiring patterns and MACS levels using analysis of covariance, correcting for age. Similarly, partial correlations, correcting for age, were utilized for investigating the relation with AHA-score.

Results

Children with pWM lesions showed significantly shorter characteristic path length (F=9.9,p=.005) and higher global efficiency (F=4.6,p=.043) compared to pGM lesions. Surprisingly, though considering current sample size, different CST wiring patterns did not significantly alter global structural graph measures. Regarding motor outcome, children with lower MACS, i.e. better manual ability, had significantly shorter characteristic path length (F=12.2,p<.001), higher global (F=10.1,p=.001) and local efficiency (F=9.7,p=.001) and higher cluster coefficient (F=9.0,p=.001). Similarly, higher scores on the AHA, i.e. better bimanual performance, were  correlated negatively with characteristic path (r=-.65,p=.001) and positively with global efficiency (r=.58,p=.003), local efficiency (r=.50,p=.014) and cluster coefficient (r=.48,p=.019).

Discussion

We observed higher characteristic path length and lower global efficiency in children with pGM lesions compared to children with pWM lesions, which corroborates the observation that these children typically present with more extensive lesions.

Finally, all graph measures were highly related to manual ability in daily life (MACS) and a standardized test for bimanual performance (AHA). This confirms our hypothesis of everyday  motor functioning in children with uCP being dependent on remodelling of the whole structural connectome.

Objective

To apply an MRI based anatomical scoring system and compare the scores in CCA patients between four groups (isolated complete
and partial, and associated complete and partial). We aimed to assess and apply the scoring system to a varied cohort of CCA patients
subjected to fetal MRI and correlate the results with patient outcome.

Methods

Prospective collection of all cases of CCA who underwent fetal MRI between 2011 and 2019 in a tertiary care center. CCA cases were
grouped in four groups according to MRI findings (in consensus by two neuroradiologists): isolated complete and partial and associated
complete and partial. A 15-point score was applied based on T2-weighted single shot fast spin echo images directed at the fetal brain,
T2-weighted fluid attenuated inversion recovery (FLAIR) and echo planar imaging. Scoring was independently performed by two
neuroradiologists with experience on fetal MRI. Patient outcome was collected by a structured systematic phone questionnaire applied to the
parent/legal guardian. Pearson correlation was used to correlate the MRI score with the outcome and comparison of scores between groups was
performed using a two-way ANOVA.

Results

Average gestational age was 27 GW (range 19+4 to 37+1 GW). A total of 130 cases were included (Isolated n= 66 [Partial n= 25,
complete n= 41], associated n= 56 [Partial n= 31, complete n= 25]). Average Score for isolated CCA was 2.91±1.95 (Partial= 2.08±2.01,
complete= 3.52±1.68) and higher for associated CCA 6.47±3.08 (Partial= 6.13±3.42, complete= 6.83±2.67). There was a significant difference in
average total score between isolated and associated cases (p< 0.01). There was no moderation effect. These groups showed statistically
significant difference between complete and partial isolated cases (p= 0.001) but not between complete and partial associated cases (p=
0.33).

Conclusion

Based on anatomical score alone, it is possible to categorize isolated complete, isolated partial and associated CCA. These
results showed good correlation with neurological outcome in the surviving children. Adding the fetal MRI score to the everyday assessment
of CCA may aid in patient counselling.

Introduction

When cancer is diagnosed during pregnancy, both fetal and maternal outcomes need to be considered. Reassuring findings on short-term outcomes after prenatal cancer treatment exposure have led to more women being treated nowadays. However, in a recent event-related potentials (ERP) study, we found indications of prenatal chemotherapy exposure affecting the neurodevelopment of executive functioning.

In this preliminary analysis, we investigate the impact of prenatal chemotherapy exposure on functional neurodevelopment using resting-state functional MRI (rs-fMRI).

Methods

We recruited 25 nine-year-old children in the prenatal-exposed group, all born to mothers who were diagnosed with cancer during pregnancy and treated with chemotherapy from the second trimester of pregnancy onwards. Children in the control group, born to healthy mothers, were matched at a 1:1 ratio with regard to age, gestational age at birth (GA) and gender. Children having major prenatal or neonatal complications with known neurological sequelae, except for prematurity or intra-uterine growth restriction, were excluded. All children underwent MRI scanning between 2015-2019, on the same scanner (3T Philips Ingenia CX, 32-channel phased-array head coil).

Each participant underwent resting-state functional (T2*-weighted EPI, TE/TR=33/1700ms, resolution=3.59x3.59x4mm, FOV=230x230x120mm, 250 volumes) and T1-weighted structural (MPRAGE, resolution=.98x.98x1.2mm, TR/TE=9.6/4.6ms, FOV=160x256x256mm) MR imaging. All images were preprocessed using fmriprep (v1.5.0, https://github.com/poldracklab/fmriprep), including a fieldmap-less approach for correcting EPI distortion artefacts. The native anatomical space was used as the output space for the preprocessed images. Afterwards, nuisance regression and temporal band pass filtering (.009-.1Hz)  was applied using the denoiser toolbox (https://github.com/arielletambini/denoiser). For each participant, white matter and cerebrospinal fluid signal, framewise translations and rotations (along the principal axes) as well as the motion components identified by ICA-AROMA (aggressive setting) were regressed out. Further analysis was conducted using the CONN toolbox (v18.b). All images were segmented (GM, WM and CSF) and normalized to MNI. Next, a Gaussian smoothing kernel of 6mm isotropic was applied to the functional images. Based on the a priori hypothesis of possible impact of chemotherapy on attention and executive functioning, nodes of the default mode (DMN), fronto-parietal (FPN), dorsal attention (DAN) and salience networks (SN) were selected for statistical analyses. Functional coherence between these regions was estimated using bivariate correlations. ROI-to-ROI statistics were calculated using a model with group as main predictor (prenatal-exposed vs. controls) and GA as covariate. Significant effects of group and GA were assessed at p<.05 (false discovery rate corrected for multiple comparisons).

Results

No significant differences in functional coherence were observed between prenatal-exposed and control children. Longer gestation was associated with stronger connectivity (see figure), between the bilateral frontal eye fields (FEF) and between the left posterior parietal cortex (PPC) and the medial prefrontal cortex (MPFC).

Discussion

To our knowledge, this is the first study to investigate resting-state brain activity in children who were prenatally exposed to cancer treatment. Based on previous findings from an ERP experiment, we hypothesized an impact on attention and executive functioning to be reflected in the functional connectivity measures. However, no significant differences were observed in comparison to matched controls. On the other hand, we found (late-)prematurity to be linked to lower connectivity between the bilateral FEF components of the DAN, as well as between left PPC component of the FPN, and the MPFC component of the DMN. This corroborates the current literature in which alterations in these networks have been observed in premature children.

In conclusion, this preliminary analysis did not establish an effect of prenatal chemotherapy exposure on functional neurodevelopment, as measured by rs-fMRI. Future work will encompass an expansion of this dataset. These findings corroborate existing literature on positive short-term outcomes, stating that cancer during pregnancy can be treated without major health effects on the offspring.

Introduction

Cerebral palsy (CP) is caused by a non-progressive injury to the developing brain. While brain abnormalities in CP have been shown to relate to motor outcome, its association with cognitive development is less well established.

Aim of the study

To investigate the relation between brain abnormalities and motor function to cognitive outcome in children with CP.

Methods

We analyzed the available brain MRI scans of 75 children with CP of a prospective longitudinal study investigating developmental trajectories of children with CP in The Netherlands: the PERRIN study. Data of two PERRIN cohorts, 0-5 and 5-9, were used, including children with CP, Gross Motor Function Classification System (GMFCS) level I-V, aged 1-7 years at time of inclusion. Median age at time of neuro-imaging was 17 months (IQR 6-25). We classified the pattern of brain damage in the following qualitative categories:  Normal; Developmental brain disorder; Predominant white matter injury; Predominant grey matter injury and Non-specific lesions. We also scored the severity of damage of 12 brain regions from grade 1 (normal) to grade 3 (severe or diffuse damage). These semi-quantitative scales were summed up to subscores for ‘extent of white matter damage’ (grades of white matter signal abnormality, white matter reduction frontal and occipital, cysts and corpus callosum, ranging from 5 to 15) and ‘extent of grey matter damage’ (grades of thalamus, putamen, cortical grey matter and globus pallidus, ranging from 4 to 9). For cognitive outcome, we created two groups: children with intellectual disability  (intelligent quotient (IQ) <70) and children with normal intelligence  (IQ ≥70). IQ tests were performed at age 4.5 years (Snijders-Oomen Nonverbal intelligence Test for the 0-5 cohort) and at age 5 or 7 years (Raven’s Coloured Progressive Matrices for children of the 5-9 cohort). GMFCS level was assessed at age 2.5 years (0-5 cohort) or age 5 or 7 years (5-9 cohort). 
Statistical significance was set at alpha 0.05. Multivariable logistic regression with backward stepwise elimination was used to analyze prognostic factors for the presence of intellectual disability.

Results

The majority of the 75 patients had a spastic type of CP: 41 bilateral spastic (55%) and 25 unilateral spastic (33%). The most common GMFCS level was level I (43%) and the distribution over the remaining four levels was relatively even. The most common brain damage pattern was ‘predominant white matter injury’ (60%), followed by ‘predominant grey matter injury’ (22.7%). Multivariable logistic regression with backward stepwise elimination showed that at a young age, more extensive white matter damage is a prognostic factor for later intellectual disability (OR 1.66 for the extent of white matter subscore, 95% C.I, 1.19-2.33). This MRI based model can correctly distinguish between presence and absence of intellectual disability in 73.3% of the cases. When adding functional information, a model containing GMFCS level (OR 3.12, 95% C.I. 1.70-5.75) and the extent of white matter damage (OR 1.54, 95% C.I, 1.00-2.39) can explain intellectual disability correctly in 88.8% of the cases.

Conclusion

Brain MRI characteristics and the GMFCS level can help to identify children with cerebral palsy at risk for intellectual disability at an early age.

Interpretation

It is important to collect brain MRI and the GMFCS level of cerebral palsy patients at a young age, since these characteristics are prognostic factors for cognitive outcome at school age. The information could therefore be useful for clinical decision making.

Purpose: We aimed to evaluate if proton magnetic resonance spectroscopy (¹H-MRS) and diffusion tensor images (DTI) quantified in very preterm infants (VPIs) at term equivalent age (TEA) enhance the predictive role of conventional MRI for their neurodevelopmental outcome (NDO) at the corrected age of 12 months using feed-forward neural-networks (fNNs).

Methods and Materials: From 300 VPIs born before 32 completed gestational weeks who received an MRI scan at TEA between September 2013 and December 2017, 173 were excluded due to missing or poor-quality spectroscopy data and/or missing neurodevelopmental tests at 12 months corrected age. The data sets of 127 VPIs were considered for motor and cognitive development, of whom 13 and 7, respectively were categorized as delayed. We evaluated five metabolite ratios and two DTI characteristics, each in six areas of the brain. We performed a feature selection algorithm for receiving a subset of characteristics that were prevalent for the VPIs with developmental delay. To reduce bias by unbalanced classes, only VPIs that shared approximate values of those prevalent characteristics were considered for further calculations. We finally constructed predictors using fNNs.

Results: Predictors constructed by fNNs achieved a true positive rate of 85.7% and a positive predictive value of 100% for prediction of cognitive developmental delay, and a true positive rate of 76.9 % and a positive predictive value of 90.9 % for prediction of motor developmental delay.

Conclusion: ¹H-MRS and DTI quantified at TEA in VPIs add to the predictive value of conventional MRI for motor and cognitive development at the corrected age of 12 months. The proposed approach applying fNNs is promising for the use in clinical practice for identifying those VPIs that would mostly benefit from early intervention services.

Objectives

Corpus callosal agenesis (ACC) is one of the most common brain malformations, generally associated with a good outcome when isolated. However, around 1/5 of patients are at risk for neurodevelopmental delay, which currently available clinical and imaging parameters are insufficient to correctly identify. Predicting outcome in children with ACC still remains a challenge. Until now, prenatal imaging was unable to reliably identify those 20% of “isolated” ACC cases, who reportedly suffer from significant cognitive impairments postnatally. In this prospective study, a novel fetal MRI based scoring system was applied to predict neurodevelopmental outcomes in complete and/or partial ACC cases, which were classified as “isolated” by state-of-the-art prenatal neuroimaging.

Methods

We created a scoring system, based on anatomical fetal brain features, tailored to assess isolated CCA cases, using standard MRI images. It consisted of eight categories: gyration, opercularization, temporal asymmetry, lamination, hippocampal changes, basal ganglia, ventricular enlargement, and presence of blood products, translating minor anatomical features in a 0-3 points score, with a maximum
attainable score of 15 points. MRI features were scored independently by two neuroradiologists, blinded to the neurodevelopmental outcome, using a standard fetal MRI protocol (T2, T1, DWI, EPI). Outcomes were tested by a neuropediatrician and/or psychologist using Bayley Scale - Third Edition (BSID-III) and the Kaufman Assessment Battery for Children – Second Edition (KABC-II), depending on the age of
the children. Normal development by the BSID-III and the KABC-II was defined as a development quotient (DQ) score ≥85, moderate to severe developmental delay as a DQ score < 70. For correlation, the scoring in each individual test was grouped into “below average”, “average” and “above average”.

Results

Twenty-one children fulfilled the inclusion criteria (Nine (42.8%) female). A total of 36 fetal MRI examinations were evaluated in 20 pregnant women. Mean gestational age was 28.3 gestational weeks (SD±4.7; range 20-38 gestational weeks). All children in our cohort were born after 35GW and there were no perinatal complications. Scores in the current cohort ranged from 0 to 8 points, with a median of 3 points. Inter-rater agreement was excellent (ICC= 0.966; 0,934-0,982 95% CI). Postnatal neurodevelopmental evaluation was performed at an average of 2.05 years (SD± 1.61, range 0.53-5.77 years). There was a significant negative correlation between MRI score and outcome: cognitive (r=-0.461; p=0.005), motor (r=-0.415; p=0.012), and language skills (r=-0.472; p=0.004). All patients could be classified by MR Score in three predicting outcome groups: good, moderate and severe. Children with a score ≤3 (12/21) had average or above average neurodevelopment.

Conclusions

Prenatal diagnosis still fails to detect 1/5 of cases with isolated ACC that have significant developmental delays. The proposed MR based score improves the prognostic value of fetal MRI, using sequences that are available at all centres, possibly improving pre-natal counselling of patients with partial and/or complete isolated callosal agenesis.

Introduction: Sensorimotor perceptions are among the first that fetuses experiment in the mother’s womb, and preterm newborns already show some somatotopy in sensorimotor cortices for different parts of the body [1]. In the adult brain, somatotopic territories demonstrate different cytoarchitectures [2], suggesting links between the functional specialization and the network structural organization. Here we aimed to characterize the microstructural properties of sensory and motor regions in the developing brain, in relation to somatotopy.

Methods: We analyzed 3T-MRI data of 40 healthy infants between 37 and 44 weeks of post-menstrual age from the “developing Human Connectome Project”. Based on T2w anatomical images [3], we generated cortical surfaces, identified sulci and registered the newborns’ data using BrainVISA software [4]. We sub-divided the pre- and post-central gyri in all newborns, based on anatomical criteria intended to reflect somatotopic territories [5]. Based on pre-processed diffusion-weighted images (multi-shell: b=0, 400, 1000, 2600s.mm^-2) [6], we identified the cortico-spinal and thalamo-cortical tracts using the bundle-specific tractography approach [7] to face with fiber crossings in the corona radiata [8]. We computed maps of diffusion tensor imaging (DTI longitudinal λ_// and transverse λ_┴ diffusivities) [8,9] and neurite orientation dispersion and density imaging (NODDI volume fraction of intra-neurite compartment ν_in and orientation dispersion index ODI) [10]. For each region and tract, we evaluated age-related changes in the indices, and we made comparisons between regions or tracts using paired Student’s T-tests.

Results: Thanks to very precise registration over the group (Figure 1), we could reliably define 3 sensory and 3 motor regions in all newborns (Figure 2a). The tractography procedure was successful in reconstructing the 6 tracts in 29 newborns out of 40 (Figure 2a). ν_in was related to λ_// and λ_┴ in the cortical regions (N=40) and tracts (N=29) respectively. As a function of the newborns’ age, we observed λ_// decreases and trends of ν_in and ODI increases in regions, as well as λ_┴ decreases and ν_in increases in tracts. Microstructural differences were observed between the motor and sensory regions and tracts, and along the somatotopic organization (Figure 2b). NODDI indices were globally more sensitive than DTI diffusivities for the comparisons between regions or between tracts.

Discussion: This diffusion MRI study in newborns suggests that early microstructural differences between sensorimotor regions and tracts might relate to functional specialization for different body parts. Further comparisons with functional mappings should confirm this hypothesis. The network shows intense maturation over this age range, which might relate to the early sensory and motor acquisitions in infants. Nevertheless, disentangling between microstructure complexity and maturation requires further comparisons with adult data since mature cortices also show microstructural differences [11]. Such an approach could allow us to identify early disruptions in the maturation of the sensorimotor network, for instance in newborns at risk of cerebral palsy.

References: 1.Dall'Orso S et al, Cerebral Cortex 2018. 2.Kuehn et al, Cerebral Cortex 2017. 3.Makropoulos et al, Neuroimage 2018. 4.Lebenberg et al, Brain Structure and Function 2018. 5.Germann et al, Cerebral Cortex 2019. 6.Bastiani et al, Neuroimage 2019. 7.Rheault et al, Neuroimage 2019. 8.Dubois et al, Brain Plasticity 2016. 9.Lebenberg et al, Neuroimage 2019. 10.Batalle et al, Neuroimage 2019. 11.Fukutomi et al, Neuroimage 2018.

Acknowledgements: This research was supported by grants from the Fondation de France (2017), the Médisite Foundation (2018) and the IdEx Université de Paris (ANR-18-IDEX-0001). Data were provided by the developing Human Connectome Project, KCL-Imperial-Oxford Consortium funded by the European Research Council under the European Union Seventh Framework Programme (FP/2007-2013) / ERC Grant Agreement no [319456]. We are grateful to the families who supported this trial.

Background

Myelin Oligodendrocyte Glycoprotein Antibodies (MOG-Ab) positivity has been widely identified in Acquired Demyelinating Syndromes (ADS) such as Acute Disseminating Encephalomyelitis (ADEM), Optic Neuritis (ON), and Longitudinally Extensive Transverse Myelitis (LETM). We aimed to determine the range of radiological patterns in comparison with distinctive imaging features of MOG-Ab-encephalomyelitis in published cohorts (Hacohen Y et al 2017, Dubey D et al 2018).

Methods

We present a cohort of ten patients with MOG-antibody associated demyelination from Royal Manchester Children’s Hospital. MR brain, spine and orbits were obtained based on the presenting complaints. Imaging findings at the first presentation and during further relapses were assessed in agreement by consultant paediatric neuroradiologist and neuroradiology fellow in correlation with clinical history. The course of disease was classified as either monophasic, prolonged monophasic, or relapsing. Prolonged monophasic disorder was defined as radiological progression/worsening within one month of the acute presentation.  

Findings/Discussion

The age range at the first presentation was 3-13 years with a M:F ratio of 5:5. Six patients presented with monophasic disease, two with prolonged monophasic, and two with relapsing course with at least two distinct episodes. Classical ADEM imaging pattern with patchy white matter lesions was found in four patients. Two patients presented with transverse myelitis. Brain and spinal cord were affected simultaneously in five patients. Four patients were presented with ON, one of them with isolated ON (normal MRI Brain and Spine). The figure demonstrates imaging features of MOG-Ab associated demyelination.  

We found imaging similarities with comparison cohorts: propensity for poorly demarcated lesions in the cerebellar peduncles; involvement of the cerebral white matter and cortex, thalamus and basal ganglia, brainstem; long-segmental ON, and LETM with preponderance for central grey matter. Two patterns of LETM were observed – contiguous holocord and multilevel long-segmental. “Leukodystrophy-like” extensive brain white matter lesions have not been documented in our cohort. Follow-up imaging showed near or complete lesional resolution associated with parenchymal volume loss. 

Cerebellar involvement was demonstrated in four patients, two of them developed cerebellitis. One of our patients with cerebellitis presented with relapsing cerebellitis associated with a patchy “lace-like” enhancement pattern and cortical enhancement. In another patient, cerebellitis was accompanied by massive cerebellar oedema with resultant supratentorial hydrocephalus. Cerebellar involvement has been mentioned in Chinese cohort of 26 patients (Mao L et al 2019) with no details or imaging provided. 

Conclusion

Our study demonstrates the wide range of neuroimaging patterns in paediatric MOG-Ab positive ADS, mostly in concordance with comparison cohorts (Hacohen Y et al 2017, Dubey D et al 2018).  Cerebellitis has not been specifically mentioned in published cohorts, and is presumably a new imaging pattern in MOG-Ab-encephalitis.

Background: the neuronal ceroid-lipofuscinoses (NCLs) are a group of inherited, neurodegenerative, lysosomal storage disorders affecting children and young adults. They are characterized by progressive intellectual and motor deterioration, seizures, retinopathy and early death. There are several forms of NCL according to the responsible mutated gene (numbered from 1 to 14) and the age at disease onset. Late infantile NCL type 2 is one of the most common forms of NCLs, caused by mutations in the NCL2 gene, resulting in dysfunction of the lysosomal enzyme tripeptidyl peptidase 1 (TPP1). First symptoms of the NCL2 generally occur between ages of two to four years and include motor decline with muscle coordination loss (ataxia), intellectual disability, epilepsy and vision loss. Symptoms of NCL2 progress rapidly as the child gets older, leading to death around the middle teenage years.

While NCLs remain incurable with symptomatic and palliative treatment, it has recently been shown for NCL2, that, replacement of the dysfunctional lysosomal enzyme through intraventricular infusion of a functional enzyme (cerliponase alfa) can effectively slow down the progression of the disease. Therefore early diagnosis of NCL2 is needed to be able to start the treatment as soon as possible. Unfortunately there are no early pathognomonic clinical signs of NCL2.

Purpose: the aim of our study was to describe brain magnetic resonance imaging (MRI) abnormalities during the course of late infantile NCL2 in order to evaluate the role of MRI for early diagnosis.

Materials & Methods: we retrospectively reviewed seven brain MRI obtained from six children (aged from two to six years old) with biochemical and genetic confirmation of NCL2. We recorded the qualitative alterations (anatomic alterations and signal intensity abnormalities) on conventional (T1, T2, FLAIR) and functional (DWI, spectroscopy) sequences concerning specific anatomic structures (cortex, white matter, basal ganglia, corpus callosum, visual pathways and brain stem) in the supra- and infratentorial regions 

Results: a specific imaging pattern was described in all patients including 1. early diffuse and progressive cortical - subcortical atrophy, predominating in the supratentorial region 2. bilateral, parieto-occipital periventricular white matter hyperintensity, including optic radiations, on T2/FLAIR sequences 3. thalamic abnormalities characterized by loss of volume and low T2 signal of the anterolateral part of the thalamus and 4. bilateral hypomyelination of the posterior limb of internal capsule, a sign not yet described in the literature.  

Conclusion: in the absence of pathognomonic clinical signs of NCLs, brain MRI might have a potential role in the early detection of NCL2 by providing a suggestive imaging pattern including diffuse atrophy of the cortex and white matter, with signal intensity abnormalities in the thalamus and the supratentorial white matter including posterior limb of internal capsule and optic radiations.

Introduction and aim:

The term interferonopathy was introduced in 2011 (1), to define a group of Mendelian disorders associated with upregulation of type I Interferon (IFN). The human type I IFN family includes 13 IFN-α subtypes that have antiviral, antiproliferative, and immunomodulatory effects. (2) As some interferonopathies, especially those diagnosed prenatally and in childhood mimic congenital infections, the umbrella concept of “pseudo torch” manifestations has been coined for mimics of torch infections. The common factor is an elevation of IFNI levels that may have different pathophysiological backgrounds. The number of genes that may be involved in the generation of these syndromes is constantly rising (3). Prenatal pathological and imaging features have been described (4). However, MR patterns have not been evaluated in detail. Aim of this retrospective study is to demonstrate MR patterns in 7 cases with manifestations, interpreted as congenital infection-like phenotype (CILP).

Patients and Methods:

7 Fetuses, aged from gestational weeks (GW) 22-32 with 10 in-vivo and 2 postmortem MRI examinations were included. In 2 cases pathohistological workup was available, in 1 fetus with additional NCAM-immunohistochemistry-based structure tensor analysis of developing fibers (5). Results of genetic examinations in accordance with the spectrum of changes associated with pseudotorch was proved in 4 cases (from 2 families). One patient was under interferon therapy because of hepatitis C during pregnancy. One had a florid Crohn’s disease. One case showed characteristical signs on imaging studies with negative torch parameters. In all cases the presence of torch-infections (especially CMV) had been excluded. Prenatal MRI examinations were performed at 1.5T or 3T fieldstrength. Protocols included T2-weighted (w) TSE sequences, T1-w, FLAIR, Epi/SWI, DWI and DTI Sequences in 2-3 planes, with a slice thickness of 2-4mm through the fetal brain, and T2-w, SSFP, T1-weighed, EPI/SWI, DWI sequences of the fetal body and extracerebral organs. Postmortem examinations were done at 3T, using isovoxel (CISS) T2-w sequences, T1-w, SWI, DTI Sequences over the fetal brain and body with 3-5mm slice thickness.

Results: Abnormal lamination was present in all exams, abnormal cortical folding in 9/10, Calcifications in 7/10, hemorrhage in 3/10, diffuse edematous changes in 2/10, hepatosplenomegalia in 7/10, and palcental changes (premature aging and/or hemorrhages) in 7/10

Conclusion:

The presented fetal MRI case series extents the known morphological findings of CILP to an “acute” manifestation with brain edema +/- hemorrhage. In addition to testing for infectious agents and genetic workup, other factors such as interferone therapy should be considered. The strength of fetal MRI to visualize abnormalities of lamination and its advantage to identify acute DWI changes of the fetal brain may lead the path to the reduction of exposure to existing “teratogenic” effects and potentially to the initiation of novel therapies at early stages, preventing further harm to the developing central nervous system.  

1. Crow YJ. Type I interferonopathies: a novel set of inborn errors of immunity. Ann N Y Acad Sci. 2011; 1238:91–98.

2. Banchereau R et al. Annu Rev Immunol. 2017 Apr 26;35:337-370.

3. (3) Meuwissen M et alJ Exp Med. 2016 Jun 27;213(7):1163-74

4. Bourgon L et al, Prenatal Diagnosis. 2019;39:806–810

5. Mitter C et al. Front Neuroanat. 2015; 9: 164.

INTRODUCTION

Several approaches exist for treating patients with refractory epilepsy due to hypothalamic hamartoma (HH) with different risks, morbidity and success. Real-time MRI-guided laser ablation is a recent and promising technique for HH and we have recently started to use it since its 2018 EMA (European Medicines Agency) approval. We present a descriptive analysis of our experience, safety and short follow-up in 5 pediatric patients.

METHODS

Patient selection

Patients were selected based on drug-resistance epilepsy and hamartoma type/size according to Delalande classification. We decided to start with class I to III in order to get more experience and not make treatment times very lengthy.

MRI Imaging, planning  and treatment

All MRI studies (preop, intraop and postop) were performed in a 3T  Philips Ingenia equipment. Fiber trajectory was planned with Voxim® software and performed with stereotactic-robotic arm Neuromate-Renishaw®. Treatment was designed to achieve disconnection, and if possible, complete ablation, with several pull-outs if needed.

Ablation was performed with the Visualase^® system. Volumetric T1 was initially acquired to check catheter position and select the two required planes by the system for monitorization. The water-pump was switched on and an ablation test was performed. Therapy was continuously monitored by real-time MR-thermography during ablation times  (Figure1) and with DWI, FLAIR and T2 during in-between cooling times to identify complications and assess disconnection/ablation. A complete MRI was performed at the end and at 24h.

Analyzed data

We specifically recorded hamartoma type and seizures, MRI findings, blood&urine analysis and visual field test prior and after treatment.

Results

We treated five patients from April-October 2019,  5-17 years old, 2-9 months follow-up. HH types were II (2) and III (3).  All patients had mostly gelastic seizures.

Intraoperative MRI changes included lesion increase, restricted diffusion, intralesional bleeding and high FLAIR/T2 signal (Figure2). In the surrounding tissues high FLAIR/T2 signal and tumefaction was also found, in the 24h persisted or were more extensive (Figure3). Currently, at six months follow-up MRI (2 patients), lesion tumefaction diminished but bleeding and high FLAIR/T2 signal persisted and surrounding tissue changes improved.

Regarding safety no hormonal, electrolyte, osmolarity or visual anomalies were found. but we have had a serious complication, a mild hemiparesia currently recovering due to tissue damage in the fiber path (Figure4). We presume it might be related to momentary water-pump malfunctioning that briefly occurred during one ablation pulse.

We have also had a focal cortical/subcortical bleeding in the brain surface of fiber placing as described in SEEG positioning.

The same patient also had self-limited high fever that we attributed to central fever. This patient also had the most extensive high-signal alteration in the surrounding tissues.

Although not significant, currently all patients are seizure free under medication.

CONCLUSIONS

1. Real time MRI-guided laser treatment for HH is a complex technique that requires tight multidisciplinary teamwork to achieve the outmost precision in treatment planning, execution and monitoring for patient safety and treatment success.
2. The system’s MRI thermography is a very useful technique for monitoring estimated lesion and surrounding tissue temperature but conventional MRI sequences are needed both to evaluate parenchymal safety out of treatment planes and assess tissue treatment-related changes that may ensure complete lesion disconnection/ablation.
3. In our short experience it seems a safe technique, but accurateness of fiber path planning and positioning is crucial as well as cautious and constant supervision of correct pump functioning.
4. In terms of efficacy all lesions seemed properly disconnected in the postop MRI, all surrounding tissue abnormalities improved in the long term, and currently all patients are seizure-free under medication but longer follow-up is mandatory.

Hematopoietic stem cell transplantation (HCT) is a partially effective therapy for patients with metachromatic leukodystrophy (MLD), a disorder characterized by deficient arylsulfatase A (ASA) activity and sulfatide storage. We examined the time course of white matter abnormalities on magnetic resonance imaging (MRI) after HCT, and investigated brain tissue to define the therapeutic effects of HCT on tissue level.

The MRI scans of fifteen successfully transplanted MLD patients with at least one available MRI examination before and three after HCT, were retrospectively evaluated with the MLD-Loes score. Brain tissue was obtained at autopsy from two transplanted and six non-transplanted patients, and two age-matched controls. The presence of donor cells, ASA, and sulfatide storage and digestion was examined by immunohistochemistry and microscopy. Myelin content, oligodendrocyte numbers and macrophage phenotypes were also assessed. An unpaired t-test, linear regression or Mann-Whitney U-test was performed to evaluate differences between the transplanted, non-transplanted and control group.

MRI white matter abnormalities initially deteriorated, but improved (n = 7) or stabilized (n = 8) from approximately one year post-HCT onwards (mean follow-up = 82 months). In the autopsy obtained white matter of transplanted patients, we found metabolically competent donor macrophages that were immunopositive for ASA. These macrophages preferentially expressed markers of alternatively-activated, anti-inflammatory cells compared to macrophages in non-transplanted patients. Besides, transplanted patients showed higher numbers of oligodendrocytes and remyelination.

In successfully transplanted MLD patients, brain white matter abnormalities may improve on MRI for years. HCT leads to the presence of donor macrophages with the metabolic ability to degrade sulfatides in the white matter. Besides, their anti-inflammatory phenotype may support oligodendrocyte survival and differentiation, thereby enabling remyelination.

Background: Mucolipidosis type IV (MLIV) is an autosomal recessive lysosomal disease that mainly affects brain white matter. It is caused by variants in the MCOLN1 gene, which encodes for transient receptor potential ion channel mucolipin-1 (TRPML1). It has been described mainly among Ashkenazi Jews, where the carrier frequency is estimated to be 1 in 100.  

MLIV causes severe impairment in infant motor and cognitive development, progressive vision loss, gastric achlorhydria often resulting in iron deficiency anemia. Impaired myelination and a dysplastic corpus callosum are the major hallmarks of post-natal brain MRI. Human brain pathology data in MLIV is limited to only two autopsy cases, that revealed gliosis, abnormal white matter and numerous storage inclusions in all types of brain cells. Fetal MRI has been described in the literature demonstrating a thin corpus callosum.

Disease pathogenesis is still unclear. Defects in transport along the lysosomal pathway, affecting membrane sorting and/or late steps of endocytosis have been postulated. Impaired iron release from the endolysosomes has been suggested to contribute to the pathogenesis of MLIV.

Objective: To describe novel fetal brain MRI and pathology findings in a fetus with a proven, postmortem diagnosis of MLIV.

Methods: Fetal MRI was performed at 31 and at 35 weeks of gestation due to a sibling with progressive spastic quadriparesis and visual impairment and a brain MRI demonstrating white matter abnormalities without a known diagnosis. Termination of pregnancy was performed following abnormal MRI findings. An autopsy and whole exome sequencing were done.

Results: Fetal brain MRI showed progressive curvilinear hypointensities in the periventricular white matter on T2-weighted images, predominantly in the frontal lobes, accompanied by a thin corpus callosum observed on the second fetal MRI.

Fetal brain pathology revealed hypercellular white matter composed of reactive astrocytes and microglia, multifocal white matter abnormalities with mineralized deposits predominantly in periventricular regions, and numerous aggregates of microglia with intracellular iron content and reactive astrocytes predominantly in the frontal lobes. A thin corpus callosum was observed. WES revealed a paternally inherited c.1135-1 G>C variant in MCOLN1. Ashkenazi Jews founder mutations analysis revealed a maternally inherited g.4127_10560del6434 0.14 variant in MCOLN1.  

Conclusions: This is the first demonstration of white matter abnormalities on prenatal brain MRI and a neuropathological description of a fetus with MLIV. Interestingly, brain pathology in the mouse model shares similar findings. These findings can aid in understanding the pathomechanism of the disease which includes prenatal white matter involvement with dramatic activation of glial cells and impaired iron metabolism.

RATIONALE Vanishing White Matter (VWM), caused by recessive mutations in any of the EIF2B1-5 genes,  is one of the more common leukodystrophies. The clinical spectrum is broad. Patients with a lower age of onset (AoO) generally have a faster disease course. A large set of MRI scans of patients with VWM has so far not been analyzed.

METHODS Genetically confirmed and clinically phenotyped VWM patients were divided according to AoO: <1 y (group 1), 1-2 (2), 2-4 (3), 4-8 (4), 8-18 (5) and >18 (6). MRI scans were available from 1985 on. We applied three techniques to quantify white matter (WM) decay. I. The ventricle-skull ratio (VSR, i.e. the sum of the maximum transverse diameter of the frontal horns, occipital horns and the cellae mediae, divided by the maximum transverse inner diameter of the skull) was measured to determine brain atrophy. II. We visually scored the aspect of the WM (normal, FLAIR hyperintense, and rarefied or cystic) as percentages of the total cerebral WM, resulting in an aspect score (0-30, 0 being 100% normal, 30 being 100% cystic). III. Using SIENAX (FMRIB Software Library) and a lesion prediction algorithm (LPA, LST), T1 and FLAIR images are semi-automatically segmented to determine volumes of normal, FLAIR hyperintense, rarefied or cystic WM, and CSF. Violation of normality led us to report medians and Kruskal-Wallis analyses to determine differences in disease duration, VSR, and aspect scores between groups.

RESULTS 422 scans of 252 patients were analyzed. Disease duration at the time of the MRI scan and VSR were different between groups (both p<0.001). Group 1 had the shortest (0.26 y) and group 6 had the longest (5.8 y) disease duration at MRI. Groups 2 and 3 had the lowest (0.85) and group 6 had the highest VSR (1.10). The aspect score was assessed in 303 scans of 200 patients. There was a difference between groups (p<0.001); group 1 had the highest (17.8) and group 6 had the lowest aspect score (12.9). First analyses suggest a faster disease course in patients with a lower AoO. Results of the semi-automatic segmentation are currently in progress and will be presented at the meeting.

DISCUSSION The aspect score indicates that patients with a lower AoO have more WM rarefaction and cystic degeneration, even after a short disease duration. Patients with a higher AoO show more WM atrophy, as indicated by a higher VSR. Our research underlines the heterogeneity of VWM. This information can be used to inform patients and serve as a reference in future clinical trials. 

Background
Metachromatic leukodystrophy (MLD) is a neurodegenerative disease, the hallmark of which is progressive demyelination of brain white matter due to sulfatide accumulation. Depending on age of onset, patients can be categorized into three groups: late-infantile, juvenile and adult. These groups differ in rate of deterioration. While primarily a white matter disease, MLD also affects grey matter, which is already evident at diagnosis. The extent and nature of this involvement during disease evolution are unknown, nor is it clear how grey matter abnormalities relate to cognitive function. 

Methods
3D T1-weighted MR images were obtained at both 1.5T and 3T. At baseline, 20 patients from all three groups were compared to age matched controls (N=47). Patients scheduled for treatment with hematopoietic stem cell transplantation (HSCT) (N=11, 5 juvenile, 5 adult, 1 late-infantile) and untreated patients (N=9, all juvenile onset) were analyzed in 2 separate groups, and also compared to each other. These analyses focused on cortical thickness, total grey matter volume (GMV), deep grey matter volume (DGMV), particularly the thalamus, and ventricular cerebrospinal fluid (vCSF) volumes. This was followed by longitudinal analyses of both the treated and untreated patient groups (median follow-up time 2.5yrs, range 1.1-13.9). The longitudinal investigation focused on the same brain structures and, only for the treated patients, on cognitive function, which was evaluated using IQ scores (age appropriate developmental, nonverbal or full IQ scales). The change in cognition over time was related to brain atrophy.

Results
Cross-sectional analyses at baseline showed smaller volumes for nearly all analyzed structures in untreated patients, compared to controls. Patients scheduled for treatment showed smaller thalamic and deep grey matter volumes for juvenile subjects and smaller thalamic volumes for adult subjects, both compared to controls. Untreated patients compared to patients scheduled for treatment showed lower thalamic and DGMV, also at baseline.
Longitudinal analyses showed progressive atrophy in the untreated patient group: a decrease in GMV and DGMV, along with an increase in vCSF volume. Within the treated group, the same analyses showed increasing vCSF volumes and an initial decrease in cortical thickness in both juvenile and adult MLD patients, which normalized later in the disease course. In the adult patients, the initial decrease was accompanied by a significant decrease in total GMV and thalamic volume.
Finally, the percentage change in IQ scores of treated patients positively correlated with changes in DGM, and thalamus volumes, and negatively correlated with  vCSF volumes.

Conclusion
Our results show progressive atrophy of both cortical and deep grey matter in untreated MLD patients. Treated patients initially also show cortical thinning. This may be due to the ongoing corticosteroid treatment as part of the HSCT protocol at the early stages of follow up, as it partially normalizes at a later time point. However, the atrophy as shown by DGM volume decrease or vCSF volume increase persists, and is correlated with cognitive decline. 

Abstract:

Introduction:

Organic acidurias are a group of rare inherited in born errors of metabolism disorders that result from deficiency in the activity of a specific enzyme that catabolize amino acids, carbohydrates or lipids. This leads to accumulation of mostly dicarboxylic acids in multiple organs, including the brain. Due to its low transport capacity through the blood-brain barrier, and at high concentrations it may become neurotoxic. Organic Aciduria are characterized by a predominant or even exclusive neurological presentation. Different subtypes of organic aciduria include, most commonly methylmalonic acidemia, Propionic acidemia, Isovaleric aciduria, glutaric aciduria type I, l-2-hydroxylgutaric aciduria and d-2-hydroxyglutaric aciduria types I and II. Neuroimaging although nonspecific plays an important role along with genetic and laboratory investigation in confirming the diagnosis and ruling out other diseases with similar presentations. Neuroimaging is also important in the follow up of these vulnerable group of patients especially those prone to complication related to neurotoxicity or malignancies for example l-2-hydroxylgutaric aciduria confirmed to develop CNS tumors.

Method:

Salmaniya Medical Complex is the largest tertiary teaching hospital in Bahrain with the only specialized pediatric metabolic disease unit on the island with referrals from other hospitals. A retrospective review of patients confirmed to have organic aciduria through biochemical and genetic studies from 2000 to 2018 was performed. Neuroimaging studies included Computed Tomography (CT) brain and Magnetic Resonance Imaging (MRI). The neuroimaging studies of patients were reviewed.

Result:

Over the period from 2000 to 2018, 29 patients were confirmed to have organic acidurias, there were 13/29 (44%) males and 16/29 (56%) females. Methylmalonic aciduria was the most common disease seen in 12/29 (41%), 7/29 (25%) had Propionic Acidemia, 2/29 (7%) had Isovaleric Aciduria, 3/29 (10%) had Glutaric Aciduria type I, 2/29 (7%) both siblings had D-2- Hydroxyglutaric Aciduria. Patients diagnosed L-2-Hydroxyglutaric Aciduria were 3/29 (10%) with an adult presentation of the disease at 25 years and his younger sister at 12 years of age. One of the patient (12years male) with L2 Hydroxyglutaric Aciduria presented with space occupying lesion proven by histopathology as Glioblastoma IDH-1 wild type WHO grade 4.  

Most of our patients 15/29 were diagnosed within the first week of life. 5/29 within the first year of life and 8/29 were older than 3 years at the time of diagnosis. 1/29 patient was 25 years old at time of diagnosis. We present a rare presentation of malignant transformation in a 12 years old boy with L-2 Hydroxyglutaric Aciduria.

All patients had Brain CT/ MRI imaging or both.

Conclusion:

We have found that the neuroimaging appearances of Organic Acidurias can be nonspecific, however they can provide a useful guide in the diagnostic work up or may confirm the clinical suspicion or narrow the differential diagnosis. It is important to diagnose these diseases early in order to facilitate early initiation of targeted metabolic treatment and to prevent permanent neurotoxicity and malignant transformation in some susceptible patients.

Pediatric cerebral vasculitis is an inflammatory process involving the blood vessel walls lead to vessel narrowing and stenosis.  It could be primary or secondary to systemic disease , drugs, or radiation therapy.  The initial presentation of cerebral vasculitis could be variable, most often the patients presents with acute symptoms of vascular insuffecincy. 

Cerebral vasculitis remains difficult to diagnose despite of the availability of multiple modalities including MRI, MRA and angiography.  Vessel wall biopsy is the gold standard study for definitive diagnosis.

Vessel wall imaging (VWI) is an emerging MRI technique to image the arterial wall of intracranial arteries. Many previous studies showed arterial wall imaging is helpful technique for diagnostic and prognostic consideration in cerebral vasculitis.   

 Our study aim is to identify the yield of contrast enhanced high-resolution MR imaging for identification of vessel wall abnormality as early diagnosis for intracranial vasculitis in our pediatric patient.  

MATERIALS AND METHODS:

Retrospective observational study of children between Jan 1^st, 2016 to Jan 1^st , 2018 was performed at KASCH. Children with clinical suspicion of acute ischemic stroke (AIS) who underwent VWI were identified.

 Inclusion criteria: Age < 14 years , all patients had no previous cardiac disease or chronic illness. All patients referred to MRI to roll out  acute ischemic stroke. All patients had vessels wall imaging study (VWI) at the time of the initial diagnosis.

Retrospective chart review for all the cases identified were performed.  Patients demography, clinical presentation, and final diagnosis were recorded.

The MRI images were evaluated blindly by a pediatric neuro-radiologist. The present of brain parenchymal abnormality on T2 WI, fluid-attenuated inversion recovery and DWI images were recorded. 

 The vessels involvement were evaluated on time-of-flight MRA and VWI . The evaluation criteria for VWI  include:  Present or absent of enhancement ,and the patterns of enhancement in region of interest as concentric or eccentric.

 Vessel wall Imaging Protocol :

Both a high contrast-to-noise ratio (CNR) and a high spatial resolution are needed to visualize the thin arterial VW and to characterize VW lesions.

We used GE  Optima MR450w – 1.5T , or GE Discovery MR370. VWI sequences include: 3T 3D T1 WI SE – pre and post contrast images ,2mm slice thickness – spacing 0, FOV 12 cm and Matrix 256 x 192.

RESULTS:

The radiographic database identified 8 patients with clinical suspicion of AIS, and vessel wall imaging. Ages ranged from 2 years to 13 years. all patients (100 %) were male. All the patients presented with clinical symptoms and signs of intracranial vascular insufficiency.

MRI imaging demonstrate Abnormalities in 7/8 (88.9%) patients on T2-weighted and FLAIR images. Abnormalities were seen in on DWI sequences and ADC maps  confirm the present of acute ischemic infarction.

Abnormal vascular narrowing /stenosis was present in 6/8(75%) patients with an abnormal TOF-MRA. The VWI study was positive in 5/8 patients

( 62%) and negative in 1 patient only with positive TOF-MRA and final diagnosis of Myoamyoa disease.

The pattern of the vessel wall enhancement was smooth and concentric in  all patients . The distribution of the vessel abnormality and enhancement in VWI were most commonly involved  the anterior circulation distribution (  85.7%) ,  and less commonly in the posterior circulation (14.3%).

Three patients had disease progression in the follow up study (33%) at an average time interval of 1 year.

Conclusion:

Vessel wall imaging ( VWI )  study is strongly  associated with  intracranial inflammatory arteriopathy in our population. We recommend to apply VWI in conjunction to TOF-MRA for previously healthy pediatric patient presented with clinical symptoms and signs of acute ischemia which aid in early diagnosis which may prevent further brain damage.

Introduction

Migraine is the most common neurological disorder and the third most common disease worldwide. However, the underlying mechanisms contributing to its development is not completely understood. Symptoms may arise from a combination of dilation-independent vascular events and neurogenic mechanisms interacting throughout the brain and within the trigeminovascular system in the meninges

Materials and method

We report here a case of a patient with a suspected familial hemiplegic migraine who presented an increased recurrences of events from one per month to one every other day. Three MRI acquisitions were performed after the appearance of a strong crisis which included a paresthesia and aphasia along with headaches. Two MRI were performed close to the crisis, while the last one was done one month later.

Perfusion maps were estimated using a 3D pseudo continuous arterial spin labeling (3D-pCASL) sequence (bolus/PLD=1600/1400ms). For MRI 2 and 3, two additional dynamic acquisitions were performed. A 2D dynamic phase contrast imaging (PCMRI) and a 3D pulsed arterial spin labeling sequence (3D-PASL) to confirm 3D-pCASL results with similar parameters (bolus/TI=1600/1400ms).

Results

Pseudo-continuous ASL sequence exhibits physiologically impossible negative values at MRI1 (Figure 1). No control perfusion acquisition was available.

Wrong ASL values can still be seen at MRI2 between hemispheres. The control 3D-PASL reveals a correct CBF with no asymmetry. However, PCMRI data shows a strong phase lag between left internal carotid (ICA) and basilar artery (BA) and more importantly right ICA (Figure 1, PCMRI) with systolic peak located respectively at 548, 604 (+56) and 641 (+93) ms. 

The final MRI exam highlights a normal perfusion with both 3D-pCASL and 3D-PASL and no more phase lag.

Discussion

Cerebral blood flow decrease and increase has been widely investigated in migraines. More than a real CBF alteration, we here see a striking artefact on pseudo-continuous arterial spin labeling: negative but “consistent” CBF values in term of contrast. This is an important observation because the way we see them could be erroneously interpreted as asymmetrical cerebral blood flow by non-specialist radiologists. In fact, in the absence of any anatomical vascular abnormalities, such finding might be a marker of the alteration of the blood flow kinematics in the inputs arteries.

Kinematics was thus investigated using PCMRI. The blood flow parameters obtained in the ICA and BA during second exam are in range of the expected values. Previous studies didn’t show any alteration of these parameters during migraine attacks. However, their detailed kinetics was not investigated. Regarding our data, their profiles across the cardiac cycle highlight an important phase lag between the two ICA. The pCASL uses a labeling scheme which can be sensitive to the cardiac cycle¹², and therefore explains the observed CBF asymmetry. During the last exam, CBF is correct and velocity curves appear resynchronized. One can suggest that, at MRI1, CBF was not asymmetrical, but the result of an altered pulse wave velocity (PWV) leading to an ASL mislabeling. PWV is the velocity at which the arterial pulse propagates through the circulatory system, seen here as the phase lag. PWV depends mainly on arterial vascular tone.

We know the influence of the parasympathetic system on the cerebrovascular tonicity. During CSD, cortical neurons and astroglia release substances which alter the cerebrovascular tone to compensate the high demand in necessary energy to restore ionic gradients and neuronal integrity. The transitory nature of the observed modifications suggests, for the first time, a reversible alteration of the vascular tone of the ICA in migraines subjects’. This alteration seems to follow the status of the patient from a severe migraine to a baseline state. 

BACKGROUND: Postoperative paediatric cerebellar mutism syndrome (POPCMS) is a complication affecting 8%-31% of children undergoing posterior fossa tumour resection. Injury to the proximal efferent cerebellar pathway during surgery is increasingly implicated in its causation but the exact pathogenesis remains unclear. Identifying MRI changes in the brain stem and cerebellum of these children can further the understanding of the pathogenesis and anatomical substrates involved.

METHODS: Longitudinal analysis of follow –up MRI in 40 children with (N=7)and without (N=33) POPCMS following posterior fossa tumour resection was performed using a fully automated quantitative analysis of the brainstem and cerebellum in 3D space using the following steps:  registration, segmentation, longitudinal interpolation of the MR images and machine learning techniques for feature selection and data classification. Multi-slice 2D MR image slices were interpolated in space and time to produce a 4D volumetric MR image dataset providing a longitudinal representation of the cerebellum and brain stem at specific time points across treatment. The deformations within the brain over time and the changing grey-level intensity of areas within the brain over time were analysed using machine learning techniques in order to identify biomarkers that correspond with the development of POPCMS following tumour resection.

 RESULTS: Longitudinal analysis and machine learning of the first 7 month MRI data following surgery identified features at specific locations in  lobule V, I-IV, VIII a of the anatomical right cerebellar hemisphere and at the junction of vermis and right I-IV lobules following surgery that are strong correlates of POPCMS. 3 features were noted at 1 month and 1 feature each was noted at 3, 4, 5 and 7 months following surgery.

CONCLUSION: This novel fully automated quantitative analysis of MRI in children following posterior fossa tumour resection reveals biomarkers that can help to further the understanding the pathogenesis of POPCMS. This is one of the few studies that support the involvement of the right cerebellar hemisphere in the development of POPCMS. These biomarkers could also serve as imaging surrogates to the severity of the injury and prognosis. This method of analysis will be applied to larger data from multicentre trials to increase the accuracy of the results.

PURPOSE

MRS is a helpful MR modality in the assessment of certain neurometabolic and neurodevelopmental disorders. Normal MRS spectra have been well-defined in older children and adults. However, many of these disorders present and are evaluated during early childhood. MRS spectra of children evolve during early brain maturation and the normal spectrum appears as a disease based on criteria derived from older individuals. Easily accessible normal values according to the age of the child being evaluated are not widely available to radiologists interpreting these studies. The main purpose of this study is to provide a guideline for the normal MRS metabolite values across an age spectrum early in life, so deviations from normal can be more readily identifiable and prevent misinterpretation of the MR spectroscopy.

METHOD AND MATERIALS

MR spectra from 60 children ranging in age from 0 to 200 months were obtained. Patients had a normal brain MRI. Quality control measures were applied and suboptimal spectra due to various artifacts were excluded. The spectra were analyzed on a standard processing platform and quantitative analysis was also performed using LCModel software. Metabolite values across the age of rage were evaluated. The major metabolites included were n-acetyl-aspartate (NAA), choline (Cho), creatine(Cr), myoinositol (mI), and lactate (L). Minor metabolites were also evaluated if the spectra were of high quality. Average values and confidence intervals were calculated across the age range. Non-linear regression models for assessment of changes over time were constructed for each metabolite.

RESULTS

Plots of change in the major metabolites across the early years of life were depicted with confidence intervals. There was a predicted increase in NAA and a decrease in Cho. Very small amounts of lactate were present normally. Changes in other metabolites were also depicted. A general estimation of the changes in the metabolites could be derived.

CONCLUSION

MRS values change dramatically in the first few years of life. Knowing the normal values in the early stages of life is challenging. Having a set of normal values with an indication of the normal variation based on a young patient's age is very helpful to the interpreting physician using MR spectroscopy in selected patients.

CLINICAL RELEVANCE/APPLICATION

Normative quantitative brain MRS metabolites is of paramount importance in the evaluation of neurodegenerative processes in children.

Background: 

“SyMRI“ generates different MR contrasts and characterizes tissue properties based on a single acquisition of a Multi-Dynamic Multi-Echo sequence (MDME). The aim of this study was to assess the diagnostic accuracy of T1- and T2-weighted contrasts, generated by “SyMRI“ for neonatal brain imaging. Additionally, the feasibility of “SyMRI“ in the assessment of myelination in term born and preterm neonates was assessed.

Methods: 

32 term born and preterm neonates (16/32 inconspicuous cases) were examined at the approximate due date [gestational age + days to MRI = corrected gestational age (CGA)] using a standardized neonatal MRI protocol (1.5 Tesla, T1 SE, T2 SE, DWI, SWI, T1 3D sequence). MDME sequence (FOV: 200 x 165 x 109 mm, Voxel: 0.9 x 1 x 4 mm, Matrix: 224 x 158 x 22 slices, TE: 13 ms, TR: 3309 ms, acquisition time: 5 min. 22 sec.)-based post-processing was performed using “SyMRI“ (Synthetic MR AB, Storgatan 11, Linköping, SWEDEN. Version 8.0.4). Two raters independently assessed neonatal brain examinations on T1-/T2-weighted images, generated by “SyMRI“ and on standard T1-/T2-weighted images, in order to determine the presence or absence of lesions. The sensitivity and specificity of both methods were calculated. 

Myelination was assessed by scoring seven brain regions on quantitative maps, generated by “SyMRI“ and on standard T1-/T2-weighted images, acquired seperately. Analysis of covariance (ANCOVA, covariate: CGA at MRI) was used for group comparison.

Results: 

The sensitivity and specificity of conventional and “SyMRI“-generated MR contrasts was calculated on the basis of 16 pathological cases and 16 inconspicuous cases. Both methods did not show any differences [sensitivity: 0.94, confidence interval (CI): 0.72 – 0.99); specificity: 1, CI: 0.81 – 1)].

In 25/30 (83.33%) cases [18 preterm (mean gestational age: 178.17 d, ± 12.56) and 7 term born neonates (mean gestational age: 279.14 d, ± 8.86)] myelination assessment could be performed. 

Quantitative T1-/T2-maps generated by “SyMRI“: ANCOVA results showed significantly lower myelination scores in preterm compared to term born neonates [T1: F(1, 22) = 7.420, p = 0.012; T2: F(1, 22) = 5.658, p = 0.026]. The myelination score showed a positive correlation with the CGA at MRI (T1: r = 0.662, n = 25, p ≤ 0.001; T2: r = 0.676, n = 25, p ≤ 0.001). 

The myelination score based on standard T1-/T2-weighted images did not correlate with the CGA at MRI. No significant differences between preterm and term born neonates were detectable.

Conclusions: 

T1-/T2-weighted images, generated by “SyMRI“ showed the same diagnostic accuracy as conventionally acquired T1-/T2-weighted contrasts.

Substantial differences concerning the myelination in term born and preterm neonates are detectable on quantitative T1-/T2-maps, generated by “SyMRI“. The method allows to detect myelination differences more sensitively than standard MR sequences. 

In addition to semi-quantitative imaging data, “SyMRI“ provides diagnostic images and leads to a more efficient use of available imaging time in neonatal brain MRI.

Introduction

Monitoring brain maturation, neuronal proliferation, neural migration and myelination using magnetic resonance imaging (MRI) has gained interest over the past years for routine follow-up of various pediatric diseases. To overcome complex and subjective assessments of morphology, automated brain segmentation tools have been optimized for this cohort to allow quantification of global or regional volumes in an acceptable timeframe. The establishment of normal development trajectories has thus become crucial for the desired routine use of such tools. Furthermore, since the compilation of a representative healthy pediatric neuroimaging database is challenging, pooling of data from different hardware settings is needed. In this work, volumetric measures of 61 brain structures were automatically extracted from two independent cross-sectional cohorts of 2- to 16-year-old healthy subjects scanned at 1.5T or 3T, respectively. We further tested whether field strength systematically biases the estimation of normative ranges.

Material and methods

Two independent cohorts of healthy individuals between 2 and 16 years old were included in this cross-sectional study (demographic details can be found in Figure 1). Clinical indications for inclusion were headaches, seizures, or minor neurologic disabilities. The two cohorts of 30 and 34 subjects were scanned at 1.5T (MAGNETOM Aera, Siemens Healthcare, Erlangen, Germany) and at 3T (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany), respectively, using a 20-channel head coil without sedation. Whole-brain 3D T1-weighted images were obtained with the magnetization prepared rapid gradient echo (MP-RAGE) sequence using acquisition parameters tailored to the respective field strength (see parameters in Table 1) and were all reviewed by a specialist in pediatric neuroradiology. Automated brain segmentation was performed using the MorphoBox prototype software optimized for pediatric applications. Reference ranges accounting for the normal evolution of brain volumes (V) with age were established for each of the 61 brain regions using a mixed-effects model (fixed effect: age; random effect: field strength):

E{V_j} = (B₀+u_0,j)+B_agelog(age)

with j=(1,2) referring to the two field strengths (1.5T and 3T), B being the coefficients of the fixed effects, and u_j the coefficients of the random effect. The goodness of fit of the mixed-effects model was compared against the goodness of the same model without the random effect of the field strength with a likelihood-ratio test to verify whether the bias between the two field strengths was statistically significant. P-values smaller than 0.05 were considered significant after Bonferroni’s correction for multiple comparisons.

Results

Brain segmentation was successful in all 64 normal subjects (see examples in Figure 2). No systematic bias between 1.5T and 3T normative ranges was found in 58 out of the 61 structures. Significant differences were found in only three brain structures: total brain (gray matter (GM) + white matter) volume (bias = 0.074 ml, p=0.003); global GM (bias = 11.94 ml, p=0.005); cortical GM (bias = 2.01 ml, p=0.008). Although being statistically significant, the observed bias is very small, and the resulting evolution of normative ranges, especially in brain and cortical GM volumes, can be practically considered comparable (Figure 3).

Discussion and conclusion

Evolution of brain growth with age was remarkably consistent between the two cohorts and independent from the field strength for most of the brain structures. The small biases observed could be explained by slight differences in in-plane resolution which induces more partial-volume effects at 1.5T compared to 3T and could be corrected by compensatory algorithms. Overall, the findings of this study encourage the pooling of data from different field strengths to enrich pediatric databases. Resulting normative brain volume growth models would provide a very valuable tool for clinicians and researchers to objectively assess abnormal brain development. We nevertheless anticipate that this still requires research to establish consensus normal trajectories.

Background: Children with pre- and perinatal brain damage will, when the motor pathways are affected, develop cerebral palsy (CP) or, when the visual pathways are affected, have cerebral visual impairment (CVI). Children with lesions affecting the motor and the visual pathways will suffer from CP and CVI. In individuals with focal injuries to the optic radiation (OR), detected with diffusion-weighted MRI (dMRI) and fiber tractography, topographically matched thinning of the macular ganglion cell layer (GCL) can be observed with optical coherence tomography (OCT), representing secondary injuries from retrograde trans-synaptic degeneration (RTSD). The focal thinning of the GCL is associated with corresponding visual field (VF) defects. In early brain injuries, the true lesional extent is difficult to judge on conventional MRI, e.g. involvement of the OR, and fiber tractography is rarely performed in the clinical routine. OCT is a relatively simple examination that can be performed in young children. The combination of homonymous VF defects and matching thinning of the retinal GCL could indicate retro-geniculate injuries, e.g. in individuals with a high risk of CVI, such as spastic CP.

Aims: To investigate the correspondence between primary injuries to the OR and VF function in individuals with spastic CP. Thereafter to study the possibility to use OCT and GCL topography to identify retinal RTSD and to predict VF function. 

Methods: Among the participants in three studies on spastic CP in our institution, twelve were invited to this study. Ten individuals (six females, median age 21 years, range 17 – 38) completed all examinations. None had any known ocular disease besides possible axonal loss related to brain injury. MRI data, including dMRI, were collected from the previous scanning protocols. The OR was reconstructed with probabilistic constrained spherical-deconvolution fibre tractography, and assessed by observation in relation to the expected topography. Cases where the number of streamlines/streamline densities in the whole or parts of the OR tracts, were clearly fewer/lower than expected were considered abnormal.  Perimetry was performed with Humphrey Field Analyser, and focal VF defects and the VF index (a global metric that represents the entire VF based on the pattern deviation) were noted. The macular ganglion cell and inner plexiform layers (GCL+IPL) was measured with OCT. The GCL+IPL thickness is analysed in six pie chart sectors centred in the macula, which are normally comparable in thickness. Focal thinning was noted as abnormal thin minimum GCL+IPL sector thickness/es and asymmetry as the difference between the thickest and thinnest GCL+IPL sectors. 

Results: The MRI lesion patterns included malformations, white-matter damage of immaturity and focal infarcts. In six cases (1-6) fiber tractography revealed different lesional involvement of the ORs. In these cases, the OCT showed focal thinning of the topographically corresponding parts of the GCL+IPL. The focal thinning correlated, indeed, with homonymous VF field defects, either hemifield defects or uni- or bilateral quadrantanopias. The relationship is illustrated for case 1 in the figure. Four cases (7-10) had normal ORs on fibre tractography. One these cases (7) had a generally reduced GCL+IPL thickness and no asymmetry, a general reduction of VF sensitivity but no focal VF defects. Another case (8) showed a similar reduction of the VF sensitivity without focal VF defects and had normal minimum GCL+IPL thickness and low asymmetry. Two cases (9-10) presented with normal GCL topography and normal VF sensitivity.

Conclusion:  This multiple-case study confirms the topographical correspondence between injuries to OR and secondary thinning of the GCL+IPL, and corresponding focal VF defects. The fact that OCT can distinguish retinal RTSD in the individuals with lesions affecting the OR, makes the method a feasible screening tool in individuals with spastic CP.

Germinal matrix hemorrhages typically occur postnatally in very preterm neonates (gestational age earlier than 28 weeks). Nonetheless, they may be observed in late preterm and term neonates, supposedly due to prenatal origin. In this clinical scenario, its real incidence is still unknown. We searched our database and teaching files for key words similar to “grade IV germinal matrix hemorrhages” and “term neonates”. Afterwards, we reviewed clinical data, in order to confirm gestational age and then looked for prenatal exams. We eventually came up with a series of 6 cases of late preterm or fully term neonates (gestational age at birth from 36 to 40 weeks), who presented to MR imaging at various ages (from 21 days old to 18 months old), under investigation for ventriculomegaly, motor delay, hemiparesis or cerebral palsy, who showed variable degrees of sequelae of grade IV germinal matrix hemorrhage. Prenatal imaging exams were available in some cases, which retrospectively were considered abnormal. We briefly discuss and compare imaging findings and then review some of the literature.

The drugs can cause many alterations affecting physical and mental health in humans. In youth, the risk is multiple, drug addiction can cause cognitive distortions that result in behavioral disorders and cognitive impairment. Materials & Methods: We performed a randomized case-control analytical study N = 115 into two groups (aged between 10 and 18 years); G1 = 55 minors addicted to drugs recruited at the Addictology Center and the Child Protection Center; and the control group G2 = 60 non-addicted minors recruited from two schools in Marrakech. we have respected the anonymity, confidentiality of participants. For the evaluation of congnitive functions we used two tests: Bel test to evaluate visuo-attention deficits, Unilateral spatial neglect and organizational strategies; and Digit Span test (DS) is used as a test of working memory. Results: We found that 73% of the addicts represent omissions (> 6 bells). Three types of hemianopsiadominate in G1: right and left spatial visuo-neglect (36% vs 17%), right spatial visuo-neglect (24% vs 13%), and central spatial visual neglect (4% vs 2%). DS Test (direct order) shows that short-term verbal memory skills are most affected in the G1(11% high school, 5% CM2, 4% CM3 and 1% CM1), in (inverse order): deficits in G1 (24% of high school, 10% CM3, 8% CM2 and 2% CM1). Conclusion: the evaluation of drug effects on neuro-cognitive abilities in young people has shown cognitive impairment; the young age, the chronic character and the frequency of consumption are elements of gravity of the addiction.