key: cord-0050667-wtnzseeb
authors: nan
title: ESNR 2020
date: 2020-09-25
journal: Neuroradiology
DOI: 10.1007/s00234-020-02534-5
sha: 105b43d2239f53afaeeff33bd6fb74719b9f692a
doc_id: 50667
cord_uid: wtnzseeb

nan

The ECNR is the fundamental Neuroradiology course aimed at neuroradiologists, established or in training. It is based on cycles of four courses (modules), each lasting five days, dedicated to diagnostic and interventional neuroradiology. The full cycle is considered complete after the attendee has participated in all four modules, which can be done in a single cycle or in different cycles. The scientific content of the course is determined by the ESONR Committee of the European Society of Neuroradiology, taking into account international standards and guidelines for training in diagnostic and interventional neuroradiology. The participating educators will be internationally renowned European experts, to be selected on the basis of their scientific background and educational skills to ensure high-quality lectures and interactive case discussions.

The following topics have been chosen, each to be covered in five full days of lectures and workshops:

• Anatomy, congenital malformations and genetics. Introduction This study validated a previously described diffusion-MRI phenotype as a potential predictive imaging biomarker in patients with recurrent glioblastoma receiving bevacizumab (BEV).

A total of 396/596 patients (66%) from the prospective randomized phase II/III EORTC-26101 trial (with n=242 in the BEV and n=154 in the non-BEV arm) met the inclusion criteria with availability of anatomical and diffusion MRIsequences at baseline prior treatment. Apparent diffusion coefficient (ADC) histograms from the contrast-enhancing tumor volume were fitted to a double Gaussian distribution and the mean of the lower curve (ADClow) was used for further analysis. The predictive ability of ADClow was assessed with biomarker threshold models and multivariable Cox-regression for overall and progression-free survival (OS, PFS).

Huntington's disease (HD) is a neurodegenerative disease with reported iron accumulation in specific brain regions degenerated in HD. Given the primary atrophy of basal ganglia in Huntington, altered iron is expected in the anterior thalamic radiation (ATR), a projection tract which is part of the cortico-basal ganglia-thalamocortical loop. However, the significance of this iron disruption and the impact to disease progression are currently unclear. In this study, we aimed to characterize the iron levels along the ATR and its relation with microstructural and metabolic properties and with clinical progression in HD.

Thirty-one HD individuals (premanifest (N=12) and manifest (N=19)) and 24 healthy controls were scanned at 3T MRI scanner. Relaxometry T2*-weighted (iron sensitive) imaging, diffusion -tensor imaging (microstructural integrity) and spectroscopy (metabolites profile) were used to characterize the ATR in HD and controls. Cross-sectional analyses included group-level comparisons and correlation studies.

Neuroradiology (2020) S12 62 :S1-107 (Suppl 1) S Result HD patients exhibited increased mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) in both ATR, with a gradient from subcortical (more affected) to deep white matter (WM) (less affected). Both ATR of premanifest individuals showed high iron levels, regionally localized in the more inferior portion, and increased AD. Indeed, AD negatively correlated with iron levels. Moreover, increases in iron levels and loss of integrity in both ATR were significantly associated with HD progression, as measured by CAP and UHDRS functional and cognitive scores. Lastly, metabolically, HD patients displayed diminished neuronal viability (NAA) and reduced cell energy (creatine) compared to premanifest individuals and controls. Moreover, in HD carriers, glutamate positively correlated with iron levels and conversely creatine negatively correlated with MD and AD.

The iron increase in premanifest individuals may reflect an increase in oligodendrocytes in an attempt to actively repair ATR. Iron was correlated with glutamate and the increase in excitatory activity (glutamate), probably as a compensatory mechanism, is detrimental for neurons by excitotoxicity, which in turn may induce an iron accumulation enhancing toxicity by ferroptosis. Therefore, this early iron increase in both ATR may cause a high level of toxicity, which may contribute in HD patients with disease progression to further extensive and severe WM disintegration, with a gradient of damage that supports the dying back hypothesis in which neuronal death begins in the axon, and to a reduction of NAA and creatine.

Neuroradiology S13 (2020) 62 :S1-107 (Suppl 1) S

The histological composition of thrombi is a potentially important factor in the management of patients with acute stroke as it might influence pharmacological and/or interventional treatment strategies. Mechanical characteristics of the occluding thrombi are primarily depending on the proportion between fibrine and red blood cells (RBC). Spectral CT imaging with dual-energy technology enables tissue differentiation depending on the keV specific absorption. We sought to evaluate the feasibility and the potential accuracy of spectral CT in the imaging of thrombus composition in an In-Vitro-Model using blood samples with defined fibrin/RBC-ratios.

Five types of defined ovine blood thrombi with gradually increasing amount of RBC (0% -100%) were scanned with spectral CT imaging (CT-Revolution, GE healthcare Germany, Solingen),using keV-values between 40 and 140 keV. The density of the thrombi was quantified in Hounsfield units (HU) . Subsequently, the specific absorption curve was established for each thrombus and statistically correlated using a Kolmogorov-Smirnov test.

The absorption curves showed a statistically significant difference for thrombi with moderate to high levels of RBC, (30 -100% RBC fraction, p-value < 0.05) with increasing differentiability at keVlevels above 80 keV. Thrombi with a RBC content of more than 40% could be differentiated best whereas thrombi with a minimum RBC content (0-5%) could not reliably differentiated from thrombi with only a little RBC content (10-40%).

Spectral CT-imaging can differentiate thrombi with a defined proportion of RBC and fibrin. Best discrimination of the histological composition is achieved in thrombi with a RBC-content of more than 40% and at high keV-levels with a threshold above 80keV. The available data allow for a potential further development of the CT-technique for future studies, in which stroke treatment might take into account the histological nature of the thrombus.

Neuroradiology (2020) S14 62 :S1-107 (Suppl 1) S

To describe the neurological manifestations and neuroimaging features of hospitalized COVID-19 patients.

In this retrospective, multicenter study from Italy, we reviewed the electronic medical records and imaging of hospitalized COVID-19 patients from February 29, 2020 to April 4, 2020. Our inclusion criteria included patients with acute neurological manifestations requiring neuroimaging.

A total of 725 COVID-19 patients were reviewed. Out of these, 108 (14.9%) fit our inclusion criteria. A total of 107 brain CT, 20 brain MRI, 17 head and neck CT angiography (CTA) and 3 whole spine MRI were reviewed. The commonest neurological manifestations were altered mental status 64 (59.3%), acute cerebrovascular disease 34 (31.5%), headache 13 (12.0%) and epilepsy 10 (9.2%). The neuroimaging hallmark of these patients were acute ischemic 34 (31.5%) and hemorrhagic cerebrovascular disease 6 (5.6%). Additional neuroimaging findings included acute encephalopathy, posterior reversible encephalopathy syndrome (PRES), hypoxic-ischemic encephalopathy (HIE), cerebral venous thrombosis, cranial nerves and caudal equina enhancement such as Guillain-Barré syndrome and variants, exacerbation of demyelinating disease and nonspecific encephalitis. There was a statistical significant association (P<0.006) between the prevalence of altered mental status and age of patient.

Neurological complications with associated abnormal neuroimaging findings are not uncommon in Neurologists and neuro-radiologists should be aware of the broad-spectrum of neuroimaging and neurologic disease that can be associated with COVID-19.

Please note: Most of this work was published in Journal Radiology https://doi. org/10.1148/radiol.2020201933 (10) 7/108 (6) 4/108 (4) 3/108 (3) 1/108 (1) Intracranial hemorrhages Large Small Subarachnoid 6 /108(6) 2/108 (2) 1/108 (1) 3/108 (3) Enhancement (MRI with and without IV contrast) Cranial nerves* Cauda equina* Leptomeningeal Parenchymal 1/10 (10 ) 2/10 (20 ) 0/10 (0) 0/10(0) Acute encephalopathy † 1/20 (5) PRES 1/20 (5) Nonspecific encephalopathy 2/20 (10) MS plaque exacerbation ‡ 2/20(10) T2/FLAIR Signal Hyperintensity Nonspecific but likely chronic white matter disease Basal ganglia Subcortical 7/20 (35) 2/20 (10) 3/20 (15) Cerebral venous thrombosis (CTA) 2/17 (12) Neuroradiology S15 (2020) 62 :S1-107 (Suppl 1) S Numbers in Parentheses are Percentages. * Miller-Fisher syndrome, a regional variant of Guillain-Barré syndrome. A 62 year-old-man presented with bilateral facial nerve palsy, ophthalmoplegia, areflexia and polyradiculopathy, rRT-PCR assay of the CSF was negative for SARS-CoV-2. † Acute encephalopathy. A 60 year-old-man without history of seizures presenting with first time convulsion (Figure 2 ), rRT-PCR assay of the CSF was negative for SARS-CoV-2. ‡ Exacerbation of Multiple Sclerosis (MS). A 53 year-old-woman presented with seizures and altered mental status. 

To investigate impaired perfusion as a potential mechanism for white matter hyperintensities (WMH) in patients with carotid artery stenosis (CAS), using advanced Magnetic Resonance (MR) imaging, including pseudocontinuous arterial spin labelling (pCASL) perfusion.

Fifty-two patients were recruited from an ongoing prospective clinical trial, the second European Carotid Surgery Trial (ECST-2). MR images were analysed, blinded to clinical information, for the degree of CAS, circle of Willis collaterals and presence and severity of arterial transit artefacts (ATAs) on pCASL. ATAs were rated according to the following scale: 0, no or minimal ASL signal; 1, moderate ASL signal with ATAs; 2, high ASL signal with ATAs; and 3 normal perfusion without ATAs. Each score was assigned to 34 brain regions per patient adapted from ASPECTS criteria, with a total of 1768 anatomic sites evaluated across all subjects. Automated WMH segmentation was performed and compared with visual rating scores. All images were subjected to rigorous visual quality control.

The median age of the study population was 73 years [IQR: 65.5 -80]) and 35 (67.3%) patients were men. ATAs were more common in patients with 70% or greater CAS (p=0.0001) and were associated with the absence of primary collaterals on the main stenotic side (p=0.022). Symptoms were more frequent among patients with ATAs than those without (58.3% versus 25%, respectively; p=0.023). Patients with ATAs had greater total WMH lesion volume, with lesions distributed mainly on the outer layers of WM (L2, L3 and L4; p= 0.024; p=0.011 and p=0.019, respectively). The hemispheric asymmetry of WMH lesion volume was more prominent at the juxtacortical layer (L4) in patients with ATAs than those without (ΔWMHL4 248.565 mm3 versus ΔWMHL4 77.49 mm3, respectively; p=0.008). However, no significant match was found between the hemisphere with greater WMH lesion volume and the side most affected by ATAs at both global and regional WM analysis.

Although ATAs were more common among patients with severe CAS and greater WMH burden, a significant link between this surrogate marker of impaired perfusion and WMH lesion volume was not established.

Disease and illnesses have plagued humanity since the earliest days. Over time there have been numerous pandemics, some with particularly high mortality rates: The Plague (14th century), Spanish Influenza (20th century), HIV/AIDS (20th century), The Antonine Plague (6th century) are among the deadliest ones. The current pandemic, COVID-19, is the third coronavirus outbreak in the past two decades, and was its rapid spread made COVID-19 so damaging although it doesn't have the death toll of the previous. COVID-19 is disseminated primarily via respiratory droplets and has the potential to cause severe respiratory distress, resulting in pneumonia, acute respiratory distress syndrome (ARDS), multiorgan dysfunction, and death. The neuroinvasive potential either as a result of direct viral replication in the CNS either or by an overactive host immune response has been described. To analyze the importance of Neurorradiologgy the authors reviewed scientific papers published on the involvement of COVID-19 in the CNS and correlated with Portugal multicentric experience.

Bibliographical research in the databases PubMed, UpToDate, and websites of the European Society of Neuroradiology, American Journal of Neuroradiology and the American Academy of Neurology keywords: COVID-19, coronavirus, central nervous system, neurology, neuroimaging. Review of reports from different national institutions and neuroradiology departments and correlation with clinical data.

The new coronavirus SARS-CoV-2 has a high tropism for the CNS and has a variable manifestation both in terms of the clinical presentation and the severity of it. Most COVID-19 patients don't have any neurological symptoms, however, one of the first reported symptoms was anosmia. Furthermore, many recent articles expose other possible presentations such as stroke, vasculitislike pattern, cerebral venous thrombosis, hemorrhagic venous infarction, Miller Fisher, and Guillain-Barre syndrome, ADEM or even acute necrotizing encephalopathy.

The COVID-19 pandemic is still an ongoing disease and it is impossible to predict its future impact. This review is meant to provide a quick and complete way to access its various neurological presentations. Introduction Condylar venous dural arteriovenous shunts (CdDAVs) are rare fistulas location often requiring complex treatment strategies.

Our aim was to report the role of transvenous reverse pressure cooker embolization technique using coils and DMSO-based polymers and present a systematic literature review on the management of these lesions. We reviewed all CdDAVs treated by transvenous reverse pressure cooker approach in our center between 2008 and 2019. We describe technical details, perioperative morbidity and outcomes of twelve patients.

All patients had anatomical and clinical cure. Two patients developed transient hypoglossal nerve palsy; no other complications were reported. A systematic literature review was performed; we found 128 reported similar patients: 83 patients were treated through transvenous approach, 17 by transvenous and transarterial, 19 by transarterial embolization, two by transarterial embolization and surgery; seven patients were treated conservatively. Treatment morbidity was 8% (more common for transvenous embolization) and mortality was 0%.

Transvenous reverse pressure cooker approach is safe and effective for the treatment of CdDAVs. Modern imaging techniques like high resolution VasoCT are very helpful to accurately diagnose and identify the fistulous site and to plan treatment. Aggressive symptoms of this type of fistula depend mainly on the presence/absence of a bridging vein and on the changes of the drainage pattern. Neuroradiology S17 (2020) 62 :S1-107 (Suppl 1) S Introduction: We aimed to compare the diagnostic performance to detect facial neuritis of enhanced compressed-sensing volume-interpolated breath-hold examination (CS-VIBE) compared with 3D T1 magnetization-prepared rapid-acquisition gradient-echo (MPRAGE).

Method: Institutional review board approval was obtained and informed consent was waived for this retrospective study. Between February 2019 to September 2019, 30 patients who had undergone enhanced temporal MR for the evaluation of facial palsy with both conventional MPRAGE (scan time: 6 minutes 8 seconds) and CS-VIBE (scan time: 2 minutes 48 seconds) were included. All images were independently reviewed by three radiologists for presence of facial neuritis, enhancement degree of facial nerve (canalicular, labyrinthine, anterior genu, tympanic and mastoid segment) by visual analysis. We also accessed signal to signal-to-noise ratio (SNR) of pons and contrast-to-noise ratio (CNR CSF-nerve) for three segments (canalicular, labyrinthine, and anterior genu) of the facial nerve. Overall image quality, artifact and facial nerve discrimination also visually analyzed. Sensitivity and specificity of both sequences were calculated with the reference of clinical diagnosis. Kruskal-Wallis test were used to compare SNR, CNR and image qualities.

Result: For the detection of facial neuritis, CS-VIBE achieved 97.8 % of sensitivity and 98.9 % specificity and conventional MPRAGE demonstrated 100.0% sensitivity and 98.9% specificity in pooled analysis. CS-VIBE show comparable enhancement degree in canalicular segment and labyrinthine segment (p-value > 0.05) and significantly higher enhancement degree in anterior genu, tympanic segment and mastoid segment (p-value < 0.001 in anterior genu and tympanic segment and p-value = 0.001 in mastoid segment) in overall analysis. SNR of pons was significantly lower in CS-VIBE, compared to conventional MPRAGE (p-value < 0.001 for all). However, in analysis of CNRCSF-nerve, CS-VIBE show significant higher in all analyzed segment regardless of observers and affected side (p-value < 0.05 for all). CS-VIBE also show better overall image quality, artifact and facial nerve discrimination compared to conventional MRPRAGE (p-value < 0.001 in overall analysis for all).

Conclusion: CS-VIBE achieved comparable diagnostic performance for the detection of facial neuritis compared to conventional MPRAGE, and could be acquired within half of the scan time. A 77-year-old female who had right facial palsy (House Brackmann grade 4) a day ago perform enhanced temporal MR. The image quality of conventional MPRAGE (A) is severely degraded because of severe motion artifact during MR scanning, even though enhancement of canalicular segment (arrow) and anterior genu (arrowhead) lead to the diagnosis of facial neuritis. However, labyrinthine segment (curved arrow) shows severe blurring. In contrast, CS-VIBE (B) is welldemonstrated canalicular (arrow), labyrinthine segment (curved arrow) and anterior genu (arrowhead) without blurring.

* Abbreviations: CMB, cerebral microbleeds 2.50 ± 0.62, 3 (2-3) 2.62 ± 0.49, 3 (2-3) 0.07 Disease (+) side 2.87 ± 0.34, 3 (3-3) 2.98 ± 0.15, 2 (2-3) 0.002 Disease (-) side 1.92 ± 0.50, 2 (2-2) 2.17 ± 0.40, 2 (2-2) <0.001 Tympanic Segment Overall 1.79 ± 0.54, 2 (1-2) 2.10 ± 0.44, 2 (2-2) <0.001 Observer 1

1.75 ± 0.47, 2 (1-2) 1.98 ± 0.43, 2 (2-2) 0.002 Observer 2

1.68 ± 0.47, 2 (1-2) 2.07 ± 0.36, 2 (2-2) <0.001 Observer 3

1.93 ± 0.63, 2 (2-2) 2.25 ± 0.47, 2 (2-3) <0.001 Disease (+) side 2.01 ± 0.46, 2 (2-2) 2.27 ± 0.49, 2 (2-3) <0.001 Disease (-) side 1.57 ± 0.52, 1 (1-2) 1.93 ± 0.29, 2 (2-2) <0.001 Mastoid Segment Overall 1.92 ± 0.44, 2 (2-2) 2.03 ± 0.35, 2 (2-2) 0.001 Observer 1

1.95 ± 0.47, 2 (2-2) 2.05 ± 0.39, 2 (2-2) 0.17 Observer 2

1.78 ± 0.45, 2 (2-2) 1.93 ± 0.31, 2 (2-2) 0.003 Observer 3

2.03 ± 0.37, 2 (2-2) 2.12 ± 0.32, 2 (2-2) 0.13 Disease (+) side 2.00 ± 0.45, 2 (2-2) 2.10 ± 0.40, 2 (2-2) 0.63 Disease (-) side 1.84 ± 0.42, 2 (2-2) 1.97 ± 0.28, 2 (2-2) 0.005

The Orbital multidisciplinary team meeting (MDT) serves as a forum for the discussion of difficult or complex cases predominantly from Oculoplastics, Skull base and ENT sinonasal teams. Our Orbital MDT takes place in the regional tertiary eye hospital treating patients from Greater Manchester in addition to referrals from other external units. It is held every 4 to 6 weeks with the imaging reviewed by a neuroradiologist with a specialist interest in Orbital-Head and Neck radiology. A snapshot view of a typical MDT within our institution is provided with interesting case images covering essential learning points for complex orbital radiology.

The findings will be presented in the form of a case-based pictorial review utilising a spatial approach.

The surgical management of orbital pathology such as orbital apex lesions or invasive neoplasms can be technically challenging and associated with significant risks to vision or ocular motility. CT and MR imaging aid the surgeon in delineating the anatomical extent of the presenting pathology. Input from a neuroradiologist in evaluating the distinctive imaging features of the lesion within the MDT, further helps narrow the differential diagnosis and positively impact on the subsequent treatment plan. In this presentation, we will review variety of pathologies including vascular malformations, neoplasms such as meningiomas, invasive orbital infection and inflammatory disorders such as

Computed Tomography (CT) of the paranasal sinus is a useful tool in the work-up of chronic rhinosinusitis as it provides important information regarding the extent of mucosal disease. The Lund-Mackay score system is the mainstay on CT evaluation of rhinosinusitis. This study aims to determine whether there is a correlation between symptoms and image findings in chronic rhinosinusitis and to assess the relevance of a CT score and staging system in these patients' management.

During one year (2019), we prospectively studied adult patients from the otorhinolaryngology practice that fulfilled clinical criteria for chronic rhinosinusitis. Patients were given a questionnaire on the day of the CT scan to assess their clinical condition at the time of the exam, using the Sino-Nasal Outcome Test (SNOT-22) scale validated to Portuguese population as well as some additional clinical data. CT scans were performed and prospectively analyzed by 2 neuroradiologists using the Lund-Mackay CT score (LMS) and the Kennedy staging system (KSS). We also collected information on normal variants. .20 for KSS). Nevertheless, in multiple linear regression, the presence of nasal obstruction and hyposmia were the symptoms that better predicted an increment on LMS. The presence of polyps also correlated well with LMS (β=11.9, p<0.001), but not with overall symptoms. Interestingly, we found that the obliteration of ethmoidal infundibulum and frontal recess showed a better correlation with headaches than the sinus obliteration.

CT scans yielded good detail on sinus involvement in rhinosinusitis. Currently used radiological scores were simple to use and easily reproducible but did not correlate well with clinical presentation. CT sinus grading systems need to be improved to better predict the symptoms.

Meniere's disease (MD), also known as idiopathic endolymphatic hydrops, is a disorder of the inner ear characterized by fluctuating hearing loss, occasional episodic vertigo, tinnitus and aural fullness. Pathophysiology of MD is controversial: endolymphatic hydrops is the marker of the disease, but it is questionable if it is causative or consequence. Increase in endolymphatic fluid in the scala media causes displacement of the Reissner membrane into the vestibular scala with increasead pressure on the organ of Corti and degeneration of nerve fibes and ganglion cells in the cochlea. Given the close anatomical and functional relation of modiolus with the membranous labyrinth and cochlear nerve fibers, a distortion of the bony structure of modiolus might be expected in the disease. The purpose of this study was to evaluate the cochlear modiolus size in patients with MD and in healthy subjects, using magnetic resonance imaging (MRI).

We performed a cross-sectional study comparing inner ears of patients with definite MD and healthy controls. The area of the cochlear modiolus was measured using transverse sections in which the cochlear modiolus was visualized at its maximum size on thin-section heavily T2-weighted three-dimensional fast spin-echo images by an experient radiologist, blinded to the clinical diagnosis. Differences between MD patients and controls were evaluated with the Mann-Whitney U test.

Results are presented as median (interquartile range [IQR]). A p-value lower than 0.05 was considered significant. Additionally, a second radiologist assessed modiolus areas and results were compared to assess inter-observer variability. Results MR images of 12 ears of 6 healthy asymptomatic controls and 18 ears of 9 patients with definite MD were obtained. The median area of the modiolus in healthy controls was 2.60 (2,52;2,95) and in MD patients was 3.62 (3.38;3.79), which was statistically significant (p< 0.001). Interobserver variability was less than 0.6 mm2 for all ears (range, 0.01-0.56 mm2; mean, 0.3 mm2 ±0.15).

Findings in this study suggest that MD is associated with modiolar changes, with increase of the area of this bony structure. This is possibly somehow related with reactive hiperostosis in the cochlea as a consequence of chronic increased pressure forces into cochlear nerve fibers and ganglion cells. As such, assessment of the modiolus area may be a useful imagiological sign in MD patients and our findings suggest a high inter-observer agreement, which favors its applicability in clinical practice. Further studies are needed to corroborate and explore these findings.

Previous studies in patients with Tourette syndrome (TS) or obsessivecompulsive disorder (OCD) have not clarified whether these two neurodevelopmental disorders represent clinical conditions due to different pathophysiological mechanisms or distinct clinical phenotypes of a common disease spectrum. Here, we used resting-state functional magnetic resonance imaging (rs-fMRI) to compare functional connectivity (FC) in cohorts of drug-naive pediatric patients with TS without any comorbidity (TS), TS patients with OCD (TS+OCD), and OCD patients.

Sixteen drug naïve patients with TS (15 males, mean age: 9.7 ± 2.1), 14 with TS+OCD (10 males, mean age: 10.2 ± 2.1), 10 with OCD (7 males, mean age: 10.9 ± 2.5) underwent 3T rs-fMRI. Eleven agematched children with episodic tension headache who were headache-free during the MRI scan (2 males, mean age: 9.9 ± 1.3 years) serves as controls. Tic severity and OCD symptoms were scored using the Yale Global Tic Severity Scale (YGTSS) and 

Overall, our findings suggest common FC changes in TS and TS+ OCD drug-naive pediatric patients. In contrast, OCD is characterized by a distinctive pattern of FC changes prominently involving the CBN and FPN. Pathophysiological mechanisms responsible for obsessive-compulsive symptoms in TS differ from those underlying pure OCD.

The complex interaction between genes, brain, and behaviour in language disorder is well established. Yet to date, the imaging literature in the language disorder field has continued to pursue heterogeneous and relatively small clinical cross-sectional samples of participants. In our current work, we aim to study the relationship between caudate volume, fractional anisotropy (FA) for caudate nucleus and language development at 24 months of age.

A total of 148 subjects were included in our study. All subjects had T2 weighted imaging while 109 subjects had both T2 weighted and diffusion weighted imaging. Region of interest (ROI) analysis of the FA maps was undertaken by a pediatric neuro-radiologist. ROIs were placed at the caudate heads for FA value quantification. The ratio of right-to-left caudate FA values and asymmetry quotient as a function of FA values (AQFA) were calculated for each subj ect. Neurodevelopmental outcomes of all subjects were assessed by certified examiners performing the BSID-III at 24-month. Regression analysis was performed to examine the effects of caudate volume and FA values on language development, with postmenstrual age at scan, gender, birth weight and maternal age as covariates.

There was a significant negative correlation between BSID-III language composite scores and left caudate FA (beta value = -0.3856, p < 0.01). There was also a significant positive correlation between BSID-III language composite scores and right caudate FA (beta value = 0.6661, p < 0.01). There was a significant positive correlation between BSID-III language composite scores and right-to-left caudate FA ratio (beta value = 0.7829, p < 0.01) as well as between BSID-III language composite scores and caudate asymmetry quotient for FA (AQvol) (beta value = 0.6738, p < 0.01). There was no significant correlation between BSID-III language composite scores and caudate volume.

Our study provides further confirmatory data on the neural correlates of childhood language disorder, in a large cohort of asian children imaged within the neonatal period, prior to the onset of environmental sources of variance. Our findings suggest that early neonatal caudate FA ratio and caudate asymmetry quotient for FA (AQvol) predict language development at 24 months and may potentially serve as a "heritable imaging marker" for language disorder. This allows the detection of children at-risk of developing impaired language development, allowing early intervention.

Neuroradiology S21 (2020) 62 :S1-107 (Suppl 1) S

Longitudinal volumetric changes of the pituitary and brain, in children on recombinant human growth hormone (rhGH) therapy might be seen on MRI. Objective: To assess volumetric changes in the pituitary gland, basal ganglia, corpus callosum, thalamus, hippocampus and amygdala in subject with isolated growth hormone deficiency (IGHD) and its response to treatment.

This is a longitudinal study of eight IGHD subjects (2 male, 6 female) with mean age of 10.7 ± 1 years and age-matched control group. Volume of pituitary gland, basal ganglia and limbic structures were obtained using 3T MRI voxel-based morphology. Left hand bone age was assessed using the Tanner-Whitehouse method. Follow up imaging was performed after an average of 1.8 ± 0.4 years on rhGH.

Subject with IGHD have smaller mean volume of pituitary gland, right thalamus, both hippocampus and both amygdala compared to control. After rhGH therapy, these volumes normalizes (as compared to age-matched control). The corpus callosum of IGHD subject have larger mean volume compared to the controls and do not show much volume changes in response to rhGH therapy. There were changes towards normalisation of bone age deficit of IGHD in response to rhGH therapy.

Pituitary gland, hippocampus and amygdala volume in IGHD subjects are smaller than age-matched control groups and showed the most response to rhGH therapy. Semi-automated volumetric assessment of pituitary gland, hippocampus and amygdala using MRI may provide objective assessment of response to rhGH therapy.

Head trauma is the most common delivery-related traumatic injury. Up to 46% of asymptomatic neonates show intracranial blood deposition, being subdural hematoma the most prevalent type. The aim of this study is to describe imaging findings attributable to birth-related traumatic head injury (BTHI) and its relation with known maternal, fetal and obstetric risk factors.

The study was approved by the institutional review board. All consecutive head CT and MRI studies performed during 2019, in neonates within the 5 first weeks of life, were reviewed by a neuroradiologist with 4 years experience and a neuroradiology resident. Imaging findings were anatomically categorized (eg.: extracranial, skull, intracranial). Neonates with non-traumatic hemorrhagic lesions and with post-natal trauma history were excluded. Medical records were retrospectively reviewed for selected risk factors, including method of delivery, maternal comorbidities, somatometric data and Apgar scores. Risk factors were assessed for relationship to traumatic head injury by using Fisher exact test, Mann Whitney U test and independent t-test.

Thirty term (n=22) and preterm neonates (19 male) underwent CT or MR imaging with a mean age at imaging=11,4 ± 1,6 days (0-35 days); only one was referred with suspected traumatic injury. Vaginal delivery occurred in 40% of neonates (all term, 5 vacuum-assisted). Twenty-one (70%) had presumed BTHI; 39 lesions were identifiedintracranial lesions (77%; included tentorial hemorrhage (n=15), supratentorial subdural hematoma (n=8), subdural retrocerebellar hematoma (n=6) vs extracranial (21%) vs cranial (2%). BTHI was more prevalent in term neonates (81 vs 19%) and significantly associated

Chronic adhesive arachnoiditis (CAA) is a complex disease process that can result in significant neurologic symptoms, severely compromising the quality of life of affected patients. CAA has been described as both a static and progressive disease, and its course remains highly variable. Suspected underlying causative factors include trauma, prior surgery, nontraumatic subarachnoid hemorrhage, meningitis, among others. Radiologic and pathologic features do not always correlate with clinical findings, and no reliable laboratory tests or electromyelography findings exist.

We performed a review of existing literature and report two rare cases from our institution database, with the clinical picture and spinal neuroimaging.

A man aged 56 years presented with 2-year progressive history of low back pain, paraparesis and gait disturbance, getting worse within last couple of months. Three years ago, he had a long hospital stay in the intensive care unit due to meningoencephalitis, and had a CSF culture specimen positive for Listeria monocytogenes. The second case, a 49 years man had a similar clinical presentation, namely 9-months of back pain, paraparesis and sensory myelopathy. The patient also had a history of meningoencephalitis with a CSF culture positive for Listeria monocytogenes, 10 years before. Both patients underwent an MRI study that confirmed chronic adhesive arachnoiditis, with multiple arachnoid septations affecting the cervical, thoracic, and lumbar segments. The arachnoid cysts had variable mass effect on the cord, and both showed lumbar nerve root clumping and enhancement. There were areas of cord atrophy, swelling and increased T2 signal, and focal foci of pial, dural and arachnoid enhancement. No enhancement of the spinal cord was observed.

To our knowledge, these are the third and fourth reports of delayed spinal complications following Listeria monocytogenes meningoencephalitis. Infectious etiologies of CAA include tuberculous meningites, pyogenic infections, cysticercosis, candida tropicalis, cryptococcosis, syphilis and listeriosis. The numerous causes of CAA, inconsistent time from insult to clinical presentation, and poorly defined and varying imaging characteristics may result in a missed or delayed diagnosis. The imaging appearance of interstitial edema and cord swelling may mimic the appearance of an underlying intramedullary tumor, leading to unnecessary and nondiagnostic cord biopsy. Recognition of the imaging characteristics of CAA is essential for making a correct diagnosis and avoiding unnecessary interventions.

Figure 2 

We retrospectively evaluated the conventional MR images of biopsy proven or newly diagnosed cases of PCNSL and GBM (that presented with solitary space occupying masses) at our Hospital in Dhahran, from Jan. 2019 to Jan. 2020. All MRIs were performed on a 1.5 T GE scanner, routine sequences with contrast studies were acquired with (single or multi-voxel) spectroscopy where possible. Tumor location, morphology, and specific MRI features and patterns were recorded, compared and presented. Already diagnosed and treated patients were excluded.

Primary CNS lymphomas were found T1 hypointense, T2 iso-to hypointense, vivid homogeneous gadolinium-enhancing lesion with restricted diffusion, exhibiting subependymal extension and crossing of the corpus callosum. These typical findings were observed in untreated non-immunocompromised patients (n=3). Primary CNS lymphomas in immunocompromised patients (HIV/AIDS or post renal transplant) showed more heterogeneous intensity, with central non-enhancement/necrosis and hemorrhage (n=2). MR spectroscopy showed large Choline peak with reduced NAA and in few cases lactate peak. All GBM lesions (n=4) showed heterogeneous contrast enhancement. These were found more commonly centrally necrotic, and more commonly demonstrating evidence of hemorrhage.

PCNSL and GBM have specific features, and imaging patterns. Recognition of these may help to differentiate PCNSL from GBM at the time of initial presentation. 

A total of 33 patients with glial brain tumors of supratentorial localization (Grade II -8, Grade III -11, Grade IV -15). All patients underwent intraoperative fluorescence diagnostics with chlorine E6 of the 2nd generation (1 mg / kg), and the contrast was assessed in preoperative MRI images in T1 mode and PET with methionine. The contrast assessment on MRI was carried out on a 4point scale (0 -no contrast (3 (9.1%) patients), 1 -focal contrast (10 (30.3%) patients), 2 -uniform contrast (8 (24.2 %) of patients), 3 -annular accumulation of contrast with hypodensity in the center of the tumor (12 (36.4%) patients)).

A comparative analysis of visible fluorescence with preoperative MRI with contrast (T1 mode) yielded a direct correlation (p> 0.05) from contrast accumulation on MRI and the degree of glioma anaplasia. The higher the degree of anaplasia, the more pronounced the contrast according to MRI and the brighter the visible intraoperative tumor fluorescence. The accumulative ability of contrast on MRI with gliomas of a low degree of anaplasia was most characteristic of oligodendrogliomas, gliomas with the presence of a cystic component. Thus, the sensitivity of the method for Grade II gliomas was 68.7%, specificity -60.5%, and for Grade III -IV gliomas -85.3% and 77.1%, respectively. A comparative analysis of PET with methionine and chlorin e6-mediated fluorescence yielded a direct correlation between the accumulation index, fluorescence intensity, and the degree of tumor anaplasia (p <0.05). A higher accumulation index (1.7-3.9) and brighter fluorescence (+++) were noted in glioblastoma (Grade IV). A low accumulation index (0.9) and very weak fluorescence (+) were observed with Grade II tumors. Thus, with low-grade gliomas, the sensitivity of PET was 88%, and for chlorin-mediated navigation, 79%. The boundaries of the tumor using various techniques ranged up to 7 mm.

The higher the degree of anaplasia, the more pronounced the contrast according to MRI and the brighter the visible intraoperative fluorescence of chlorin e6 in the tumor. A higher PET radiopharmaceutical accumulation index and brighter fluorescence were noted with Grade IV. A low accumulation index and very weak fluorescence were observed with Grade II tumors. Chlorin e6 and contrast on MRI are practically not visualized in the area of tumor necrosis, as well as a decrease in the activity of accumulation of the radiopharmaceutical according to PET + CT.

Pre-surgical language mapping with functional magnetic resonance imaging (fMRI) is routinely conducted to assist neurosurgeons in preventing damage to brain regions responsible for language. Functional differences exist between the monolingual versus the bilingual brain [1, 2] , whereas clinical fMRI tasks are typically conducted in a single language. The presence of a secondary language is a potential source of error in the inferred language map. Recently, a white paper established the gold-standard protocol for language fMRI, which includes a sentence completion and a silent word generation task [3] . However, the impact of bilingualism on the results of these gold-standard tasks is currently unexplored. Our objective was to test whether bilingualism influences the results of clinically-relevant tasks, requiring ad-hoc implementation.

We prospectively recruited sixteen healthy right-handed subjects (mean age 42.37 years; nine males, seven females) without any neurological history. Eight subjects were monolinguals (English-speaking) and eight bilinguals (native Spanish (L1), acquired English (L2)). All bilinguals were fluent in English. All subjects underwent task-based fMRI with the same 3T magnet and gold-standard language tasks (English and Spanish-translated). Starting from the active clusters on fMRI maps, we constructed pairwise correlation matrices and inferred the connectivity matrix of individual functional networks for monolinguals and bilinguals (divided into L1 and L2) [4] . We named the persistent functional architecture across subjects the 'common network'. Finally, we run k-core centrality measure to characterize any arising differences between subjects both individually and at group level [5] .

Our results showed a persistent sub-network "core" architecture consisting of Broca's area, pre-supplementary motor area, and premotor area, with the highest score on k-core centrality measures [6] . Moreover, Wernicke's area was engaged by the network core with variable extent across groups (8/8 bilingual subjects speaking Spanish, 6/8 monolingual subjects, and 4/8 bilingual subjects speaking English), with different k-core occupancies. The bilingual L2 group showed weaker connection strengths between core fROIs compared to the same task in L1.

Our results may influence fMRI task choice and interpretation, with possible impact on therapeutic planning. To optimize pre-operative language mapping for bilingual patients, we suggest to consider employing pre-operative language tasks in either L1 or both L1 and L2 to ensure a robust and accurate language map. This recommendation is based on the higher clinical task engagement of language processing systems related to L1 compared to L2, as evidenced by higher common network link weights and more frequent involvement of the "core" architecture.

[1] Seo, R., Stocco, A., and Prat, C. S. The bilingual language network: differential involvement of anterior cingulate, basal ganglia and prefrontal cortex in preparation, monitoring, and execution. Neuroimage 174, 44-56 (2018).

[2] Li, L., Emmorey, K., Feng, X., Lu, C., and Ding, G. Functional connectivity reveals which language the "control regions" control during bilingual production. [4] Del Ferraro, G., Moreno, A., Min, B., Morone, F., Pérez-Ramírez, Ú., Pérez-Cervera, L., Parra, L. C., Holodny, A., Canals, S., and Makse, H. A. Finding influential nodes for integration in brain networks using optimal percolation theory. Nat. Commun. 9, 2274 (2018).

[5] Morone, F., Del Ferraro, G., and Makse, H. A. The k-core as a predictor of structural collapse in mutualistic ecosystems. Nat Phys 15, 95 (2019).

[6] Li, Q., Del Ferraro, G., Pasquini, L., Peck, K. K., Makse, H. A., and Holodny, A. I. Core language brain network for fmri-language task used in clinical applications. Network Neuroscience (2019).

The image displays k-core centrality measure functioning: the network stability is tested by progressively removing links to highlight "core" network components which remain stable after multiple links removal. Fig. 2 -Network construction passages: we started building the functional language network from active clusters on fMRI maps (step 1), by computing a correlation matrix from BOLD signal time-series interdependence (step 2). Functional links were inferred by thresholding pair-wise correlations between pair of voxels with a penalization parameter set to optimize brain integration (all the clusters are connected) and sparsity (minimal wiring), (step 3). The product of the workflow is a representation of the language network activated by the employed task (step 4).

show a persistent sub-network "core" architecture consisting of Broca's area, presupplementary motor area, and premotor area, with the highest score on k-core centrality measures. The language "core" component is present in every subject; however, bilinguals speaking English (acquired language, image b on the top) show less stable connection between Wernicke's area and the network "core" with respect to monolinguals (a) and bilinguals speaking their native language (image b on the bottom). Fig. 4 -The difference in connectivity between monolinguals and bilinguals is well depicted by K-core diagrams: bilingual patients speaking their native language (B) display strong connection between Wernicke's area and the network "core", with similar k-core occupancy of monolinguals (A). This result may suggest to perform pre-operative language tasks in either L1 or both L1 and L2 to ensure a robust and accurate language map in bilinguals.

Neuroradiology S25 (2020) 62 :S1-107 (Suppl 1) S

Cerebral venous thrombosis is a recognized complication in hematological malignancies like acute lymphoblastic leukemia and acute lymphoblastic lymphoma (ALL). While it may be related to the disease itself by activation of blood coagulation via procoagulant substances or by impairment of fibrinolytic or anticoagulant pathways, it may also be related to the chemotherapy instituted, particularly Lasparaginase.

We present four selected cases from our Department's Neuroradiology database, with illustrative CT and MRI imaging showing sinovenous thrombosis (CSVT) following therapy with L-asparaginase in patients suffering from hematological malignancies.

L-asparaginase is an enzyme that catalyzes the conversion of L-asparagine, necessary for lymphoblasts and other neoplastic cells, decreasing it's plasmatic concentration, leading to inhibition of RNA and DNA synthesis with the subsequent blastic cell apoptosis. The incidence of thrombotic complications during L-asparaginase therapy varies from 1% to 37%, being cerebral venous thrombosis its leading form of presentation, reported in 1-3% of treated patients. Clinical features of headache, altered consciousness, focal neurological deficit, and seizures developing during or immediately after treatment with Lasparaginase should alert the treating physician and prompt to a correct evaluation of the patient.

The key management of this condition is early diagnosis by imaging as delayed institution of anticoagulation may in some cases be futile. Rapid replenishment of coagulation factors may be achieved with fresh frozen plasma although antithrombin concentrates are preferred for this. For therapeutic anticoagulation, low molecular weight heparin is given initially and this may be continued or it may be substituted by oral anticoagulants for 3-6 months, leaving endovascular treatment as last resort gateway. We conclude that, diagnosis of CSVT in ALL patients being treated with L-asparaginase requires a high index of clinical suspicion in the presence of seizures, focal neurological deficit, and features of raised intracranial pressure. Early diagnosis demands a low threshold for imaging, and MRI should complement CT study. Identification of relevant findings such as venous infarcts, the empty delta sign, and absent flow in the dural sinuses on CT and MR venography enables proper diagnosis and management. Differentiating between glioblastomas and primary central nervous system lymphomas (PCNSLs) were limited to tumor entity and the knowledge on peritumoral edema is critically lacking. This study aimed to develop a machine-learning classification models that improved technical feasibility, generalizability, and diagnostic performance based on radiomics features include peritumoral edema characteristics. Combined with peritumoral edema region, a total of 1724 radiomics features were extracted from conventional MRI sequences for each patient. Paired 25 feature selection methods and 16 machine learning classifiers with 10-fold cross-validation were used to build classification models in training cohort (469 patients), and optimal paired methods were selected by diagnostic performance in test-Chinese Glioma Genome Atlas (CGGA) cohort (95 patients) and test-The Cancer Genome Atlas (TCGA) cohort (140 patients). The diagnostic performances were compared between paired classification model based on different feature groups. The optimal paired classification model comprised Least Absolute Shrinkage and Selection Operator and Linear Programming Boosting, which achieved area under the curve (AUC) and accuracy of 0.9981 and 0.9684, respectively, in test-CGGA cohort. In test-TCGA cohort, this classification model achieved AUC and accuracy of 0.9882 and 0.9357, respectively. Combined with the features extracted from peritumoral edema region, the paired classification model showed a significantly higher accuracy and precision than the model only based on the tumorenhanced region in both two test cohorts. The radiomics-based paired classification model provided a reliable diagnostic performance and practical generalizability in differentiating glioblastomas from PCNSLs, which also benefited from combination of peritumoral edema regions. Figure 1 . Flow chart of cohort assignment. The patients who underwent the MRI examination during the first 6th days of each month were assigned to the time-independent test cohort and the remaining were assigned to training cohort.

Neuroradiology (2020) S26 62 :S1-107 (Suppl 1) S showed the diagnostic performances of classification methods based on different ROI regions in test-CGGA cohort and test-TCGA cohort. The groups with significant differences were connected (one-way ANOVA, p < 0.05). Paired classification methods based on TER-FG and C-FG showed significant higher AUC, accuracy and specificity and F1-score than those based on PEIR-FG in both two test cohorts, while the C-FG based methods showed significant higher accuracy and precision than those based on TER-GR. PEIR-FG, peritumoral edema included region feature group; TER-FG, tumor enhanced region feature group; C-FG, combination feature group. 62 :S1-107 (Suppl 1) S Introduction Meckel's cave is a CSF-filled dural recess in the sphenoid bone and it can be involved in both intrinsic or extrinsic diseases and result in trigeminal neuralgia1-3. Classically, trigeminal neuralgia presents as recurrent episodes of brief but intense pain triggered by minimal stimulation in 1 or more branches, numbness, and weakness in the muscles of mastication3,4. Trigeminal neuralgia is typically a clinical diagnosis, however Magnetic Resonance Imaging (MRI) can be used in cases of diagnostic dilemma,5,6. This report demonstrates a rare case of unilateral trigeminal neuralgia with underlying unilateral hypoplasia of the trigeminal nerve and Meckel's cave.

Retrospective analysis. A 47-year-old female presented with a protracted history of left trigeminal nerve neuralgia, including hyperaesthesia in the mandibular (V3) distribution of the trigeminal nerve, exacerbated for the past two weeks. She had been symptomatic for many years, experiencing intermittent episodes of craniofacial pain in all three divisions of the trigeminal nerve (V1-V3), but most severe in the V2 and V3 distributions. She proceeded to have an MRI brain, which showed hypoplasia of both the left trigeminal nerve and Meckel's cave.

Result Figure 1a -c: High resolution T2-weighted sequence shows a collapsed left Meckel's cave with absent of CSF signal (*) in image A and C. The cisternal segment of the left trigeminal nerve is also asymmetrically smaller in calibre (arrows) in image A and B. Figure 2 : Axial T1-weighted MRI image shows subtle asymmetric atrophy of the left masseter and pterygoid muscles.

The significance of our case is the imaging findings without a clear pathogenesis. While the clinical presentation and imaging findings are similar to those found in the case studies describing congenital trigeminal neuropathy, our patient presented symptomatically in adulthood, suggesting an acquired aetiology rather than congenital. In light of these findings, we hypothesise that our patient's symptomatology is due to an acquired inflammatory insult, and based on the imaging, the cause of our patient's trigeminal neuralgia is the resultant hypoplasia of Meckel's cave.

Although an exceedingly rare finding, radiologists should be aware of trigeminal nerve and Meckel's cave hypoplasia when evaluating patients with trigeminal neuralgia. As imaging is required for this particular cause of trigeminal neuralgia, radiologists should be familiar with the subtle imaging findings on MRI. A timely and accurate diagnosis can aid in reducing the morbidity associated with trigeminal neuralgia. Trigeminal nerve and Meckel's cave hypoplasia are best treated through a multidisciplinary approach.

An artifact is a feature that may be present in an image which is not part of the original structure. It is a consequence either of failure of the hardware or software; physiologic motion (as cerebrospinal fluid flow) and/or ferromagnetic objects within the patient (as neurosurgical clips). Its importance relies within the possibility of misdiagnosis of subjacent pathologies. Furthermore, some specific artifacts may aid physicians when considering possibilities for differential diagnosis. To this extent, the authors aim at simplifying the approach to artifacts in brain MRI, at recognizing and working with artifacts and at how some features related to artifact may contribute to differential diagnosis.

Publications from the last 10 years were reviewed in search of artifacts that have been described in brain and spinal cord MRI scans. 20 different artifacts were found and then actively searched for in MRI scans from a university hospital in Vitória, Brazil from the last 6 years. All the findings were compiled and meticulously studied in order to provide a concise description of the main characteristics of each one. The most emblematic data was selected to compose this study.

Result 1-Artifacts contributing to diagnosis (Figs 1, 2) 2-Artifacts that must be distinguished from truly pathologic lesions (Figs 3, 4, 5)

MRI is currently the main diagnostic imaging method to evaluate central nervous system parenchyma as it generates images with high spatial resolution. With its widespread application, some obstacles arise, such as artifacts. These elements may severely disturb interpretations of findings, thus it becomes of major relevance that physicians are familiar with their existence. Correction of artifacts is feasible, though it is imperative their exact identification to allow further repairment. Besides, knowing specific features that correlate with an artifact or a specific pathology may support a diagnostic hypothesis, for instance. Attaining to do so is a relevant, though not simple, manner of improving diagnostic accuracy.

Neuroradiology S29 (2020) 62 :S1-107 (Suppl 1) S

The appropriate recognition and improvement of artifacts allows for more accurate interpretation of scan findings, thus reducing probability of misdiagnosis. As new sequences emerge constantly, it is important that physicians remain acquainted with the limitations of the diagnostic modality they are applying in their everyday medical practice. 

Drug-resistance accounts for one third of patients with focal onset epilepsy. Epilepsy surgery may lead in significant seizure reduction, remaining though underutilized. In drug-resistant focal epilepsy, localization of the epileptogenic zone is required for surgical planning prior to a focal cortical resection. Patients Neuroradiology (2020) S30 62 :S1-107 (Suppl 1) S with MRI-negative epilepsy have a lower chance of having surgery than those with MRI-demonstrated lesions and, if surgery is performed, odds of becoming seizure-free after surgery are 2.5 times higher in patients with MRI-defined lesions. Structural magnetic resonance imaging (MRI) is of fundamental significance to the identification of the epileptogenic lesion. Unfortunately, conventional imaging often fails to reveal these lesions, demanding the practice of harmonized neuroimaging of epilepsy structural sequences-(HARNESS)-MRI protocol and post-processing methods such as Volumetry.

We selected 21 patients with a medical history of drug-resistant epilepsy and a "non-lesional" conventional MRI (qualitatively evaluated). They were further examined with advanced MRI techniques by 3D-Τ1-weighted MPRAGE (voxel-size 1mm x 1mm x 1mm), before and after intravenous administration of 15ml Gadovist, 3D-SPACE-Τ2 weighted and 3D-SPACE-FLAIR-fat saturation (voxel size 0.9mm x 0.9mm x 0.9 mm) volumetric sequences. Echo-planar ΒOLD (Blood Oxygenation Level Dependent) for task-based and resting state fMRI (RS-fMRI), as well as diffusion tensor imaging (DTI-FT) sequences were acquired in order to define lateralization of language and memory, and if possible, to visualize epileptogenic zones. For postprocessing of task-based fMRI and DTI-FT, BrainMagix (Imagilys-Belgium), for RS-fMRI GraphICA (Brainet-Canada), and for Volumetric Quantitative Evaluation SurferMagix (platform which integrates Free-Surfer into BrainMagix), were used.

All 21 patients had a "non-lesional" MRI. Applying advanced and quantitative imaging techniques, MRI-negative was transformed to MRI-positive in 16 patients. After surgery, 15 out of 16 patients, were free of disabling seizures (Engel class I), with one-year follow-up. Surgery was also conducted for the patients with the non-lesional MRI. Among these 5 patients, only one was free of disabling seizures, during one-year of follow-up.

By revealing lesions unseen with conventional imaging, the HARNESS-MRI protocol combined with advanced and quantitative imaging techniques have the potential to transform MRI-negative into MRI-positive cases. Thereby, applying the technical advances and developments in neuroimaging more systematically in everyday clinical routine, we succeed in offering the life-changing benefits of epilepsy surgery to a greater number of patients. Neuroradiology S31 (2020) 62 :S1-107 (Suppl 1) S Introduction Immunotherapy has revolutionised the treatment of metastatic disease from a variety of different primaries, but is frequently associated with immune-related adverse events (irAEs). Immunotherapy-related hypophysitis (IH) is common but challenging to recognise when the presentation is non-specific or masked by glucocorticosteroid therapy, commonly required in the first 3-6 months after immunotherapy for the treatment of other irAEs. Imaging, therefore, plays an important role in the diagnosis. Incidental imaging evidence of IH may also be identified on routine follow-up imaging without relevant symptomatology, or precede the develop of symptoms.

This pictorial review illustrates the imaging features of IH and some of the important differentials in oncology patients.

The key MRI feature of IH is diffuse, modest enlargement of the pituitary gland. Indeed, the pituitary gland in IH is often only borderline enlarged, or may even measure within the normal range. Thus, comparison should be made to pre-treatment baseline imaging if the pituitary gland measures towards the upper limit of normal on neuroimaging performed within 6 months of commencing immunotherapy. The temporal evolution of this enlargement should correlate with treatment, being normal pre-treatment, enlarging after commencing immunotherapy, and subsequently returning to baseline or smaller. IH is usually associated with homogeneous enhancement of the pituitary gland, and the pituitary stalk may be thickened. Larger pituitary size, deviation of the pituitary stalk, the presence of a discrete lesion surrounded by normal pituitary tissue, sellar expansion and clival invasion are not typical of IH and suggest alternate diagnoses. Correlation with FDG-PET, if already performed as part of routine restaging, is also valuable. The normal pituitary gland does not demonstrate significant FDG activity, thus new and significant FDG uptake indicates pathology.

Immunotherapy-related hypophysitis is a common but challenging diagnosis clinically, thus radiologists should specifically assess the pituitary gland within the first months after commencing immunotherapy. As the pituitary enlargement occurring with hypophysitis may be subtle, it is important to appreciate enlargement compared to baseline, with subsequent regression. Recognition of these imaging changes should prompt longitudinal biochemical evaluation for evolving pituitary dysfunction. Radiologists should also be aware of features suggestive of alternate diagnoses, including metastases and incidental lesions. 

The apolipoprotein E (APOE) ε4 allele is the main genetic risk factor for Alzheimer's disease (AD), but the underlying pathogenic mechanism for the increased risk remains elusive, as well as the association to cerebral small vessel disease (SVD). SVD is prevalent among patients with cognitive impairment and is thought to have an important role in the pathophysiology of dementia. We aimed to investigate the association between the APOE ε genotype and magnetic resonance imaging (MRI) markers of SVD in a memory clinic population.

This is a cross-sectional study with 520 patients undergoing dementia investigation, including an MRI brain scan and APOE genotyping. MR images were assessed for markers of SVD: cerebral microbleeds (CMBs), white matter hyperintensities (WMHs), brain infarcts, enlarged perivascular spaces, cortical superficial siderosis (cSS), and intracerebral haemorrhage. Cerebrospinal fluid (CSF) was analysed for routine biomarkers in 399 patients, including amyloid β (Aβ) 42, total tau and phosphorylated tau.

Neuroradiology (2020) S32 62 :S1-107 (Suppl 1) S Result APOE ε4 carriers with AD had increased number of CMBs when looking at all and lobar brain regions (P<0.001). Decreased numbers of CMBs were seen in APOE ε2 (P<0.05), ε3 and ε3/3 carriers (P<0.001) when looking at all brain regions. Increased number of CMBs in deep and infratentorial regions were seen in APOE ε2 and ε3 (P<0.05). In APOE ε4/4 carriers, CMBs, cSS, WMHs and enlarged perivascular spaces were associated with decreased CSF Aβ42 in the whole cohort, AD and MCI (P<0.05).

CMBs and Aβ42 brain deposition are associated with the APOE ε4 allele, whereas APOE ε3 and ε2 are associated with markers of hypertensive arteriopathy. The association between APOE ε4 allele and cerebrovascular disease in the shape of CMBs may partly explain the relationship between APOE ε4 and AD. Further longitudinal assessment is warranted to better assess neuroimaging markers and development of SVD and AD with the APOE alleles.

There are two major categories of reporting error in radiology, perceptual and cognitive errors. Perceptual errors make up the majority, accounting for 60-80% of these1,2. In many industries and other medical specialties, standardised practices have been shown to reduce bias and substantially increase quality3. Structured reporting has been suggested as doing the same in radiology4. We propose a systematic approach for interpreting MRI Brains in order to reduce perceptual errors and to maximise reporting speed and confidence.

Having outlined the order that standard MRI Brain sequences should be assessed, we chose examples of specific commonly missed pathology in each of these sequences. Images were taken from cases seen in our institution.

We recommend the following order/structured approach to assessing MRI 

Using this systematic approach to interpreting brain MRIs, we hope to reduce missed pathology and increase a radiologist's confidence in reading these scans.

Neuroradiology S33 (2020) 62 :S1-107 (Suppl 1) S

Meningeal spread in lymphoma is not an infrequent finding in Neuroradiologist's practice. On the other hand, primary dural lymphoma (PDL) accounts for less than 1% of all central nervous system lymphomas, being non-Hodgkin marginal zone B-cell lymphoma the most frequent histology subtype. The radiological aspects from meningeal primary lesions do not widely differ from those found in secondary involvement. Hence, morphology and systemic disease exclusion are essential for PDL diagnosis.

We report a case of a rare subtype of primary dural lymphoma in a 60 year-old woman.

A 60 year-old woman, with no previous medical conditions, suffered a head trauma without loss of consciousness, performing thereby a head computed tomography (CT). This revealed a dural mass on the anterior portion of the superior sagittal sinus with associated brain oedema. On the MRI a dural bifrontal thickening was present, with mainly hypersignal on long TR sequences, isosignal on T1, with no restricted diffusion. After gadolinium injection, an almost complete enhancement was seen. The adjacent frontal brain parenchyma had extensive areas of T2 hypersignal, with facilitated diffusion, in relation with vasogenic oedema. The lesion was biopsied and the histology revealed an extensive diffuse and nodular infiltration of predominantly small lymphoid cells. Immunohistochemistry was positive for CD20, BCL2, BCL6 and CD10, and negative for CD5, CD23, cyclin D1 and EBER. The diagnosed was made for non-Hodgkin follicular lymphoma. A full body CT, lumbar puncture and bone marrow biopsy were performed revealing no signs of systemic disease. The final diagnosis was a primary dural follicular lymphoma.

Primary dural lymphoma is an extremely rare entity and should not be forgotten by the Neuroradiologist when considering the differential diagnosis for a dural mass. The diagnostic work-up requires a biopsy as well as exclusion of other body sites involvement with full body CT, lumbar puncture and bone marrow biopsy. This case represents an important and uncommon cause of a frequent imaging finding, that should not be forgotten as a possible aetiology. 

This study aimed to identify the molecular biomarkers associated with the natural growth rate of low-grade gliomas (LGGs, WHO grade II).

Preoperative MR images of 5274 patients with LGGs were retrospectively reviewed. Patients with at least two preoperative MR images taken at an interval of more than 90 days were enrolled in this study. The mean tumor diameters of each scan were assessed on T2-weighted images. A multivariate linear mixed-effects regression model was applied to assess the effect of biomarker expression on tumor growth rate. The slope of the linear mixed-effects model regression line indicated the velocity of tumor diameter expansion.

In total, 45 patients were included in the study. The velocity of diameter expansion (VDE) for all patients was 3.0 mm/year. Higher expression levels of mutant p53 (+3.19 mm/year, p = 0.004) and epidermal growth factor receptor (EGFR, +7.5 mm/year, p < 0.001) were found to be significantly associated with higher tumor growth rate. Higher expression level of alpha-thalassemia/mental retardation syndrome X-linked protein (ATRX, -4.18 mm/year, p = 0.016) was found to be significantly associated with lower tumor growth rate. Biomarker status of Ki-67, isocitrate dehydrogenase 1 (IDH 1), telomerase reverse transcriptase (TERT), 1p/19q, O6-methylguanine-DNA methyltransferase (MGMT) promoter, matrix metalloprotein (MMP-9), vascular endothelial growth factor (VEGF) and phosphatase and tensin homolog (PTEN) showed no independent significant impact on the tumor growth rate.

Status of biomarkers were significantly associated with the tumor growth rate in LGGs, and we used a statistical model to quantitively identify the effect of molecular biomarkers to the tumor growth rate in LGGs.

Neuroradiology (2020) S34 62 :S1-107 (Suppl 1) S :S1-107 (Suppl 1) S 

The aim of this study was to evaluate the impact on decision making and treatment planning of brain metastases (BM) using gadopiclenol, a new high relaxivity macrocyclic gadolinium-based contrast agent (GBCA).

This is a post-hoc analysis of a phase IIb study, on data from patients who underwent two separate magnetic resonance imaging (MRI) examinations, one with gadopiclenol and one with gadobenate dimeglumine, both at 0.1 mmol/kg. Patients with ≥1 BM detected in any of both scans were subjected to a blinded reader analysis. For each patient, treatment plans (stereotactic radiosurgery [SRS] or whole-brain radiotherapy [WBRT]) were determined for both MRIs, with the gross tumor volume (GTV) indicating the contrastenhancing aspects of the tumor.

A total of 13 adult patients (31% females) presenting with at least one brain metastasis were analyzed. Among these patients, gadopiclenol depicted additional BM as compared with gadobenate dimeglumine in 7 patients (54%) and the treatment plan was modified in 2 patients (15%) ( Table 1 ). In one of those patients, a single metastasis was seen only with gadopiclenol, changing the decision from no treatment to SRS. In the second patient, gadopiclenol detected 5 additional metastases, changing the decision from SRS to WBRT. The mean GTVs were relatively comparable between gadopiclenol and gadobenate dimeglumine, except for one patient for which only one lesion was detected with both contrast agents but with a GTV larger with gadopiclenol than with gadobenate dimeglumine (24.1 vs. 22.0 cm3, respectively).

Gadopiclenol improved detection of BM, which led to a change in treatment decisions in 2 out of 13 of patients (from no therapy to SRS and from SRS to WBRT). Except for one patient, the mean GTVs were similar between gadopiclenol and gadobenate dimeglumine. Gadopiclenol at the dose of 0.1 mmol/kg could potentially improve impact of diagnostic information on patient management, both on diagnostic thinking and therapeutic decisions. the dorsal surface of the clivus. The majority of retroclival EP is incidentally found and therefore asymptomatic, and its diagnosis can be challenging. Although EP is T1-hypointense, T2-hyperintense and has no contrast enhancement, its MRI appearance can be diverse. Chihara et al. (2013) proposed a classification based on its imaging characteristics: a cyst-like component on the dorsal surface of the clivus (type A) that can have an associated T2-hyperintense lesion within the clivus (type B). Atypical imaging features (incomplete or variant) should be recognized and approached with caution. Precise knowledge of these imaging features is of paramount importance to reach an accurate diagnosis and to distinguish EP from clival malignant lesions such as chordoma. The aim of this work is to review the radiological characteristics of EP that may rule out differential diagnoses and allow prompt diagnosis.

We reviewed imaging studies (MRI and CT) of 6 cases of incidentally diagnosed retroclival EP in order to access its radiological characteristics, particularly location, size, signal intensity and contrast enhancement. We compared retroclival EP imaging features we found with reviewed literature, its proposed classification and its main differential diagnosis.

Cranial MRI studies revealed midline retroclival classical EP features in all cases: 2 were diagnosed with type A (both presenting an osseous stalk) and 4 with type B. Largest diameter of lesions varied from 1.1 to 1.7 cm and when follow-up studies were available, no significant changes were recorded. No enhancement after contrast administration nor atypical imaging characteristics were found.

Discussion EP generally appears well circumscribed, homogeneous and located in the intradural space of the prepontine cistern. It may have a retroclival attachment (osseous stalk). Both EP and chordoma share similar origins and MRI signal characteristics, however chordomas are malignant lesions with a destructive behavior, mostly extradural, usually symptomatic due to local mass effect and show characteristic contrast enhancement and bone destruction. Distinguishing these two entities is of uttermost importance for proper management. Benign notochordal cell tumor, fibrous dysplasia, arachnoid cyst, chondrosarcoma and metastasis may share radiological features with EP and should be considered differential diagnosis.

Retroclival EP and clival chordoma are notochordal remnants with completely different prognosis. It is therefore critical to be acquainted with EP imaging findings in order to achieve a correct diagnosis and proper management. 

Internal carotid artery dissection (ICAD) is a frequent cause of ischemic stroke in young adults. However, the predominant mechanism is unclear. Territorial infarcts are more likely to be consequence of thromboembolism, whereas border-zone infarcts are more likely to be of hemodynamic origin.

Retrospective review of imaging findings of patients with ICAD causing ischemic stroke admitted in our centre since January 2010 in order to identify the most common mechanism of stroke. ICADs secondary to a major trauma were excluded.

Result 25 patients (64% males; mean age of 48 ± 14 years-old); 26 ICADs causing ischemic stroke. ICADs were unilateral in 23 of 26 cases (88%) and accompanied by asymptomatic dissection of the opposite ICA in 1 (4%) and of the vertebral artery in 2 cases (8%). One patient suffered 2 consecutive strokes caused by dissection of the left followed by the right ICA. Cerebral ischemic strokes were studied by MRI and CT in 20 (77%) and by CT in remaining 6 cases. Territorial infarcts were found in 25 cases (96%), with additional border-zone infarcts in 2 (8%); only one patient had an isolated border-zone Neuroradiology S37 (2020) 62 :S1-107 (Suppl 1) S infarct. 3 patients showed signs of chronic infarct. Middle cerebral artery (MCA) territory was affected in 24 of the 25 territorial infarcts [18 (75%) cortical and subcortical, 4 (17%) cortical and 2 (8%) subcortical] with an additional anterior cerebral artery (ACA) territory infarct in 4 patients; one patient had an isolated ACA territory infarct. MCA occlusion (n=11) or stenosis (n=6) were detected in 17 cases and ACA occlusion (n=1) or stenosis (n=3) in 4. Border-zone infarcts affected areas between ACA/MCA territories (n=1), MCA/PCA territories (n=1) or both (n=1). Intravenous thrombolysis was done in 8 patients (32%) and 11 patients (44%) underwent mechanical thrombectomy. One patient suffered recurrent stroke 6 weeks after the first episode. ICA occlusion (n=16) or sub-obliterative stenosis (n=7) were detected in 23 cases; 2 patients had a stenosis > 50% and 1 < 50%. The degree of obstruction in the dissected ICA did not differ between patients with different patterns of infarct.

In our study, the vast majority of ischemic lesions in patients with ICAD are territorial infarcts, while border-zone infarcts are infrequent. These findings support previous studies suggesting that thromboembolism, rather than hemodynamic infarction, is the essential stroke mechanism in ICAD. In this way, prevention of arterial embolism can play a crucial role in the management of these patients.

Giuseppe Pontillo 1 , Sirio Cocozza 1 , Maria Petracca 2 , Chiara Paolella 1 , Teresa Perillo 1 , Elena Augusta Vola 1 , Vincenzo Brescia Morra 2 , Giuseppe Palma 3 , Arturo Brunetti 1

Modifications of magnetic susceptibility, seemingly reflecting iron accumulation/depletion, have been consistently demonstrated in the deep gray matter (DGM) of multiple sclerosis (MS) patients, but uncertainties remain concerning the underlying neurobiological processes and their clinical relevance. We used quantitative susceptibility mapping (QSM) and longitudinal relaxation rate (R1) relaxometry to disentangle the contribution of atrophy, iron and myelin changes to DGM damage in MS, simultaneously exploring their relationship with clinical disability.

In this prospective cross-sectional study, from December 2013 to April 2015, 91 patients and 55 healthy controls were imaged with 3T MRI to compute QSMs and R1 map, from which iron and myelin concentration maps were estimated by applying an external model. Modifications of DGM iron and myelin (mean concentration and total content) were investigated at both global (bulk analysis) and regional (voxel-based and atlas-based thalamic subnuclei analyses) levels. Significantly altered MRI features were tested as disability predictors in hierarchical linear regression models.

Compared to controls, MS patients showed reduced thalamic(p<0.001) and increased pallidal(p<0.001) iron concentrations. No differences emerged regarding DGM structures' myelin or iron content in the basal ganglia, while an actual iron depletion was found in the thalamus(p<0.001). At the voxel-based analysis, patients showed increased iron density in the basal ganglia(p≤0.001) and reduced iron and myelin local amount in thalamic postero-medial regions(p≤0.004) (Figures 1 and 2 ), corresponding to reduced iron and myelin content in the pulvinar(p≤0.001) at the subnuclei analysis. Thalamic volume(B=-0.341,p=0.02), iron concentration(B=-0.379,p=0.005) and content(B=-0.406,p=0.009) significantly predicted disability, as well as pulvinar iron(B=-0.415,p=0.003) and myelin(B=-0.406,p=0.02) content, independently from atrophy.

Quantitative MRI suggests an atrophy-related iron increase within the basal ganglia of MS patients, along with an absolute reduction of thalamic iron and myelin, which correlates with disability. Atrophyindependent depletion of thalamic iron and myelin may represent a sensitive marker of subclinical inflammation. FIGURE LEGENDS Figure 1 . Image shows voxelwise analysis of iron maps. Clusters of significant between-group difference regarding unmodulated (top panel) and modulated (lower panel) iron maps for both the MS>HC (redyellow) and MS<HC (blue-light blue) contrasts are presented, superimposed on the QSM template in the MNI space. Figure 2 . Image shows voxelwise analysis of myelin maps. Clusters of significant between-group difference regarding modulated myelin maps for the MS<HC (blue-light blue) contrast are presented, superimposed on the QSM template in the MNI space. No significant differences emerged for the MS>HC contrast or for unmodulated myelin maps.

Neuroradiology (2020) S38 62 :S1-107 (Suppl 1) S

Introduction

The aim of this study was to evaluate the diagnostic preference between gadopiclenol, a new macrocyclic gadolinium-based contrast agent (GBCA) with high relaxivity, and gadobenate dimeglumine for magnetic resonance imaging (MRI) of central nervous system (CNS) according to field strength and type of lesions.

This is a post-hoc analysis of a phase IIb study, on data from patients who underwent two separate MR examinations, one with gadopiclenol (0.05 or 0.1 mmol/kg) and one with gadobenate dimeglumine (0.1 mmol/kg). Diagnostic preference was assessed by 3 blinded readers in a global matched-pairs fashion and analyzed according to field strength (1. 5T [n=78] or 3T [n=33]). Separate analyses were performed in patients with intra-axial lesions (IAL) (n=46) and the subgroup of those with brain metastases (BM) (n=27).

When comparing gadopiclenol 0.1 mmol/kg to gadobenate dimeglumine, all 3 readers expressed a preference for images with gadopiclenol in most cases at 1.5T (47.2% to 88.9% vs 0% to 5.6%) and at 3T (41.2% to 82.4% vs 0% to 5.9%). When comparing gadopiclenol 0.05 mmol/kg to gadobenate dimeglumine, the 3 readers expressed no preference in most cases (50% to 76.2% at 1.5T and 37.5% to 81.3% at 3T), a preference for gadopiclenol in 7.1% to 23.4% of the cases at 1.5T and 12.5% to 56.3% at 3 T and for gadobenate in 16.7% to 26.2% at 1.5T and 6.3% at 3T. In patients with IAL, the 3 readers expressed a preference for images with gadopiclenol 0.1 mmol/kg in the majority of the cases (52.4% to 90.5% vs 0% to 14.3% of preference for gadobenate) and no preference between gadopiclenol 0.05 mmol/kg and gadobenate (36% to 68%). Similar results were observed in patients with BM.

Whatever the field strength, readers expressed in majority a preference for CNS MRI with gadopiclenol at 0.1 mmol/kg compared to gadobenate at 0.1 mmol/kg and no preference between gadopiclenol at 0.05 mmol/kg and gadobenate at 0.1 mmol/kg. Similar results were observed for the subgroups of patients with IAL and more specifically with BM. Gadopiclenol at 0.05 mmol/kg could provide similar diagnostic efficacy as currently marketed GBCAs.

The VASARI (Visually AccesSAble Rembrandt Images) lexicon was developed for The Cancer Imaging Archive (TCIA) to aid standardisation of reporting glioma MR imaging features in pragmatic imaging biomarker studies. Published work assesses expert rater agreement on TCIA data. Using a cohort of glioblastoma cases from real world clinical practice, we evaluated the agreement of VASARI features with raters of varied experience.

Proven glioblastomas presenting serially through a regional neuro-oncology center from 2012-2015 with known, but blinded, molecular diagnosis and outcomes were included. MRIs were included if FLAIR, diffusion and preand post-contrast T1W sequences (ideally volumetric) were available. A novice rater (medical student) was familiarised with the VASARI lexicon on a training set, and then assessed all presentation MRIs using the 24 VASARI imaging features and bi-dimensional lengths. These MRIs were independently rated by a neuroradiology fellow, and an expert neuro-oncology neuroradiologist (10 years' experience). Interrater agreement was assessed using intraclass correlation coefficients (ICC) and one-way ANOVA (with p<0.01 taken to be significant).

116 MRIs were fully assessed. There was excellent interrater agreement (ICC>0.8) for tumour location (feature1; ICC=0.90), lesion side (f2; ICC=0.92), enhancing tumour crossing midline (f23; ICC=0.86) (figure 1) and size measurements (f29 and f30; both ICC=0.91). Good agreement was seen for eloquent cortex involvement (f3; ICC=0.71), proportion of enhancing tumour (f5; ICC=0.66), proportion of oedema (f14; ICC=0.71), multifocality (f9; ICC=0.74), ependymal invasion (f19; ICC=0.79), and satellites (f24; ICC=0.75). Some features had higher agreement between the neuroradiologists than all raters (p=0.001) including: necrosis (f7), haemorrhage (f16), diffusion abnormality (f17) and white matter invasion (f21).

Many VASARI criteria, particularly gross structural features, can be reliably assessed using a novice reader, with limited training, on realworld data. Complex, interpretational features appear reliant on neuroimaging expertise. Analysis with matched molecular and outcomes data is planned to identify whether VASARI features in clinical cohorts provide useful diagnostic or prognostic biomarkers, as reported for TCIA data. Non-expert raters can reproduce standardised visual features, which could facilitate imaging biomarker research and pragmatic clinical applications, including large scale and efficient comparisons for machine learning and radiomic glioma studies.

Neuroradiology S39 (2020) 62 :S1-107 (Suppl 1) S 

Changes of brain structure and function have been described in peripheral neuropathies. Aim of our study was to systematically investigate possible modifications of major large-scale brain networks using resting-state (RS)-fMRI in CMT1A patients.

In this prospective cross-sectional study, from May-2017 to May-2019, we acquired 3T MRI brain scans of right-handed genetically confirmed CMT1A patients and age-and sex-comparable healthy controls. Patients also underwent clinical and electrophysiological examinations assessing neurological impairment. RS-fMRI data were analyzed using a seed-based approach, with thirty-two different seeds sampling the main hubs of default mode (DMN), sensorimotor (SMN), visual (VN), salience (SN), dorsal attention (DAN), frontoparietal (FPN), language (LN) and cerebellar (CN) networks.

Between-group differences in terms of functional connectivity (FC) with the explored seeds were tested voxel-wise, correcting for local gray matter density to account for possible structural abnormalities, while the relationship between FC modifications and neurological impairment was investigated using robust correlation analyses.

Result 18 CMT1A patients (34.0 ± 11.4 years; M/F: 11/7) were enrolled, along with 20 healthy controls (30.1±10.2years; M/F:11/9). In the CMT group compared to controls, we found clusters of increased FC with the visual cortex (p=0.001), SN (p<6·10-4), DAN (p<8·10-5), LN (p<7·10-4) (Figure 1 ), along with a single cluster of reduced FC with the visual cortex in the left lentiform nucleus (p=10-6). A significant positive correlation (r=0.655, p=0.006) emerged between neurophysiological impairment and the cluster of increased FC with right temporal language areas in the left inferior parietal lobule, along with an association between walking ability and the cluster of increased FC with the left supramarginal gyrus (SN) in the homolateral middle temporal gyrus (r=0.620, p=0.006).

In conclusion, our data show evidence of diffuse functional reorganization involving multiple large-scale networks in the CMT1A brain, independent of structural modifications and partially correlating with peripheral nerve damage and functional impairment. 

Varicella Zoster encephalitis is a relatively rare form of the viral infection, predominantly occurring in immunocompromised patients. This disease course can develop both during primary infection as well as reactivation. When spreading to the brain, VZV usually causes a cerebral vasculopathy with ensuing ischemic and/or haemorrhagic lesions. Brain MRI shows multiple T2-FLAIR hyperintense lesions on the grey-white matter junction, some enhancing with contrast suggesting disruption of the blood-brain barrier. This is often paired with vascular malformations involving both small and large arteries, commonly consisting of segmental stenosis with post-stenotic dilation. Case Neuroradiology (2020) S40 62 :S1-107 (Suppl 1) S presentation Our patient was an immunocompetent 70 year old woman who presented with a headache and subsequent progressive confusion and somnolence, culminating in loss of consciousness. No skin rash was found on her body. PCR testing on cerebrospinal fluid confirmed the diagnosis of VZV infection. Brain CT showed bilateral hypodense areas in the medial temporal lobes extending subcortically to the occipital lobes. Further investigations with brain MRI showed radiological signs of a limbic encephalitis ( fig.1 ,left) corresponding to the CT images. Noticeably, the MRI also showed a number of nodular lesions with peripheral enhancement, located on the grey-white matter junction ( fig.1,right) . Follow-up brain MRI showed decreased signs of active encephalitis, yet haemorrhagic infiltration of all the lesions ( fig.2 ). MR angiography showed no vascular malformations nor obvious signs of vasculitis of the major intracranial arteries.

The presence of an initial radiological limbic encephalitis, combined with the classic lesions at the grey-white matter junction, is what makes the radiological profile of this VZV encephalitis case so peculiar. To our knowledge, no other case has been reported involving these two lesion types concurrently. We compared ours with the two cases we could find in the literature involving bilateral medial temporal lobe lesions (table) . It is noteworthy to point out that two out of the three patients were immunocompetent and none presented with a skin rash.

Agapi-Alexandra Katsarou 1 , Sotirios Bisdas 2 , Eva Liouta 1 , George Stranjalis 1 , Laurent Hermoye 3 , Vasileios K. Katsaros 1

Maximal resection and preservation of neurological function are key principles in surgery for brain space-occupying lesions. Task-based, Resting-State functional magnetic resonance imaging (fMRI) and Diffusion Tensor Imaging -Fiber Tractography (DTI-FT) provide information for description and investigation of the brain tumor relations with adjacent brain activations, networks and tracts. The aim of this study is to correlate language functions, networks and connections, in order to define the lateralization of language as a pre-surgical decision and planning tool.

In 75 patients with a known brain space-occupying lesion in the frontal and/or temporal lobe(s) by conventional MRI, we obtained 3D-Τ1-weighted MPRAGE (voxel-size 1mm x 1mm x 1mm), before and after intravenous administration of 15ml Gadovist, 3D-SPACE-Τ2 weighted and 3D-SPACE-FLAIR-fat saturation (voxel size 0.9mm x 0.9mm x 0.9 mm) sequences as routine protocol for guiding neuronavigation helped surgical procedures. Additionally echo-planar ΒOLD (Blood Oxygenation Level Dependent) for task-based and resting state fMRI (RS-fMRI), as well as diffusion tensor imaging (DTI-FT) sequences, in order to define lateralization of language, were performed. For post-processing of task-based fMRI and DTI-FT, BrainMagix (Imagilys-Belgium), for RS-fMRI GraphICA (Brainet-Canada) were used. Subsequently the Laterality Index was calculated for all methods and correlation with handedness and neuropsychological tests was performed.

Result 67 out of 75 patients (89.3%) showed left language lateralization. The combination of Task-based fMRI language activations, Resting-State networks and DTI-FT tracts demonstrated a strong positive correlation (r=0.75, P<.01) between inferior frontal gyrus activation/ superior longitudinal fasciculus visualization and verbal fluency performance (COWAT-phonemic subtest) in non-aphasic patients. Statistically significant differences in inferior frontal gyrus activation/superior longitudinal fasciculus visualization was also found when aphasic patients (m=2.4±1.2) were compared to nonaphasic ones (m=4.3±1.8) (P<.01).

The combination of fMRI and DTI-FT determines fair exactly the language lateralization in patients with brain space-occupying lesion in the frontal and/or temporal lobe(s), allowing thus a more precise pre-surgical decision and planning and preservation of language function post-surgically.

Neuroradiology S41 (2020) 62 :S1-107 (Suppl 1) S

In recent years, cerebellar abnormalities have gained increasing attention as possible physiopathological substratum of idiopathic cervical dystonia (ICD), but a consistent pattern of cerebellar structural modifications has not yet been established. We systematically investigated the presence of volumetric alterations of cerebellar gray (GM) and white matter (WM) in ICD patients, as well as their clinical relevance.

In this two-centers prospective cross-sectional study, from May 2013 to December 2017, 27 patients with ICD and 27 age-and sexcomparable healthy controls underwent brain MRI including 3D T1weighted sequences for volumetric analyses. Between-group differences in terms of gray matter and cerebellar peduncles volumes were investigated using both region of interest (ROI)-based ( Figure 1 ) and voxel-based approaches, and significant volumetric changes were correlated with clinical impairment (as measured with the Tsui score) and presence of tremor.

Result ICD patients showed significant volumetric reduction of cerebellar GM in the anterior lobe and lobule VI, resulting from both ROIbased (p≤0.009) and voxel-based (p≤0.04) analyses, while small clusters of reduced WM volume were found in the right cerebellum and left midbrain (p=0.04), along with reduced volume of the bilateral superior (p=0.04) and middle (p=0.03) cerebellar peduncles ( Figure 2 ). Furthermore, higher middle cerebellar peduncles volume was associated with the presence of tremor (p=0.04).

Our data show evidence of a specific pattern of cerebellar structural abnormalities in ICD patients, with volume loss mainly involving cortical GM regions related to the somatotopic representation of the affected body parts and, to a lesser extent, cerebellar peduncles. FIGURE LEGENDS Figure 1 . Results of the segmentation of cerebellar lobules and peduncles. (A) In a 53-year-old female patient, the SUIT cerebellar atlas is aligned in the native subject space and superimposed on (from left to right) axial, coronal and sagittal reconstructions obtained from the 3D T1-weighted sequence. (B) In a 54-year-old male patient, atlas-derived cerebellar peduncles ROIs are aligned in the native subject space and superimposed on coronal (left column) and axial (right column) reconstructions obtained from the 3D T1-weighted sequence. Figure 2 . Results of the voxel-based analyses. Thresholded statistical maps (in red-yellow) for the ICD < HC contrast regarding GM (A) and WM (B) volumes are superimposed on the SUIT T1-weighted template in axial planes. Introduction Implantable devices are encountered with increasing frequency in neuroradiology, and knowledge of the normal appearances and correct neuroanatomical placement of devices is vital when interpreting radiographic, CT, or MRI studies. We present an educational review of the indications for and imaging appearances of diagnostic and therapeutic devices to enable the neuroradiologist in training to successfully interpret these studies.

At three tertiary neurosciences referral centres, a RIS/PACS search covering all imaging modalities from 2010 to 2020 was undertaken to identify relevant studies demonstrating an implantable device in the central or peripheral nervous system. Keywords included stimulator and modulator. Cases: Postoperative radiographs, CT, and MRI images are presented, demonstrating the positioning and normal imaging appearances of implanted CNS and PNS devices. Intracranial therapeutic implants such as deep brain stimulators typically target deep grey nuclei to treat movement disorders ( Figures 1& 2) . Spinal stimulators typically target the dorsal columns to achieve pain relief in patients with refractory pain syndromes ( Figure 3 ). Peripheral nerve stimulators can be part of the management of refractory pain syndromes, including occipital nerve stimulation for migraine ( Figure 4 ) and

Neuroradiology (2020) S42 62 :S1-107 (Suppl 1) S sphenopalatine ganglion stimulation for cluster headache. Peripheral nerve stimulators can also be used in the treatment of incontinence through targeting of the sacral ( Figure 5 ) or pudendal nerves, or in the treatment of epilepsy through left vagal nerve stimulation in the neck.

New implantable devices are trialled and incorporated into clinical practice at an ever-increasing pace. Some devices, such as deep brain stimulators, have variable targets, and so close liaison with neurology and neurosurgery teams is required to determine whether surgery has provided a satisfactory anatomical and clinical result. Other devices such as spinal stimulators have a high rate of device failure, and will require a detailed imaging assessment of placement before revision is considered.

Correct implantable device placement is necessary to reach a diagnosis or achieve a therapeutic effect, and the cases presented will help the reader in interpreting postoperative radiographic, CT, and MRI studies performed to assess these devices. :S1-107 (Suppl 1) S

In this work we evaluate the clinical relevance of robust fMRI approaches to assess language lateralization [1] and of advanced tractographic reconstructions [2] for pre-surgical management of patients undergoing brain lesion resections. Here we compare preoperative imaging with intra-and postoperative findings in a pilot cohort of patients.

Four patients ( Fig.1 ) underwent brain tumour surgery and were eligible for preoperative 1.5T MRI as per flow chart in Fig.2 . Motor and language fMRI employed a BOLD GRE-EPI sequence (TE/TR=40/ 3000ms, voxel=2.5x2.5x3mm^3); paradigms consisted of 6 cycles of alternating rest and task periods (30 sec); language lateralization was calculated using a threshold-independent method validated with a reference cohort of volunteers [1] . Diffusion data were acquired with a DTI SE-EPI sequence (TE/TR=86/9500ms, voxel=(2.5mm)^3, 64 directions at b=1500s/mm^2); the corticospinal tract (CST) was reconstructed using multi-fiber probabilistic tractography (MRtrix3) [3] . Presurgical brain mapping was performed using Cranial and StealthViz MEDTRONIC Software. Intraoperative motor and language mapping were performed with train-of-5 and Penfield techniques, respectively [4] . Post-resection stimulation distances [5] were compared with tractography-based CST-cavity distance measurements, following nonlinear registration of pre-and postoperative MRI (Fig.4d ).

Major preoperative and intraoperative findings are summarized in Fig.3 . Language mapping confirmed the hemispherical dominance predicted by fMRI in all patients. Patient 2 developed SMA syndrome with speech initiation deficit (expected in the dominant hemisphere, Fig.4a ), while Patients 3 and 4 had no postoperative language deficit, as expected. In all patients, CST was identified by intraoperative stimulation in locations predicted by tractography. Multi-fiber probabilistic tractography also predicted infiltration of the CST in Patients 2 ( Fig.4 ) and 4, confirmed by intra-operative ultrasound and GLIOLAN imaging. Tractography CST-cavity distances were similar to intraoperative stimulation measurements (Fig.4c ), but tended to overestimate the value (Fig.3 ).

All major preoperative imaging predictions were confirmed intraoperativelyfuture work will expand this systematic assessment to a larger cohort. Advanced pre-surgical MRI evaluations on vetted cases were feasible, and a constant dialogue between neuroradiologists, neurosurgeons, and medical physicists allowed optimization of clinical resources and cross-validation. 

Central nervous system involvement has been described in peripheral neuropathies, including different forms of Charcot-Marie-Tooth (CMT) disease. Aim of our study was to systematically investigate possible brain structural modifications in CMT1A patients, using volumetric MRI and diffusion tensor imaging (DTI).

In this prospective cross-sectional study, from May-2017 to May-2019, we acquired 3T MRI brain scans of genetically confirmed CMT1A patients and age-and sex-comparable healthy controls. Patients also underwent clinical and electrophysiological examinations assessing motor and sensory domains. Voxel-based morphometry (VBM) and tract-based spatial statistics (TBSS) analyses were performed using a nonparametric approach based on permutations, including age and sex (and total intracranial volume for VBM) as nuisance covariates. When between-group differences emerged at VBM or TBSS analyses, the first eigenvariate was extracted from the cluster and its age-and sex-adjusted standardized residuals tested for correlation with clinical and electrophysiological variables.

Result 20 CMT1A patients (34.5±11.1years; M/F:11/9) were enrolled, along with 20 healthy controls (30.1±10.2years; M/F:11/9). The VBM analysis revealed clusters of significantly increased GM volume in CMT1A patients compared to healthy controls, encompassing the bilateral cerebellar lobules III-VI and the left hippocampus (all ps=0.04) (Figure 1 ), with no differences in terms of DTI metrics at the TBSS analysis. A significant negative correlation (r=-0.502, p=0.03) emerged between ulnar compound motor action potential and the z-scores corresponding to the right cerebellar cluster of augmented GM volume ( Figure 2 ).

Our data show evidence of structural reorganization in the brain of CMT1A patients, possibly reflecting neural plasticity mechanisms in response to peripheral nerve pathology and modulating the effect of axonal degeneration on functional impairment. Discussion & Conclusion Our data show evidence of structural reorganization in the brain of CMT1A patients, possibly reflecting neural plasticity mechanisms in response to peripheral nerve pathology and modulating the effect of axonal degeneration on functional impairment. Figure Introduction CT head/Neuroimaging formulates one of the key investigations to guide the diagnosis, management and prognosis of head injuries. Head injury is the commonest cause of death and disability in people aged 1-40 years in the United Kingdom. This study evaluated whether we were appropriately using NICE Guidelines to assess head injury patients by performing CT head.

A single centre retrospective study was carried out looking at patients attending the Emergency Department in August 2019 with suspected head injury. 102 patients were identified using iCare and data collection was completed using FirstNet and PACs.

Total of 102 patients fit the criteria out of which 40 patients had CT head 20 out of 40 patients fit the criteria for CT head as per NICE guidelines Mean time to be triaged in ED was approximately 25 minutes Mean time to be seen after triage was approximately 2 hours CT head performed within 1 hour of request was 36%. CT head performed within 1 hour of triage was 7%. CT head reported within 1 hour was 75%.

This retrospective study showed that doctors and nurses are not compliant with the NICE guidelines as 50% of CT head were done when it was not warranted. CT head was not being performed within the 1st hour of head injury presentation: delay of approximately 2-hour 20minutes between triage and assessment by a medical practitioner. 75% of data suggest CT head was being reported within 1 hour, which is compliant with NICE guidelines. Moreover, it was impossible to say whether some of the risk factors such as vomiting, seizures, loss of consciousness or amnesia was not assessed or merely not documented. This data suggested that an intervention is required to improve the documentation process and compliance of doctors and nurses in order to improve the delays between triage, assessment and performing CT head. Furthermore, the cost of unwarranted CT scans and patients being exposed to unnecessary radiation could be reduced. Hence, we are implementing a quality improvement project in order to address the above issues. 

In January 2019 we treated one male patient with medically refractory ET for a second unilateral MRgFUS procedure targeting the VIM and ZI on the right. He had previously undergone a unilateral MRgFUS procedure 14 months before, targeting the VIM and ZI on the left. Tremor severity and functional impairment are assessed at baseline and at regular intervals post-treatment for 12 months using the Bain-Findley Spirals (BFS) and Clinical Rating Scale for Tremor (CRST) which includes an arm tremor score.

There was successful thermal ablation of the target tissues at the Vim and ZI at each procedure with immediate improvement of tremor suppression which was sustained over the follow up period. From baseline to 12 months post-treatment BFS spirals improved 60.0% right arm and 64.3% left arm (Figure 1 ), CRST improved 75.4% ( Figure 2 ) and within the CRST the arm tremor score improved by 66.6% in both arms ( Figure 3 ). Adverse events following bilateral treatment were mild and transient, including unsteadiness lasting 2 weeks and dysarthria lasting 3 weeks. However subtle fatiguability of speech was noted at 3 month follow up on clinical assessment by consultant neurologist but not by the patient or his wife. At 6 months follow up, speech had returned to pre-procedure status both subjectively and objectively.

This pilot study has demonstrated both the safety and efficacy of the world's first staged bilateral MRgFUS procedure for the treatment of ET with stable results at 12 month follow up and no long term adverse effects. This provides further evidence that MRgFUS is an effective and curative treatment for tremor. 62 :S1-107 (Suppl 1) S

The basal ganglia are paired deep gray matter structures that can be involved in a wide variety of diseases. However, only a few substances reduce T1 relaxation time, and they appear bright on T1-weighted images. These include methemoglobin, protein, calcium, different minerals, lipids, and melanin. Radiologists can detect Uni or bilateral abnormalities of the basal ganglia in acute and chronic pathologies, and although magnetic resonance imaging (MRI) is the imaging modality of choice, the correct diagnosis is comprehensive; clinical-laboratory.

Method 10 patients were collected, who were studied in an MRI of 1.5 teslas.

Different entities can be associated with increased signal intensity in T1 in the basal ganglia; Basal Ganglia Calcification Basal ganglia calcifications are often seen in daily practice, usually idiopathic and physiological. Pathological processes may also be the cause of dystrophic calcifications, especially when encountered in younger patients. such com; Fahr disease, sequelae of ischemia, toxin exposure, infection, metabolic causes. Intracranial Haemorrhage T1 shortening is seen in early and late subacute stages of haematoma formation. This is attributed to the presence of paramagnetic methaemoglobin. Basal ganglia haemorrhage is usually hypertensive in aetiology. Haemorrhagic Infarction Infarcts involving the basal ganglia can be complicated by haemorrhagic conversion. As above, subacute blood can be hyperintense on T1W sequences. Hypoxic Ischemic Encephalopathy There is predominant injury to gray matter where the metabolic requirements are higher. T1W hyperintensity in the basal ganglia has been described to be due to microhaemorrhages. Hemichorea-Hemiballismus Syndrome The syndrome is most commonly seen in patients with non-ketotic hyperglycaemia on a background of poorly controlled diabetes. Patients are found to have contralateral signal abnormalities in the basal ganglia, most commonly T1W hyperintensity. Wilson's Disease An autosomal recessive disorder of copper metabolism. Abnormal copper accumulation leads to neurodegeneration. Typically, T2W hyperintensity may be seen in the bilateral basal ganglia, thalami, midbrain and pons. Deposition of paramagnetic copper in the basal ganglia can also result in T1 shortening. Neurofibromatosis Type Focal areas of signal intensity are often described in patients with NF1. They can occur in the basal ganglia, brainstem, thalami, optic tracts and cerebellum. These are usually bright on T2W images and occasionally bright on T1W images. Gadolinium Deposition With the use of MRI contrast agents, there is increasing incidence of Gadolinium deposition in the brain, commonly in the bilateral basal ganglia. Gadoline causes relaxation in T1-and T2-enhanced sequences. 

Currently established prognostic models in traumatic brain injury (TBI) include non-contrast CT which is insensitive to early perfusion alterations associated with secondary brain injury. Perfusion CT (PCT) on the other hand offers insight into early perfusion abnormalities. We hypothesised that adding CT perfusion and permeability data to the established outcome predictors improves the performance of the prognostic model.

A prospective cohort study of consecutive 50 adult patients with head injury and GCS score ≤ 12 was performed at a single Level 1 Trauma Centre. PCT was added to routine control CT 12-24h after admission. Perfusion and permeability parametric maps were derived using deconvolution technique and Extended Tofts Model, respectively. ROI analysis was performed in 6 major vascular teritorries. Glasgow Outcome Scale (GOS) was used 6 months later to categorize patients' functional outcome to favorable (GOS >3) or unfavorable (GOS ≤ 3). We defined Core prognostic model, consisting of age, motor GCS score, pupillary reactivity and CT Rotterdam Score. Next, we added perfusion and permability data as predictors and compared updated models to the Core model using cross-validated areas under the receiver operator curves (cv-AUC).

Neuroradiology S47 (2020) 62 :S1-107 (Suppl 1) S

Significant advantage over Core model was shown by the model, containing both mean cerebral extravascular-extracellular volume per unit of tissue volume (VEmean) and cerebral blood volume of the least perfused arterial territory (CBVmin) in addition to core predictors (cv-AUC (95% CI) 0.75 (0.51, 0.84) vs. 0.6 (0.37, 0.74)).

The development of cerebral ischemia and traumatic cerebral edema constitute the secondary brain injury and represent the target for therapeutic interventions. Therefore, to improve the performance of the Core prognostic model we included the CBVmin as a predictor to reflect the degree of regional perfusion disturbance. In addition we included VEmean to quantify the amount of extravascular extracellular fluid that reflects blood-brain barrier permeability alteration and contributes to traumatic cerebral edema. Our results suggest that adding CT perfusion and permeability data to the established outcome predictors improves the performance of the prognostic model in the setting of moderate and severe TBI. Neuroradiology (2020) S48 62 :S1-107 (Suppl 1) S

Meningiomas and schwannomas are common extra-axial brain tumors. Discrimination is challenging in some locations when characteristic imaging features are absent. This study investigated the accuracy of percentage signal recoveries obtained from dynamic susceptibility contrast perfusion imaging (DSC-PI) in discriminating meningiomas and schwannomas.

Retrospective database research was conducted. Sixty nine meningioma and 15 schwannoma having DSC-PI between January 2016 and February 2020 were included. Time to signal intensity curves (TSIC) were analyzed and grouped as T1-dominant leakage, T2*-dominant leakage and return to baseline. Relative cerebral blood volume (rCBV), relative mean transit time (rMTT), percentage signal recovery1 (PSR1) and PSR2 values were calculated. The differences between the groups were investigated. Receiver operating characteristic curves were operated.

Result rCBV, rMTT, PSR1 and PSR2 values were statistically different between meningiomas and schwannomas. PSR2 provided the best discrimination. With the cut off value of 1.08 for PSR2, meningiomas and schwannomas were differentiated with 95.7% sensitivity and 93.3% specificity. TSICs were also different between two groups. Most of meningiomas showed T2*-dominant leakage (78.2%), whereas most of shwannomas showed T1-dominant leakage (93.3%).

DSC-PI is a useful imaging tool for non-invasive discrimination of meningiomas and schwannomas. Particularly, percentage signal recoveries discriminates meningiomas and schwannomas with high sensitivity and specificity. Cortical laminar necrosis with putaminal necrosis is rare findings in osmotic demyelination syndrome that associated with unfavourable prognosis. We report a case of a 41 years old lady with background of hypertension on Hydrochlorothiazide presented with altered sensorium and status epilepticus.

No history of chronic alcoholism or malnutrition. Laboratory findings showed significant hyponatraemia and hypokalaemia upon admission. Following rapid correction of electrolytes, patient developed multiple episodes of fits in the ward with flaccid quadruparesis. MRI of the brain showed findings of central pontine myelinolysis and extra-pontine myelinolysis. In addition, there are also cortical laminar necrosis and putaminal necrosis. As the patient was on diuretic antihypertensive, this is likely the main cause that promotes the development of hyponatraemia. The appearance of cortical laminar necrosis and putaminal necrosis in the MRI are more likely to be related to hypoxic ischemic events in this case as the brain is known to be more vulnerable due to underlying metabolic encephalopathy. The patient showed poor recovery despite treatment. Neuroradiology S49 (2020) 62 :S1-107 (Suppl 1) S Introduction Dopaminergic therapy often does not improve the 'executive' cognitive deficits in Parkinson's disease. However, executive dysfunction in these patients may be improved by targeting other neurotransmitters including serotonin and noradrenaline. One approach to investigating this is task-free (resting-state) magnetic resonance imaging (fMRI) which minimises task-related artefacts and elucidates the impact of drug treatment on brain network connectivity.

Task-free fMRI was used to investigate the effects of atomoxetine and citalopram, targeting noradrenergic and serotonergic transmission, on functional brain networks. Thirty-three patients were scanned in a double-blind, placebo-controlled crossover design, following administration of 30mg-oral citalopram, 40mg-oral atomoxetine or placebo. Seventy-six health controls were scanned once without medication to provide normative data. Executive function tests included response inhibition and verbal fluency. Network connectivity was quantified using graph analysis.

Relative to controls, Parkinson's disease patients on placebo demonstrated reduced hub degree, hub centrality and clustering coefficient. For atomoxetine, higher plasma concentrations correlated with improved verbal fluency. Atomoxetine also improved response inhibition relative to the normalisation of hub eigen centrality towards controls. In patients with more severe disease, citalopram reduced whole-brain clustering coefficient, modularity and path length.

In Parkinson's disease, atomoxetine and citalopram alter functional network connectivity in different ways. Citalopram alters connectivity relative to disease severity while atomoxetine improves response inhibition in proportion to changes in hub connectivity. These results support the use of task-free imaging to understand the effect of pharmacological therapies on network connectivity.

Gliomatosis cerebri describes a rare growth pattern of diffusely infiltrating glioma. Clinical outcomes are variable, and treatment options are limited. To better define the population of patients presenting with gliomatosis cerebri tumour growth pattern, we examined referrals over 10 years to a tertiary referral centre in the United Kingdom.

We retrospectively reviewed patients referred to a multidisciplinary team and collected baseline clinical data on presenting symptoms, imaging appearances, pathology results, tumour genetics, and survival.

Thirty-two patients fulfilled the inclusion criteria, with a median age of 60 years (Interquartile Range (IQR) 48 to 69 years). The overall median length of survival from MDT referral to death was 46 weeks (IQR 24 to 68 weeks). The most common presenting symptoms were neuropsychiatric (31%), seizure (28%), or headache (22%). 19 patients had molecular data on tumour IDH Mutation status. 13 tumours were IDH wild-type, and 6 were IDH mutant. Where definitive histological analysis was available, the commonest tumour types were WHO grade IV glioblastoma (n=9/27, 33%), WHO grade II diffuse astrocytoma (n=7/27, 26%), and WHO grade III anaplastic astrocytoma (n=4/27, 15%). 2 patients had imaging consistent with gliomatosis, but biopsy was non-diagnostic. On imaging, the median number of supratentorial lobes infiltrated by tumour was 4. Contrast enhancement patterns varied between and within tumours. In 6/8 (75%) patients who underwent MR perfusion, one or more tumour regions with moderately increased relative cerebral blood volume were identified. MR Spectroscopy was obtained in 3 patients, of whom only 1 had a result consistent with tumour, with reversal of the normal choline:creatinine ratio.

The imaging appearances, histological, and genetic findings of gliomatosis cerebri are heterogeneous and can pose challenges to securing the diagnosis. Advanced imaging, including MR perfusion, may help to identify a biopsy target.

Anatomical segmentation of magnetic resonance (MR) or computed tomography (CT) images is an important method for diagnostics and scientific research. For example, segmentations provide the basis for deriving quantitative measures such as anatomical volumes and help to identify or monitor the progression of diseases. Manual segmentation, however, requires expert knowledge of radiologists and is Neuroradiology (2020) S50 62 :S1-107 (Suppl 1) S therefore often not available or limited to small datasets. An alternative approach is to automatize segmentation using sophisticated software tools. This led to the development of several high-performance segmentation tools, which achieve good segmentation accuracy, but are usually limited to T1-weighted MRI scans and require long processing times. In our work, we explore whether these weak points of classical segmentation tools can be alleviated using fully convolutional neural networks (F-CNN): Due to the generality of neural networks they can be directly adapted to different imaging modalities, merely by adjusting the weights of the network during training. In addition, neural networks can be implemented on graphical processing units (GPUs), which reduces processing times.

We investigate the segmentation performance of F-CNNs in a combined study on four different imaging modalities, which include both MRI (T1-MPRAGE, FLAIR, DWI) and CT scans. A large database of multi-modal scans, with up to 500 healthy cases per modality, allows us to train the networks and benchmark the performance. We segment 3D volumes into 28 anatomical structures. The segmentation maps for training were generated using the open-source software FreeSurfer on the T1-MPRAGE dataset and subsequently registered to the different imaging contrasts.

We find that the best segmentation performance is obtained for T1weighted MRI scans (see Figure 1 ), but that accurate segmentations can also be obtained solely from FLAIR and DWI scans. Even for CT scans we find reasonable segmentation performance for most anatomical structures. Example slices of 3D segmentations are displayed in Figure 2 . Full segmentations of 3D volumes can be obtained in a few seconds, drastically reducing the processing time required by software packages like FreeSurfer. Therefore, our F-CNN based segmentation tool represents a useful resource in applications where only a specific MR or CT scan is available for segmentation. In addition, it also enables real-time processing of the MR/CT volume directly after or between signal acquisitions. 

The survival rate of pediatric patients with congenital and acquired disorders has improved in the last decades. Sick children have an increased probability of reaching an adult or even elder age. The purpose of this presentation is to raise awareness of the frequent presence of typically pediatric CNS pathology in adults.

The cases were obtained from our hospital center PACS. MRI and CT are used to review the imaging findings of the selected patients. The images were collected by a neuroradiologist and a neuroradiology resident between 2017 and 2019. The selected patients had imaging studies for diverse clinical reasons, and pediatric pathology were incidental findings.

A total of 17 patients were collected, with ages ranging from 18 to 90 years old (mean= 52 years old). The most common reasons for imaging were trauma and epilepsy. All the selected patients presented pediatric CNS pathology unrelated and unstated in the clinical information. We present cases of premature and term neonatal hypoxicischemic injury, germinal matrix hemorrhage sequel, congenital cerebral malformations, and spine malformations.

There is a wide range of pediatric lesions or malformations that are incidentally found in the adult population. However, with the increasing lifespan of sick children, the number of brain studies with sequels from pediatric pathology is expected to increase. Frequently, pediatric clinical history is not presented in the clinical information while requesting the examination, especially in the emergency setting.

Pediatric neurologic pathology is not an infrequent finding in adult CNS imaging studies. The neuroradiologist should be aware of this group of diseases that can be overlooked or misinterpreted.

Medical applications of deep learning (DL) suffer from the scarcity of the available training-data. Generative models (GMs) are DL tools that learn the probability distribution of a given data-set, and thus hold the potential to generate large amounts of additional training data going beyond conventional methods of data augmentation (DA). However, the scarcity of the data is itself a limitation for GMs. Thus, we propose to use image translation (IT), a sub-class of GMs where the probability distribution of a set of images is learnt conditioned on a segmentation map. This brings several advantages: First, there is a plethora of segmentation algorithms that can easily provide the necessary segmentations for MR images, thus making use of as much available information as possible. Moreover, synthesising segmentation maps is computationally cheap, e.g. by varying boundaries and sizes of anatomical structures. Furthermore, modern architectures produce state-of-the-art results with only a minimal number of parameters, allowing for efficient training with limited training data. Finally, and maybe most importantly, if used to synthesise medical images displaying pathologies, IT by construction provides perfect labels for these pathologies and thus is an ideal starting point to train segmentation networks with better performance than conventional DA.

We compare the performance of various IT networks in synthesising artificial DWI scans of brains that exhibit stroke lesions. For training, we use both healthy volumes (2771) and such with a stroke lesion (305), which have been labelled by three expert radiologists. As a second ingredient, we train a simple GM to learn the distribution of lesions and generating new 3D lesion labels. The combined pipeline ( Fig. 1) is capable of generating high quality DWI scans of stroke lesions of variable size and location, see Fig. 2 for an example. We evaluate the performance of this approach by training a U-Net to perform the segmentation of stroke lesions for various degrees of augmented training data sets. Fig. 3 shows preliminary results: Enhancing the clinical with 2771 artificial volumes generated from healthy MR scans yields a total training-set of more than 3000 volumes. The U-Net lesion segmentation is faster and achieves better results in segmenting stroke lesions compared to training on the 305 clinical cases alone.

Our results show the great potential in DA using IT and a comparative study using different IT models and various degrees of DA will be discussed and validated on an independent test-set. 

Tacrolimus in an immunosuppressive agent that belongs to the calcineurin inhibitors, on occasions causing neurological complications in the central and peripheral nervous system. Headaches and tremor represent common mild neurological manifestations reported. Encephalopathy syndromes are rare but serious complications, with white matter lesions following a PRES pattern, sometimes with a diffuse involvement resembling tumor-like or pontine myelinolysis lesions. Best efforts for this challenging diagnosis should be made as the neurotoxicity can be reversible.

Review the imaging findings of tacrolimus-induced neurotoxicity, based in two recent cases from our institution: 54-year-old women presented to the ER with a 1-month history of generalized tremor, the last 15 days with associated complains of diplopia, gait disturbance and headache. Kidney transplant in 2014, with tacrolimus and prednisolone as maintenance therapy since then. 59-year-old female admitted in the ER with sudden vertigo, gait disturbance and vomiting. Liver transplant 25 years ago and chronic tacrolimus use. Blood tests were unremarkable, and lumbar puncture showed elevated total protein level on both women. Brain MRs performed included brainstem and supratentorial non-enhancing WM abnormalities, without changes on apparent diffusion coefficient (ADC) map. MR perfusion and spectroscopy studies were normal.

Both cases showed undesirable effects linked to the use of tacrolimus, with clinical improvement after drug suspension. Imaging findings of leucoencephalopathy tend to improve at a slower rhythm. Neurotoxicity should be suspected in all patients under immunosuppressive therapies along with the most feared diagnosis, progressive multifocal encephalopathy (PML).

Joanna E Perthen 1 , Paul Dorman 2 , Danny Morland 3 , Nancy Redfern 3 , David Butteriss 1

Spontaneous intracranial hypotension (SICH) typically presents with headache and is caused by a spontaneous CSF leak within the spine. A robust treatment paradigm has yet to be agreed upon. We present retrospective data from the entire cohort of SICH patients at our regional neurosciences centre 2010 -2020. Our experience with this cohort has led to the development of a local treatment pathway. We present treatment responses for the entire cohort, to give a representative overview of this diverse patient group and enable other centres to develop robust investigation and treatment paradigms of their own.

Patients diagnosed with SICH at our neurosciences centre were retrospectively identified. Patients with a new persistent postural headache, and/or positive imaging findings were included. Our treatment pathway is presented. Responses are monitored clinically; response to a specific treatment is defined as resolution of clinical symptoms, or sufficient improvement in symptoms such no further treatment is required, for at least 6 months. A relapse is defined as recurrence of clinical symptoms after a period of at least 6 months of response. Neuroradiology S53 (2020) 62 :S1-107 (Suppl 1) S

Outcome data for our representative, diverse SICH cohort shows the success of a staged approach to treatment. The use of EBP, with myelography and targeted treatments reserved for those not responding to EBP, produces good results, with the majority of patients in our cohort attaining symptom relief. Relapse rates are low, and surgery is only rarely required. We use these data to inform our discussions with patients, enabling us to give approximate success rates for the various interventions at our institution. 

The clivus is a midline central skull base structure formed by the union of the basisphenoid and the basiocciput. Due to its location and anatomical relationships, it is of great importance and can be affected by several pathological processes. Understanding the normal development of the clivus, pneumatization of the sphenoid sinus, and bone marrow maturation is crucial to recognize potential lesions. The purpose of this work is to review the lesions involving the basisphenoid and discuss their differentiating imaging features.

We retrospectively selected cases of intrinsic basisphenoid lesions, and lesions arising above and below the adjacent skull base with basisphenoid involvement. Clinical data and CT/MRI images were collected and discussed.

The cases include metastases, a chordoma, a neurenteric cyst, an arachnoid cyst, an invasive pituitary macroadenoma, a sphenoid mucocele, and a nasopharyngeal carcinoma. One case of arrested pneumatization of the sphenoid sinus is also presented as a normal variant that may mimic disease.

Neuroimaging is a cornerstone in the diagnosis of the basisphenoid lesions, with CT and MRI complementing each other. CT has its greatest value in the appreciation of bone involvement patterns, and MRI is the best tool to evaluate the lesion structure as well as adjacent soft tissue involvement. The precise location and anatomical relationships are essential to a systematic diagnostic approach and knowledge of anatomical variants is crucial to avoid inaccurate diagnoses. However, imaging alone is usually not enough to establish a definitive diagnosis.

Developmental anomalies of the intracranial arteries are often incidental findings in digital subtraction angiographies (DSA). The posterior inferior cerebellar artery (PICA) commonly arises from a fenestrated segment of the intracranial vertebral artery (VA), and this frequent variation can be misinterpreted as or confused with a PICA of double origin. True double origin of the PICA (DOPICA) occurs when two branches of the PICA arise independently from the intracranial VA and fuse to form an arterial ring. This imaging finding is rare and has a reported incidence of 0.36-6% in various retrospective studies, totalling about 30 cases. It has important clinical implications, especially in cases of aneurysms on a DOPICA, that are even rarer. We discovered this rare variation incidentally while doing a DSA to study a spontaneous subarachnoid haemorrhage (SAH).

Description and discussion of a clinical and imagiological case.

A 54-year-old female presented with malaise, asthenia and vomiting that were initially interpreted as food poisoning. She later returned to the emergency department with the same symptoms, as well as headache and neck pain. Neurological examination revealed: GCS 15 without focal neurological deficits, but with terminal neck stiffness. Head CT scan was remarkable for intraventricular blood (H&H 1, Fisher 2, WFNS 1) and CT angiography for a partially thrombosed left PICA aneurysm, so she was hospitalized. A DSA was performed and demonstrated a double origin of the left PICA, one of them in the V3-V4 transition and the other in the proximal segment of the V4 segment, which fused distally to form a single. However, in the arterial segment corresponding to the proximal origin of the PICA, there was a large saccular aneurysm with a fusiform configuration that was oriented superiorly, had slightly lobulated contours and measured about 11mm of height, 5mm of width and 3mm of neck size (there were two other small aneurysms in both carotid arteries: intracavernous ipsilaterally and carotic-ophthalmic contralaterally). The DOPICA aneurysm was successfully treated with surgical clipping, without any complications.

This rare anatomical variation is beneficial if a complication occurs following aneurysm exclusion, that results in discontinuity of flow in the neck region of the aneurysm. Distally, the flow in PICA remains preserved by the presence of another origin, and proximally, the brainstem perforating arteries of both origins remain viable. 

Analysis of brain cortical cortical thickness [BCT] changes in mTBI revealed cortical thinning. However, less study about other brain morphometrics; cortical surface area [CSA] and subcortical brain volume [SBV] . In mTBI, apart from brain structural changes, there is functional changes as well in term of neurocognitive functions which include attention, language, memory, visual Neuroradiology (2020) S54 62 :S1-107 (Suppl 1) S spatial & executive function. The objective of this study was to evaluate the brain morphometry differences in between healthy subject and mTBI patient and its correlations with neurocognitive performances.

Method 64 mTBI patients (3 months post injury) and 20 healthy controls [HC] underwent MRI brain and Neuropsychological Assessment Battery-Screening Module (S-NAB). mTBI group was divided into recovered[R] (S-NAB >85) and non-recovered groups [NR] (S-NAB <85). Data were processed using FreeSurfer Software to obtain BCT, CSA and SBV. Relations between the significant brain morphometry and neurocognitive deficit were measured with Pearson Correlation.

There were significantly lower number of attentions, language and memory in the mTBI group compared to HC. When comparing in between R and NR groups, R group had significantly higher score of all 5 domains. mTBI group had significantly thinner cortex in the right caudal middle frontal ( There is a trend of decreasing CSA in the R vs NR in the above structures, but none shows statistically significance result. Therefore, CSA can differentiate in between the HC and mTBI groups. We found a significantly strong positive correlations (r > 0.3, p <0.005) in between CSA of left precuneus with spatial, right fusiform with memory as well as left inferior parietal with attention, memory and spatial. These CSA structures are related with its cognitive function as mentioned.

Our findings imply that at 3 months post mTBI, reduced CSA is seen in areas coinciding with neurocognitive deficit. Method 56 mTBI and 19 healthy controls (HC) were recruited. Neuropsychological Assessment Battery-Screening Module (S-NAB) performance were assessed 2 weeks post-trauma and at 3 months. The mTBI group was then divided into recovered (REC; S-NAB ≥85) and non-recovered (NREC; S-NAB <85), whereby domains affected were chiefly attention and language. DTI and NODDI were done at 3 months. Using Tract-Based Spatial Statistics (TBSS), DTI and NODDI parameters were obtained for 50 white matter tracts (WMTs). Data was analysed using SPSS. 

NODDI detected more microstructural WMT changes in mTBI 3 months postinjury than DTI, largely involving WMTs in the limbic system which has significant function in attention and language. Certain WMTs showed significant in one or more NODDI metrics, however the overall trend suggestive of reduced neurite density and alignment with increased oedema was apparent. Some findings obtained were not readily explained, suggesting neurite recovery process may be more nuanced than initially thought, which merits further research in this area.

Neuroradiology S55 (2020) 62 :S1-107 (Suppl 1) S

The changes around the sella turcica following transphenoidal surgery may pose difficulty in interpretation of post-op MRI, especially in differentiating residual tumour from post-op changes. The understanding of the normal evolution of endoscopic skull base reconstruction is essential and crucial to distinguish implanted materials from a residual neoplasm. The objective of this study is to evaluate the sphenoid/ sellar changes in MRI post transsphenoidal surgery of pituitary lesions. Our aim is to also study certain specific MRI features of different implanted materials used during TSS of over a period of 2 years and to differentiate the MRI features between post-op implant materials and recurrence of the tumour.

A total of 30 patients in our institution from January 2012 till August 2018 with pituitary macroadenoma who underwent Transsphenoidal surgery (TSS) and have serial follow up MRI post-surgery, were assessed retrospectively. Data were collected according to the morphology, signal intensity on the conventional sequence (T1WI, T2WI T1W post-contrast) of the adenoma, sellar, sphenoid sinus and the implant material (post-op MRI) over a period of 2 years. Extension of the adenoma into the sphenoid sinus was also recorded.

The MRI signal of different implanted material with time showed a variable signal and was non-specific. In term of size evolution, the combination of the nasoseptal flap (NSF) and artificial graft i.e Duragen (n=15), majority of the cases showed significant size reduction of the implant material over a period of 2 to 6 months, whereas, in the case of using NSF alone (n=10), significant size reduction of the implant material was observed over a period of 6 to 12 months. In a pre-existing complex pituitary lesion, the early post-TSS MRI, there was difficulty to differentiate between the implant material or residual macroadenoma, but with time, the residual tumour remained static in size whilst the implant material showed a reduction in size and in thickness and the floor of sellar was more well-formed.

The duration of resolution of different materials differ with time with significant size reduction of the implant material over a period of 6 to 12 months with size stability noted beyond 1 year. There is no significant difference in the MRI signal of different implant material with time.

In 2019 the UK National Institute for Health and Care Excellence (NICE) updated its guidelines on the initial management of suspected TIA. After specialist assessment, MRI should be considered the first line crosssectional imaging modality. The benefits of MRI in assessment of TIA/stroke are well documented, assessing vascular territory, haemorrhage and alternative pathologies. The principle drawback is the time taken to image patients. With a potential increase in MR brain imaging to meet these new guidelines, new ways are sought to respond to this demand. In our trust we evaluate Siemens' rapid neuroimaging protocol, GOBrain, reviewing its diagnostic accuracy in TIA patients and comparing it against our standard practice.

All scans were performed on Siemens Magnetom Skyra 3T. Our standard and the GOBrain TIA/stroke protocol consists of axial T2, sagittal T1, DWI, coronal FLAIR and SWI. Initially a sample of 10 studies were reviewed using Fujifilm's Synapse 5 PACS. A standard proforma for reporting adequacy was adapted from Nael, K et al.

Patients were identified for one month of TIA clinic attendances before and after implementation of GOBrain. The time of the study was extrapolated from image time-stamps as presented on Synapse 5. Secondary measures were assessed to review current departmental practices influencing total time from patient arrival to discharge, identifying areas for further quality improvement. This included, but not limited to, the time for report verification, final diagnosis on imaging and final diagnosis by the referring clinician.

All 10 studies were deemed adequate for diagnostic purposes. Table 1 shows the results between the two cohorts.

The introduction of a rapid sequence protocol for TIA brain imaging has shown a significant reduction in scanning time (58%). Any gain Neuroradiology (2020) S56 62 :S1-107 (Suppl 1) S in scanning time can yield enormous benefits within a radiology department, increasing overall throughput and reducing cost of imaging in select groups such as ours. We have shown that there is an increased number of patients identifying with positive ischaemic changes and final end diagnosis of TIA. This suggests the increasing importance of imaging in clinical decision making. A reduction in MR scanning time must be met with an improved patient pathway enabling the shortest time possible from patient referral to final diagnosis. This evaluation has identified areas for further improvement within our department and we propose further quality analysis in collaboration with our stroke clinician colleagues to reduce this time period.

Halimaton Khadijah Mohamad 1 , Shahizon Azura Mohamed Mukari 2 , Zhe Kang Law 3 , Ramesh Sahathevan 4

The objective of this study is to evaluate the value of CT and MRI for the diagnosis of Cerebral Amyloid Angiopathy (CAA) using Modified Boston Criteria, as the reference standard. Our aim is also to study certain features of Cerebral Amyloid Angiopathy haemorrhage (CAAH) in comparison to hypertensive haemorrhage (HH) on imaging.

From January 2012 till December 2015, 31 patients admitted for primary intracerebral haemorrhage with both CT and MR imaging done, were investigated retrospectively for the diagnosis of CAAH or HH. Interrater agreement was calculated and features of both CAAH and HH were recorded for assessment.

Using Modified Boston criteria as a reference standard, a good interrater agreement was achieved using Cohen's Kappa coefficient with κ of 0.664 in the diagnosis of in vivo CAA-ICH (n=9) and HPT-ICH (n=17). The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were 100%, 77.3%, 64.3% and 100% respectively. Patients in CAAH group were older with a range of age between 70-79, compared with the HH patients. The bleeds were all superficially located with frontal lobe (44.4%) predominant; the majority of the cases (66.7%) have a small volume of haemorrhage < 30mls, according to ICH scoring. SAH and IVH extension was seen in 11.1% and 22.2% respectively, in comparison with the HH with a percentage of 11.8% and 35.8%, respectively. 22.2% had multiple haemorrhages at different lobes and 33.3% had a recurrent haemorrhage.

A modified Boston criteria with combined CT and MR imaging is a reliable, non-invasive method for diagnosis of premortem CAAH with a good interrater agreement. There are certain CT and MRI features to some extent that may assist to distinguish CAAH from HH.

The purpose of this study is to evaluate white matter microstructural changes of the irritative zone (IZ) and non-irritative zone (N-IZ) using whole brain Tract-Based Spatial Statistics (TBSS).

This is a cross sectional study of 24 cases with frontal or temporal lobe epilepsy and 24 controls. Electroencephalogram (EEG) and MRI brain using a standard epilepsy protocol were performed in all cases. The irritative zones (IZ) were represented by fixing the DTI dataset to the right side and the non-irritative zones to the left side. A set of diffusion parameters; fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD) and radial diffusivity (RD) were acquired for the affected white matter tracks (WMT) and the WMT changes were analysed using TBSS. The distribution of affected tracts was compared between IZ and non-IZ of epilepsy patients versus controls (Figure 1 ).

There are bilateral and extensive WMT changes in the epilepsy group compared to the control group. The abnormal diffusion parameters were seen involving the limbic circuit, major association fibres (superior and inferior longitudinal fasciculi, superior and inferior frontooccipital fasciculi) as well as the corpus callosum and projection fibres. Both IZ and non-IZ showed decreased FA and AD in DTI parameters.

Diffusion tensor MRI demonstrated widespread global axonal injury in both the IZ and non-IZ hemispheres.

Neuroradiology S57 (2020) 62 :S1-107 (Suppl 1) S Introduction Parkinson's disease (PD) is a clinically heterogeneous chronic progressive neuro-degenerative disease with loss of dopaminergic neurons in the nigrosome 1 (N1) territory of the substantia nigra pars compacta (SNpc). The nigrosome 1 (N1) appears to be the largest and is located in the dorsal medial aspect of the tail (caudal most part) of the SNpc. The disappearance of N1 on MRI is known to occur over time and is associated with increased iron deposition . This is a retrospective study to investigate the feasibility of nigrosome-1 detection using 3T-susceptibility weighted MRI on 3T Siemens Magnetom Prisma MRI machine and T2* gradient recalled echo 3D sequences on 3T GE MRI machine as well as the diagnostic accuracy that can be achieved.

A group of 74 controls was used to characterize the appearance of the N1 sign and train the raters. Subsequently, 61 patients with PD and 74 controls were blindly analyzed for the presence or absence of the N1 sign on SWI and T2* GRE 3D sequences. First 5 caudal slices of the SN were reviewed.

From the group of 74 controls, N1 has variable appearance. Among the controls, 64 (86.4%) cases showed the N1 sign bilaterally, 8 cases (10.8%) showed bilateral loss of N1 sign, 2 (2.7%) showed unilateral loss of N1 sign. Of the 15 PD patients whose SWI images were reviewed, 7 (46.7%) cases showed bilateral loss of the N1 sign, 2 (13.3%) cases showed the N1 sign unilaterally and 6 (40%) case showed the N1 sign bilaterally. Of the 46 PD patients whose T2* GRE 3D images were reviewed, 24 (52.2%) cases showed a bilateral loss of the N1 sign, 5 (10.8%) cases showed the N1 sign unilaterally and 5 (10.8%) cases showed the N1 sign bilaterally.

The N1 has variable appearance. It can be visualized using SWI in the non-PD controls and is not visualized in most patients with PD.

Neuroradiology (2020) S58 62 :S1-107 (Suppl 1) S

To determine ratio of hypodense regions ('Swirl Sign') versus hyperdense regions in intracerebral haemorrhage and it's association with haematoma expansion and to explore the functional outcome of intracerebral haemorrhage patients with 'swirl sign' using modified Rankin score (mRS).

Method 34 patients who had spontaneous intracerebral haemorrhage with initial (CT1) and repeat CT (CT2) within 96 hours were included in the study. Presence of 'swirl sign', its volume and haematoma volume in initial and repeat CT were calculated using semi auto-segmentation method using 'ITK snap' application.

Result 23 patients (67%) had swirl sign present in the haematoma with the mean initial and follow up haematoma volumes of 42.5 cm3 and 54.5cm3 respectively. Mean haematoma expansion documented for this group was 11.9cm3. 11 patients (33%) had no swirl sign present with the mean initial and follow up haematoma volumes of 9.7cm3 and 4.4cm3 respectively. The mean haematoma expansion for this group was 4.4cm3. The p-value of correlation between ratio of swirl volume/haematoma with percentage of haematoma expansion [(CT2-CT1)/CT1) was 0.041 with significant correlation at 0.05 level whilst the correlation co-efficient (r value) was 0.449. Score of (0-2) was defined as good MRS score whilst the score of (3-6) was defined as poor MRS score. 3 patients (13%) with the presence of 'Swirl sign' within the haemorrhage had good mRS (0-2) and 20 patients (87%) had poor mRS (3-6). 6 patients (55%) without presence of 'Swirl sign' within the haemorrhage had good mRS (0-2) and 5 patients (45%) had poor mRS (3-6).

CT swirl sign is associated with haematoma expansion in patients with spontaneous intracerebral haemorrhage and it can be used as a predictor of functional outcome of patients in spontaneous intracerebral haemorrhage. Introduction

To explore the relationships between lipid-imaging profiles to sterol regulatory element binding proteins (SREBPs) gene expression profiling of the gliomatous tumour.

Ten histologically proven glioma patients underwent a standard preoperative MRI tumour protocol with in-and opposed-phase (IOP) MRI. The lipid fractions was derived from the segmented nonenhancing tumour regions overlaid on the IOP-constructed lipid map. The resected tissue samples were collected for correlation with SREBPs gene expression profiling. The tissue sample from 3 histology normal patients were used as control. RNA extraction was performed using RNeasy lipid tissue mini kit followed by cDNA synthesis using one-step reverse transcription protocol. A Taq-Man probe based quantitative real-time PCR was performed to evaluate the level of SREBP-1 and SREBP-2 gene expression. For this analysis, TBP gene was used as reference gene.

The lipid fraction of the solid non-enhancing region was higher for grade IV (mean SLRIV=0.073) compared to grade II and III (mean SLRII=0.053, and mean SLRIII=0.057). Gene expression analysis showed that SREBP-1 expression was increased by 7-fold change in grade II. Gene expression of SREBP-2 however was increased by 9fold change in grade II, and by 13-fold change in grade III compared with controls, suggesting SREBPs were highly expressed in both grade II and grade III. The expression of SREBP-2 was lower by 3.5-fold change in grade IV compared with normal controls.

The high expression of SREBPs was observed for grade II and III implied reprogramming of lipid metabolism. The low expression of SREBPs coincide with higher lipid fraction in Grade IV glioma, which may indicate exhaustion of lipogenesis regulation. Imaging markers may act as surrogate biomarkers to reflect changes in regulation of lipid metabolism and synthesis. Keywords Glioma, MRI, lipid, lipid fraction, SREBPs, in-opposed MRI Neuroradiology S59 (2020) 62 :S1-107 (Suppl 1) S

The hummingbird sign reflects mesencephalic atrophy and raises suspicion of progressive supranuclear palsy (PSP) in patients with neurodegeneration and motor symptoms. Reports have shown high specificity for distinguishing PSP from Parkinsons disease (1). Normal pressure hydrocephalus (iNPH) can present with gait disturbance, decreased postural control and repeated falls and is an important differential diagnosis to PSP (2). In iNPH, MRI shows several morphological signs, including a decreased size of mesencephalon (3). If iNPH is misinterpreted as PSP, important surgical treatment will be omitted or delayed.

Eight blinded neuroradiologists assessed 104 sagittal MR images depicting the mesencephalon of PSP patients (n = 26), iNPH patients (n = 40), and controls (n = 38). For each subject, they assessed if hummingbird sign was present and graded the confidence of that assessment (1-5). Assessments were combined to a score ranging from -5 to +5 (assessment score), and averaged for each subject. Separately, area measurements of each mesencephalon were conducted using a freehand tool in PACS (fig 1) . Assessment scores and area measurements were compared to clinical diagnosis. Assessment scores were plotted against mesencephalic area (fig 2) .

The hummingbird sign was present in 71% of PSP patients, 57% of iNPH patients and 7% of controls (fig 3) . Average area measurements were 80,8 mm2 for PSP (SD 16,2), 93,2 mm2 for iNPH (SD 12, 9) and 133,7 mm2 for controls (SD 16, 5) . Area measurements with a cut-off value at 90 mm2 resulted in higher sensitivity (76.9% vs. 69.2%) and higher specificity (80.8% vs. 66.7%) compared to averaged visual assessments of hummingbird sign. ROC analyses showed AUC of 0.976 (assessment score) and 0.993 (area measurements) when comparing PSP to controls only, but AUC was reduced to 0.824 (assessment score) and 0.856 (area measurements) when PSP was compared to a combined group of iNPH patients and controls (fig 4) .

The hummingbird sign can suggest either PSP or iNPH, depending on other radiological findings and clinical symptoms. Area measurements done directly in PACS using a 90 mm2 cut-off had higher sensitivity and specificity than visually assessing the presence of the hummingbird sign. Assessing the hummingbird sign is subjective and problematic. There are more than 330 different species of hummingbirds, all with distinct features (fig 5) . 

To validate in our centre the proposed new Edinburgh CT criteria for cerebral amyloid angiopathy-associated (CAA) lobar intracerebral haemorrhage (ICH), published in 2018, as a reliable interdisciplinary score that can be used on daily basis in the Emergency Department.

We retrieved the images of 400 patients, with first-ever intracerebral haemorrhage diagnosed by non-contrast CT, in our Emergency Department. Two expert neuroradiologists, two general radiologists and two neurologists independently evaluated the CT imaging appearances to assess the presence or absence of the simplified Edinburgh CT predictors. Namely, the extra-axial haemorrhage (in the subarachnoid, subdural, or intra-ventricular spaces), the finger-like projections (elongated extensions arising from the hematoma, longer than they are wide) and the localization of the intracerebral haemorrhage (whether lobar or non-lobar). Intra-and inter-observer agreement, the sensitivity and specificity were calculated.

There was a significant inter-observer agreement between Neurologists, Radiologists and Neuroradiologists (p> 0,05) in assessing the presence or absence of the extra-axial haemorrhage and the localization of the intracerebral haemorrhage. Nevertheless, we found a decent inter-observer agreement between Radiologists-Neuroradiologists and Neurologists in detecting the finger-like projections.

The new simplified Edinburgh CT criteria represent a valuable score that can be used to diagnose cerebral angiopathy-associated lobar intracerebral haemorrhage in every emergency department. Indeed, when missing the expertise of the Neuroradiologists, these criteria can be used by other figures (Radiologists and Neurologists) in Emergency, to quickly diagnose the CAA associated lobar ICH.

Introduction 18 F-FDG PET-CT has a well-established role in identifying different patterns of brain metabolism associated with different subtypes of dementia and has been shown to be able to identify focal cerebral hyoperfusion at an earlier stage then the atrophy pattern on MRI. However MR brain imaging is easily Neuroradiology S6 1 (2020) 62 :S1-107 (Suppl 1) S accessible, faster and significantly less expensive in comparison to FDG PET scan. Therefore, MRI brain is often the line first investigation of patients with cognitive impairment and is able to identify different dementia subtypes. However, MRI brain is thought to be a more subjective assessment with less sensitivity and specificity in comparison to PET-CT scan. In this study we assess the accuracy of MRI brain in identifying dementia subtype in comparison to FDG-PET scan and the interobserver variation in MRI dementia reporting Method MRI brain scans of patients who had subsequently had FDG PET for cognitive impairment in the last 3 years were identified using the radiology PACS system by a radionuclide radiologist. MRI brains were reviewed by a consultant neuroradiologist with an interest in cognitive imaging and senior neuroradiology fellow. Both were blinded to patients PET CT result. The MRI and PETs were asked to rank each scan and give one of 6 diagnoses: Normal; Equivocal; Alzheimer's disease; Fronto-temporal dementia; Vascular dementia and Other dementia. MRI scans findings were compared against each other and with the FDG PET results. A statistical analysis software is routinely used with FDG PET scans in our centre to minimise the subjective analysis.

Result 30 patients were identified, age ranged from 47 and 82 years with mean age of 64.9. 13 of the 30 patients had significant perfusion abnormality on the FGD PET scans. The interobserver agreement between the 2 MRI reporters was 67%. 17 out of the 30 patients had MRI report findings consistent with the PET-CT findings, an argument of 57%. MRI scans had shown a positive predictive value of 44.4% and a negative predictive value of 61.9% when comparted to PET CT.

Analysis of MRI scans findings confirmed a relatively low interobserver agreement between reporters with a disagreement in one third of cases. Suggesting the subjectivity of reporting MRI scans in dementia subtypes assessment. MRI scan has shown a low sensitivity of 33.3% and specificity of 72.2% when compared to FGD PET. Overall findings confirm MRI is not able to replace 18F-FDG PET CT in assessment of patient with cognitive decline. Introduction Cerebral microhaemorrhages, also referred to as cerebral microbleeds are a common finding on routine MRI brain examinations and there are a wide range of differential diagnoses associated with them. They are defined as small hypointense foci of upto 1 cm detected on gradient echo (GRE) or susceptibility weighted imaging (SWI)[1]. The two most common causes of microbleeds include systemic hypertension ( Figure 1 ) and cerebral amyloid angiopathy ( Figure 2 ) each with a distinctive distribution pattern. However, there are several uncommon causes which should be considered when cerebral microbleeds are identified on MRI imaging. These include diffuse axonal injury secondary to trauma, multiple cavernous malformations, posterior reversible encephalopathy (Figure 3 ), fat embolism, previous ECMO treatment ( Figure 4 ) and cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL). The pattern of microhaemorrhages can assist in differentiating between different pathologies. For example, the location of haemorrhage within the deep grey nuclei and infratentorial compartment are likely to reflect hypertensive pathology whereas lobar peripheral microhaemorrhages are more in keeping with cerebral amyloid angiopathy. There is also a potential association with microhaemmorhagic foci in COVID-19 patients [2].

A series of cases will be reviewed in this educational exhibit highlighting the imaging patterns of microhaemorrhages. Some of the common and uncommon causes of cerebral microhaemorrhages will be discussed with example cases. There will be particular focus on the various radiological manifestations of cerebral amyloid angiopathy including inflammatory amyloid angiopathy ( Figure 5 ) and the proposed diagnostic criteria in literature.

The aim of this pictorial review is to provide an overview of the differential diagnoses related to cerebral microhaemorrhages and their clinical significance. 

The median OS was 20.4 months. Survival was significantly lower in patients > 50 years old (P = 0.002) and without gross total resection (P = 0.03). Significant differences (P < 0.05) between long survival (LS) and short survival (SS) groups were found for both DCE and DSC perfusion parameters of CE-VOIs at T0 and T1 timeponts, consistently for rCBV 90th percentile and Vp 10th percentile.

Perfusion parameters obtained from a standardized, longitudinal protocol are able to predict prognosis before and after the beginning of standard treatments for glioblastoma.

Elmire Dervisoglu 1 , Ceylan Altintas Taslicay 1 , Cengizhan Taslicay 2 , Burcu Alparslan 1 , Umay Kiraz 3 , Gur Akansel 1

Isocitrate dehydrogenase (IDH) mutation is the most important independent molecular prognostic factor in glioblastoma (GBM). In this study, it was aimed Neuroradiology S6 3 (2020) 62 :S1-107 (Suppl 1) S to investigate the contribution of radiomics and machine learning to MR imaging in predicting IDH1 mutation in GBM.

Method 78 patients (15 IDH-mutant GBM, 63 IDH-wild type GBM) were included in this retrospective study. Preoperative MR images were evaluated using Visually AcceSAble Rembrandt Images (VASARI) scoring. A total of 107 radiomics features, including shape-based, first, second and higher-order textural features, were extracted from T2W, FLAIR, post-contrast T1W, and ADC map images. Classifications were made with support vector machine, logistic regression, k-nearest neighborhood, naive Bayes, decision tree, random forest and neural network algorithms using VASARI features and radiomics features separately. Feature reduction was done with information gain method, and stratified 10-fold cross-validation was used as an internal validation technique. The performance of machine learning algorithms was evaluated with the area under the curve (AUC), sensitivity, specificity, and accuracy rates. McNemar's test was used to establish whether the classification with VASARI scoring is significantly different from classification with radiomics features. p <0.05 was considered statistically significant.

The neural network had the highest diagnostic performance in the classification made with both VASARI (0.752 AUC, 85 % accuracy, 47 % sensitivity, 94 % specificity) and radiomics features (0.826 AUC, 87 % accuracy, 60 % sensitivity, 94 % specificity). There was no statistically significant difference between the two classifications (p = 0.803). The sensitivity of the naive Bayes algorithm using radiomics features (87 %) in detecting IDH1 mutation was significantly higher than the classification with VASARI features (60%) (p = 0.002).

Machine learning using radiomics features showed higher sensitivity compared to VASARI scoring in determining IDH1 mutation in GBM.

Ceylan Altintas Taslicay 1 , Elmire Dervisoglu 1 , Cengizhan Taslicay 2 , Ozgur Cakir 1 , Busra Bayrak Yaprak 3 , Cigdem Vural 3

In this study, it was aimed to investigate the contribution of radiomics and machine learning (ML) to MR imaging in differentiating high grade meningioma (HGM)-low grade meningioma (LGM) and classifying LGM as fibrousnonfibrous.

The study protocol was approved by the institutional ethics committee.163 patients (32 HGM, 131 LGM; 15 fibrous, 116 nonfibrous) were included in this retrospective study. Preoperative MR images were evaluated using 14 semantic features from. A total of 107 radiomics features were extracted from T2W, FLAIR, post-contrast T1W, post-contrast 3D T1-FFE and ADC map images. Using semantic and radiomics features, five models were created to distinguish HGM from LGM . Classifications were made with support vector machine, logistic regression, k-nearest neighborhood, naive Bayes, decision tree, random forest and neural network algorithms. The same ML algorithms were used to distinguish fibrous and nonfibrous meningiomas. Feature reduction was done with information gain method, and stratified 10-fold crossvalidation was used as an internal validation technique. The performance of ML algorithms was evaluated with the area under the curve (AUC), sensitivity, specificity, and accuracy rates. p <0.05 was considered statistically significant.

The naive Bayes algorithm had the best diagnostic performance in all classification models. The ML model using the radiomics features obtained from post contrast 3D T1-FFE images showed the best results in distinguishing HGM from LGM (AUC:0.906). All models using radiomics features showed better results than the model using semantic features alone. When radiomics features were added to semantic features, performances of ML algoritms were increased. In addition, classification using radiomics features obtained from all sequences had the highest performance in distinguishing LGM as fibrous and nonfibrous (AUC: 944).

ML using radiomics features in distinguishing HGM from LGM showed better results compared to semantic features. In addition, radiomics based ML showed promising results to distinguish LGM as fibrous and nonfibrous.

Anterior temporal lobe resection (ATLR) conduces to remission in 80% of patients with refractory temporal lobe epilepsy (TLE); however, in 30% of cases there is also significant verbal memory loss after speech dominant hemisphere ATLR. The assessment of structures responsible for memory formation through fMRI may help balancing the risk-benefit of ATLR.

We reviewed pre-surgical clinical data and imaging results of 26 patients with refractory epilepsy. For fMRI, a verbal memory encoding paradigm was applied, consisting of words visualization with out-of-scanner recognition assessment. An event-related analysis located brain activations for words remembered (WR) followed by calculation of laterality index (LI) within a mask for fronto-temporal region.

Structural lesion was bilateral on 1 patient, left in 14 patients and right in 11 patients. These right TLE patients were referred to memory fMRI based on EEG incongruences and/or left temporal dysfunction in neuropsychological assessment -5 of them showing temporal dysfunction contralateral to the side of lesion. In 7 patients, there was no activation of mesial structures. Frontal structures were the main contributors to LI in 14 patients, with greater relevance of temporal mesial structures in 6 patients. There was no difference of LI value between gender, age at onset of epilepsy or type of lesion, but we found a correlation between LI and duration of epilepsy (-0.628, p=0.001). A stronger lateralization (>|0.5|) was associated with better performance, with a mean WR of 82% vs. 65% for poorer lateralization, and with significant correlation (p=0.006).

Our results reflect neural plasticity consequential to disruption from temporal lesions over time. This is especially relevant for left side lesions, with compensatory network reorganization, including an anticipated lateralization of verbal memory function to the right hemisphere and/or a quite unexpected impact from left frontal structures. In either, better performance was achieved with stronger lateralization. Bilaterality was seen in patients with left lesions and long disease duration, with globally poor activations and worse performance, suggesting a widespread neural dysfunction besides the primary lesion.

Laterality index in our sample was highly dependent of left frontal structures, which seem to be involved in verbal memory compensation in pre-surgical TLE. This aspect may influence the perspective of fMRI as a prognostic marker of post-surgical deficit in ATLR. Further studies are warranted to clarify the exact significance of this finding and how it may be influenced by technical issues of memory fMRI implementation in each centre. 

This study aims to investigate whether there is a correlation between isocitrate dehydrogenase (IDH) mutation status and magnetic resonance imaging (MRI) findings in glial tumors.

Between January 2015 and November 2019, all brain tumor histopathological reports were reviewed. 66 adult patients with a diagnosis of glial tumor, whose IDH mutation status was determined by polymerase chain reaction or immunohistochemistry were included in the study. Patients were classified as IDH mutation negative or positive. MRI images of patients were evaluated according to VASARI (Visually AcceSAble Rembrandt Images) MRI parameters.

In 66 cases, 49 (%74,2) were negative and 17 (%25,8) were positive for IDH mutation. 39 (%59,1) of the cases were pathologically diagnosed as glioblastoma and 27 (%40,9) were other lower grade gliomas. When all diffuse infiltrative glial tumor cases were grouped according to their IDH mutation status; having less than 33% edema regarding to entire signal abnormality, having more than 33% non-enhancing solid component, having less than 5% necrosis, having less than 5% enhancing solid component, not having avid contrast enhancement, not having thick enhancing margin, not having diffusion restriction, and younger patient age at diagnosis (<46,5) were related to presence of IDH mutation. Some of these evaluation parameters can be time-consuming and challenging for radiologists in routine clinical practise. However with the use of artificial intelligence there is the potential that they can be integrated to analysis of glial tumors.

MRI findings are useful in predicting the status of IDH mutation in glial tumours.

Neuroradiology S6 5 (2020) 62 :S1-107 (Suppl 1) S

In the past decade, major advances in magnetic resonance imaging (MRI) have changed the in vivo evaluation of parkinsonian syndromes, with a growingly important role in individual patient evaluation and pathophysiological characterization. Recent innovations in imaging techniques, including advanced sequences at high and ultrahigh field MRI and modern image analysis algorithms, have enabled visualization of structural and functional brain changes in Parkinson's disease (PD), as well as atypical parkinsonian disorders (APD), arising as important tools to improve sensitivity and specificity of clinical diagnosis, monitor therapeutic interventions and study the progression of the disease.

We reviewed PD imaging advances and suggest a standardized PD MRI protocol for use in clinical settings.

New MRI sequences, particularly those allowing neuromelanin and brain iron visualization, as well as quantitative brain volume measurements techniques, have been shown to be accurate tools for PD evaluation, even in early disease stages, and for differential diagnosis with other parkinsonian syndromes. A multimodal MRI approach of combinations of different methods and techniques sensitive to complementary tissue characteristics may provide higher diagnostic accuracy than single methods and techniques. However, current heterogeneity of MRI studies interferes with the quality of the reports and data collected with implications for the quality of care. Standardization will allow for a better and more homogeneous collection of data for image repositories and for research purposes. It is therefore important to adopt a standardized MRI protocol for the study of parkinsonian syndromes.

The history of Muratoff`s bundle is related to the (superior)-frontooccipital fascicle [(s)-FOF]. Muratoff described in dogs a fiber bundle lying directly above the caudate and beneath the corpus callosum, naming it as a "subcallosal fasciculus" (1). Later, Dejerine described the true [(s)-FOF], without differentiating it from Muratoff`s bundle, nor from the Sachs-Probst bundle that Onufrowicz had mistaken for the FOF (2-6). In post-mortem studies on different species, the Muratoff`s bundle and the [(s)-FOF] were viewed as a single or two separate entities (7,8). Krieg described the juxtacaudate white matter as projectional, anterior, medial and somesthetic nucleus components (9).

To investigate the juxtacaudate white matter lying just inferomedial to the caudate based on the work of Dejerine (2) and Krieg (9). To the best of our knowledge, no study has been performed to clarify its relation to Muratoff`s bundle using diffusion tensor imaging-fiber tractography (DTI-FT).

Method DTI Dataset was acquired on a 3T MRI from a single healthy subject. Tractography was conducted using Fibertrack package (Philips, Best, Netherlands) with Fiber Assignment with Continuous Tracking (FACT) algorithm.

Neuroradiology (2020) S6 6 6 62 :S1-107 (Suppl 1) S

We observed a dense fiber bundle, oriented in anteroposterior direction just inferomedial and parallel to the head and body of the caudate (Figure 2 and 3 , medial nucleus component of the juxtacaudate bundle, Sge`on Figure 1) . Approximately at the level of posterior aspect of the thalamus, these fibers joined the fibers running parallel to it just superolateral to the caudate, known as Muratoff``s bundle (Figure 2 ). After joining, these fibers made a perpendicular turn upwards to reach primary motor-and sensory cortex (Figure 4 ).

Juxtacaudate fiber system, defined as constituting of four major segments, is a complex white matter system playing a major role in mediating the normal basal ganglia function. Its presence is controversial and still the subject of considerable differences in interpretation. In the coronal section it has a distinct appearance, causing it to be named as "substance grise sous-ependymaire" (Sge and Sge`). Opinions differ too, as to whether it should be regarded as an association or projectional tract.

In our preliminary study, we describe the medial nucleus component of the juxtacaudate bundle and its relation to the other components of this system. The understanding of connectional neuroanatomy will help us to understand the anatomic basis of behavior. Further studies have to be done to confirm our results.

Svjetlana Jefic 1 , Daliborka Maric 2 , Dalibor Vranjes 3

There are many neurological conditions with non typical clinical presentation, but with MR finding that can provide a high level of diagnostic certainty. In such cases, the neuroradiologist should overtake the responsibility of having a leading role in the clarification of unclarified clinical situations, sometimes life-threatening.

We present two cases with different outcomes that demonstrate the importance of timely and accurate diagnosis. Case 1: A 74 years old woman with a reduced level of consciousness and suspicion of the acute cerebrovascular incident; MR study showed bilateral asymmetrical changes of both insular cortex and of the left temporal lobe including the hippocampal region in terms of cortical thickening and marked restricted diffusion (Figure 1 ). These findings, according to the literature, are not pathognomonic, but in more than 90% of cases they are suggestive of herpes virus encephalitis and that was our diagnosis. Case 2: A 10 years old girl with sudden onset of limping the right leg and an incident of the lack of bladder control a week after virus infection as an outpatient; MR study showed within the right superior and medial gyrus and along frontotemporal superior cortex both sides there were multiple discontinuing changes with cortico-juxtacortical involvement, a very subtle restricted diffusion and gyriform postcontrast rim-enhancement that spared the cortex. On the left temporal lobe, there were few patchy cortico-subcortical similar appearing lesions (Figure 2 ). Taking into account the age, recent viral infection, current clinical picture, and the pattern of changes, our diagnosis was anti-MOG ADEM with recommendation of urgent hospitalization.

Case 1: The intensive clinical treatment resulted with clinical improvement and the patient is, for our knowledge, in good condition. Case 2: In the hospital, it occurred a severe deterioration with altered mental state and quadriplegia; another MR exam was interpreted by the other radiologist as encephalitis and the girl was treated with acyclovir, ceftriaxone and supportive therapy without any improvement. The biopsy was performed with the result of demyelination; the serology showed MOG positive antibodies. After cortico-therapy and plasmapheresis the girl started to recover. Five months later, the girl felt the loss of vision of the right eye and another MR exam revealed NMO.

An accurate and timely neuroradiological diagnosis can be in some cases life-saving or, vice versa, in otherwise treatable conditions it can lead to severe sequelas, complications and even fatal outcomes.

S6 7 (2020) 62 :S1-107 (Suppl 1) S Introduction Toxic leukoencephalopathy (TLE) is an encephalopathy which is illustrated by a spectrum of clinical and histopathological features associated with structural changes to cerebral white matter injured by a leukotoxic agent. Such substances include antineoplastic drugs, immunosuppressive drugs, antimicrobial agents, environmental toxins, and drugs of abuse [1, 2] . In acute phase, acute toxic leukoencephalopathy can have a characteristic and profound MRI imaging appearacne that is potentially reversible with therapy or removal of the offending agent. The mechanism of toxicity remains speculative and histopathology demonstrates white matter vacuolization and spongiform defects as hallmarks of TLE.

The triad of toxin exposure, neurobehavioral symptoms, and classic radiographic findings should raise suspicion for opioid-induced leukoencephalopathy [1] . Given that exposure to opioids may not have been witnessed or admitted to, t he diagnosis of opioid-induced leukoencephalopathy requires a high degree of clinical suspicion.

The clinical presentation of toxic leukoencephalopathy is extremely variable, ranging from minor cognitive impairment, easily confused with psychiatric illnesses, to severe neurological dysfunction. The lack of imaging in patients with minimal or mild encephalopathic symptoms may account for why this entity may be underdiagnosed in the acute phase. 57-year-old female presented to A&E after found unresponsive at home, with London Ambulance GCS 5 E 1, V1, M3, SATs 90%, 3 days history of cough and fever. Patient admitted to Intensive Care Unit after being intubated for low GCS, CT head showed suspicion of global hypoxic brain injury. Stroke Team suggested for an MRI which suggested acute opiod leukoencephalopathy. Initially, it was assumed that patient has hypoxic brain injury due to being hypoxic after opioid ingestion, however imaging suggested that acute TLE, which is a reversible cause. It is an important learning point for healthcare professional to be aware of acute TLE as a differential. 

Sickle cell disease (SCD) is a hereditary autossomal recessive condition resulting in the formation of abnormal hemoglobin which manifests as multisystem ischemia and infarction, as well as hemolytic anemia. The first presentation is commonly a painful vaso-occlusive crisis: sudden onset of bone or visceral pain due to microvascular occlusion and ischemia, often in the setting of sepsis or dehydration. Sickle cell disease is known to have a wide spectrum of clinical presentations from completely asymptomatic to a severe overwhelming crisis.

We reviewed five cases of sickle cell disease (SCD) evaluated at our hospital for neurological complaints. All patients underwent an MR imaging examination of the brain at 1.5 T with an imaging unit (Siemens Medical Systems) using a standard quadrature head coil.

The age range was 3 to 37years, with a mean age of 18.8 years, with a prevalence of male patients (4:1). Vascular and parenchimal findings comprised territorial infarcts (large vessels and border areas), silent infarts, progressive stenosis -Moya Moya pattern arteriopathy (3), lacunar infarcts (1) and epidural hematomas (1). Bone manifestations included diploic thickening and extramedullary hematopoiesis (1) and bone infarts (1).

Sickle cell disease is among the most common causes of stroke in the pediatric population and approximately 25% of patients with sickle cell disease have a neurological event in their lifetime, many of these in childhood. Infarction or ischemia in border-zone areas is a result from either vasoocclusion within the cerebral vessels or transient reduction in perfusion pressure. The latter might accompany a hypoxic or hypovolemic episode. Extensive infarcts that occur along the distribution of the major vessels typically are caused by endothelial hyperplasia with associated vessel stenosis and occlusion, which are characteristic of cerebral vasculopathy in sickle-cell disease (Moya Moya pattern). Spontaneous epidural hematoma (EDH) is a rare occurrence in patients with sickle cell disease, with a small number of cases reported. Spontaneous epidural hemorrhage is a rare complication of sickle cell anemia and can present with acute headache and neurological symptoms. Bone infarcts, as in other distribution are related to vascular occlusion and bone thickening is a consequence of extramedullary hematopoiesis due to chronic anemia. This is an increasing global health problem, a multiorgan disease with acute and chronic complications. This review highlights the broad spectrum of common and rare imaging findings seen in patients with sickle cell disease that radiologists must recognize. We aimed to analyze how supra-aortic vessel (SAV) status influences intracranial embolus lodgment.

For this experimental analysis, we used a silicon model of the extracranial and intracranial circulations with a complete circle of Willis under a physiological pulsatile flow of a blood-mimicking fluid. Clot analogs were produced from ovine blood, predefined to contain >95 % red blood cells for the associated extreme clot fragility, so that the clots would readily fragment upon injection. We analyzed the sequence of emboli lodgments on injecting 104 clot analogs through the right internal carotid artery (ICA) as patency of the SAVs was changed using the following four SAV conditions: (a) both carotid and vertebral arteries patent; (b) left ICA occluded; (c) ipsilateral ICA occluded, and (d), both vertebral arteries occluded.

A total of 104 clot analogs fragmented into 406 emboli, with no differences among the four SAV conditions tested (P = 0.213). The first embolus lodgment was in the middle cerebral artery (MCA) in 96.2 % of cases (100/104). Location of further embolisms varied significantly depending on SAV patency (P < 0.0001). When the SAVs were permeable, emboli lodged in the ipsilateral anterior cerebral artery (ACA) in 29 % of cases (28/97), in the posterior cerebral artery (PCA) in 11 % (11/97), and in the contralateral circulation in 1 % (1/97). By comparison, ACA emboli were significantly lower when the ipsilateral ICA was obstructed (16%, 14/88), and PCA emboli were more common (25 %, 25/99) when both vertebral arteries were occluded. Previous MCA occlusion was a prerequisite for ACA embolism under all the conditions tested. Similarly, emboli lodged in the PCA more frequently after emboli had first lodged in the MCA and ACA. Finally, contralateral embolism was observed in 37 % of cases (45/122) when flow in the contralateral ICA was stopped, decreasing to 15 % (14/88) when the ipsilateral CA was obstructed (P < 0.0001).

Emboli lodgment was dependent on SAV patency. ACA or PCA embolisms were never isolated events and were associated with at least one previous MCA embolus. Moreover, contralateral thromboembolic lodgment can be seen in patients with ICA occlusion ipsilateral or contralateral to embolus side, which in cases of partial arterial recanalization may obscure t he real cause of the stroke (atherothrombotic vs. cardioembolic).

Neuroradiology S6 9 (2020) 62 :S1-107 (Suppl 1) S 

Elias Johansson 1 , Anders Wahlin 2 , Sebastian Castillo 1 , Laleh Zarrinkoob 1 , Anders Eklund 2 , Jan Malm 1 Introduction Carotid near-occlusion is a severe stenosis causing the distal internal carotid artery (ICA) to reduce in size. Separating symptomatic near-occlusion from ≥50% conventional stenosis (without near-occlusion) is important as recommended management differs. No single CT angiography (CTA) parameter is >90% sensitive and specific for near-occlusion and carotid ultrasound sensitivity is <20%. Assessment of blood flow could be a diagnostic alternative. The purpose of this study was to assess the diagnostic performance of ICA flow measurements with 4D flow MRI for separating near-occlusion and conventional ≥50% carotid stenosis.

Patients with ≥50% carotid stenosis were prospectively and consecutively examined with CTA and 4D flow MRI (n=78). CTAs were reviewed by an observer with near-occlusion expertise blinded to 4D flow MRI findings. CTA-based exclusions were contralateral occlusion (n=4), bilateral nearocclusion (n=1) and uncertain if near-occlusion or conventional ≥50% carotid stenosis (n=6). Degree of stenosis was categorized by CTA findings into conventional ≥50% carotid stenosis and near-occlusion. Flow was ascertained by 4D flow MRI. Ipsilateral ICA flow and flow fraction (ipsilateral ICA flow/ (ipsilateral + contralateral ICA flow) were analysed.

Sixty-seven patients were included, mean age 73 years, 34% women and 82% with symptomatic stenosis. 41 (61%) had conventional ≥50% carotid stenosis and 26 (39%) had near-occlusion. Among ≥50% carotid stenoses and nearocclusions, mean flow in ipsilateral ICA was 199 (SD 46) ml/min and 58 (SD 47) ml/min respectively, mean flow fraction was 46% (SD 7%) and 16% (SD 12%) respectively (both p<0.001, figure) . Diagnostic separation was slightly better for flow fraction than ICA flow, area under the curve was 0.97 (95%CI 0.91-1.0) for flow fraction. A flow fraction ≤0.3 (dashed line in figure) was 96% (25/26) sensitive and 98% (40/41) specific for near-occlusion.

Near-occlusion had lower ICA flow than ≥50% conventional stenosis. Separating near-occlusion from ≥50% conventional by flow fraction ≤0.3 was superior to all previous suggested single parameter thresholds by any diagnostic method. Main novelty was separating near-occlusion in the reference test, omitted in all previous comparisons of phase contrast MRI (with any technique, including 2D and 3D) and degree of stenosis. However, 2D and 3D phase contrast MRI will reasonably perform similarly to the 4D methods we used as analysed measurements are feasible to assess with all three techniques. Although encouraging, these findings should be validated before wide-spread use in routine practice.

Introduction Contrast-induced neurotoxicity (CIN) is an adverse reaction to contrast agents which can occur in the context of diagnostic and therapeutic neurological endovascular procedures. Our goal was to conduct a systematic review of patients reported in the literature diagnosed with CIN after neurological endovascular procedures.

A systematic search of EMBASE and MEDLINE databases was conducted. Inclusion criteria were age ≥18 years; neurological endovascular procedure performed ≤24 hours before ictus; new manifestations compatible with encephalic dysfunction; imaging performed <7 days after ictus; exclusion of other causes; manifestations attributed to pre-existing conditions.

Forty-eight patients were included, of which 60.4% were female and 60.4% had arterial hypertension; median age was 63 years. The most frequent procedure was diagnostic cerebral angiography (n=24, 50.0%) and non-ionic contrast agents were more frequently used (n=40, 83.3%). Twenty-six patients (54.2%) developed clinical manifestations compatible with CIN during or immediately after the procedure, and the most frequent manifestations were encephalopathy, cortical blindness and motor deficit. The most frequent imaging findings were cortical contrast enhancement on Computed Tomography (n=23/42), and cortical sulci effacement (n=18/48). Most patients presented complete clinical recovery (89.6%) at a median time of 3 days. Among patients with neuroimaging changes who underwent follow-up, complete regression of the abnormalities was shown in 81.5% at a median time of 5 days.

CIN is a rare adverse reaction in the context of neurological endovascular procedures. Recognition of clinical manifestations and typical imaging abnormalities, while excluding other causes, is essential for diagnosis. :S1-107 (Suppl 1) S

Mechanical thrombectomy is the standard treatment in acute ischemic stroke due to large vessel occlusion, but there is limited evidence about its efficacy in elderly patients, mainly in nonagenarians. We sought to analyse the safety and efficacy of mechanical thrombectomy in stroke nonagenarian patients and compare the results with the octagenarian ones.

We retrospectively collected consecutive patients with acute ischemic stroke due to large vessel occlusion subjected to mechanical thrombectomy, during a period of 29 months in a single tertiary medical center. Patients were divided in two sub-groups, according to their age: 80-89 and 90 years and older. Successful and futile recanalization, in-hospital complications, functional outcome and mortality at discharge and at 3 months were compared between these two groups. Multivariable analysis was performed to identify independent predictors of functional outcome at 3 months of follow-up, assessed by the modified Rankin Scale. 

No statistically significant differences were found in functional outcomes and mortality at discharge and at 3 months, recanalization and complication rates in octagenarians and nonagenarians. This way, patients should not be excluded from mechanical thrombectomy based on age alone.

Cranial dural arteriovenous fistulas (dAVF) comprise up to 15% of all intracranial vascular malformations, possibly leading to severe disability or even death, due to hemorrhagic or non-hemorrhagic deficits. Given the complex vascular anatomy, additional understanding of its pathophysiology remains crucial to predict and prevent devastating outcomes. Therefore, we set out to explore the relationship between bilateral and/or mixed intracranialextracranial feeders on hemorrhage at presentation.

We reviewed all cranial dAVF patients admitted at our hospital for the past 10 years, excluding carotid-cavernous fistula as these are rarer and have very distinct features. We performed Fisher's exact test and logistic regression for univariate and multivariable analyses, respectively.

We identified 33 cases (64% male; 62.8 ± 12.9 years old). Twelve patients presented with hemorrhage, 8 with non-hemorrhagic deficits, and 13 were asymptomatic. According to the Borden classification, 10 were type I, 1 was type II and 22 were type III. Even though 52% showed bilateral feeders and 52% displayed mixed intracranial-extracranial feeders, only 15% presented both. On univariate analysis, only bilateral feeders showed a trend towards increased hemorrhagic presentation (p = 0.07). On multivariate analysis, also including Borden's classification, this effect was attenuated though still suggestive of increased probability for hemorrhagic presentation (OR = 3.97, p = 0.11), as was Borden grade (OR 2.93, p = 0.08).

We presented a fairly large cohort of patients with cranial dAVF and showed additional angiographic features beyond the ones included in traditional dAVF classifications. Though the latter provide reasonable prognostic assessment, particularly if contextualized with the patient's symptoms, new features may provide further insight into the risk of bleeding. We showed a trend for increased hemorrhagic presentations on fistulas with bilateral feeders, though not for mixed or both combined. This effect was partly attenuated in multivariable analysis, and even though it was non-significant this could be due to small sample size, statistically speaking, as Borden classificationa documented good prognostic marker of fistula aggressivenesswas also not significant in our analysis. Additional analyses of larger cohorts, possibly multicentric, may provide further insight into this matter.

Bilateral feeders for dAVF may increase the risk of hemorrhage at presentation, even when Borden grading is accounted for. Due to a small sample size, even if rather large for dAVF, analyses on larger cohorts are needed to elucidate this matter.

Andre Pinho Araujo 1 , Francisco Sabenca 2 , Renata Souto Silva 3 , Marta Gomes Rodrigues 1 , Andre Cunha 1 , Sergio Castro 1 , Manuel Ribeiro 1

The pathophysiologic mechanisms of posterior circulation ischemic events among patients with chronic disease of proximal vertebral artery remains poorly understood. There are several presumed mechanisms in which artery-to-artery embolism plays a role. A significant percentage of patients with artery-to-artery embolism in the posterior circulation have minor or major neurologic deficits at discharge and there is few data in the literature on long-term outcome in patients with V1 segment occlusions.

Neuroradiology (2020) S72 62 :S1-107 (Suppl 1) S

We report two cases of V1 segment chronic occlusions that became symptomatic despite best medical treatment.

Two male patients with 65 and 74 years old presented with posterior circulation strokes associated with V1 segment chronic occlusions. One patient had a right P2 occlusion in the setting of acute stroke combined with left V1 occlusion and a contralateral hypoplastic vertebral artery which terminated mainly in PICA; a 6-month-prior Angio-CT revealed atherosclerotic disease in the ostium, favoring progression of the disease and consequent occlusion of this artery. The other patient presented with a top basilar artery occlusion in the setting of acute stroke associated with a right subclavian artery occlusion, a severe stenosis of the right vertebral artery ostium and a chronic occlusion of the left proximal V1 segment (with distal collateralization). In both cases there were no significant intracranial atherosclerotic vertebrobasilar disease, but there was filling of distal vertebral artery segments due to arterial collateralization from cervical branches. They had significant vascular risk factors and both were on a dual antiplatelet regimen due to previous strokes. Endovascular recanalization was unsuccessful because of the inability to catheterize the access vessels due to its chronic disease.. The 24-Hour Holter monitoring did not show abnormalities in both patients, nor did the transthoracic echocardiogram in the first patient; the second patient had a moderate to severe aortic valve stenosis and an interauricular communication.

We believe that artery-to-artery embolism due to turbulent flow in the chronically occluded proximal vertebral arteries was the cause for the posterior circulation strokes on both of our patients.

Chronic V1 segment disease is sometimes associated with distal turbulent arterial flow, which can be directly related to artery-to-artery embolism, playing a role in the mechanism of injury in symptomatic proximal vertebral artery disease. It is crucial to understand the natural history of bilateral V1 segment occlusion/severe stenosis for the long-term management of these patients. 

International guidelines support early use of decompressive hemicraniectomy (DHC) in acute malignant ischemic stroke (aMIS) in patients aged 60 years or younger. Deciding to perform DHC on older patients is controversial, since current evidence shows elderly have poorer outcomes, with higher mortality and severe disability rates.

To describe our institution experience with DHC in patients with aMIS and compare outcomes between age groups.

Retrospective study of patients subjected to DHC at our institution from January 2016 to June 2019 in the setting of middle cerebral artery aMIS. Patients were stratified in three age groups: ≤60, 61-70 and >70 years. Patient demographics, stroke onset and imaging data, procedural times and outcomes were collected. Primary endpoints were the modified Rankin scale scores (mRS) and mortality rates (MR) at 12 months follow-up.

Surgery was performed in 44 patients with ages ranging from 50 to 82 years: 50.0 % with 60 years or younger; 29.5 % with 61-70 years; and 20.5 % with more than 70 years. Median ASPECTs at hospital admission was 5. Fifteen patients (34%) had been submitted to mechanical thrombectomy (MT), 7 (16%) to rtPA and 11 to combined therapy (25% 

Real-life data mirrors clinical trials results, even with less restrictive inclusion criteria. Accurate and timely preoperative risk assessment is essential to select older patients that might benefit from surgery and avoid futile procedures.

Neuroradiology S73 (2020) 62 :S1-107 (Suppl 1) S Introduction Among various venous routes, the Inferior Petrosal Sinus (IPS) is the most widely used approach to treat arteriovenous fistulas of the cavernous sinus. Frequently observed partial or complete thrombotic IPS occlusions may represent a technical challenge. In addition, anatomic variations such as the aberrant IPS may represent a potentially frustrating obstacle during transvenous occlusion (TVO).

Five patients with, both direct and indirect arteriovenous cavernous sinus fistulas (CSFs) underwent TVO procedures using the traditional IPS approach. In all cases uni-or bilateral aberrant IPS termination into the extracranial internal jugular vein (IJV) were encountered.

In all patients the IPS approach could be successfully used to access the CS and to occlude the AV shunts. Arterial Road Mapping with obliques views and 3D-DSA rather than jugular phlebography were useful tools to identify the exact location of points of entry into the IJV allowing for retrograde catheter navigation into the CS.

The aberrant IPS is a rare anatomic variant and a potentially challenging problem for endovascular operators. However, when recognized by proper imaging techniques, it can be equally mastered and used as approach for TVO in CSFs. Introduction Partially thrombosed intracranial aneurysms (PTIAs) are a group of complex brain aneurysms which have varied presentations, an unclear prognosis and are a challenge to treat by both endovascular and surgical techniques. We aim to show that these aneurysms should be regarded as a collection of several subtypes with unique features rather than a single entity. We also evaluate the impact of flow diversion (FD) treatment in their management.

We conducted a retrospective review of 48 PTIAs from 48 patients who were managed between 2010-2019. The clinical presentation, evolution and management of the aneurysms were analyzed. Based on our data, the aneurysms were classified into 4 separate subtypes according to their morphology, location and etiopathogenesis. The impact of different endovascular treatment options in PTIA management was also evaluated.

PTIAs (including the thrombosed component) had a mean diameter of 23mm (range: 10-45mm) and were commonly found in the internal Neuroradiology (2020) S74 62 :S1-107 (Suppl 1) S carotid arteries (41.7%), vertebrobasilar arteries (29.2%), and middle carotid arteries (20.8%). They contained a thrombus that was either intramural (Type 1, 2, 3) or intraluminal (Type 4). Type 1 aneurysms, the most common PTIA subtype, evolve as a result of recurrent aneurysm rupture and pseudomembrane formation. Type 2 aneurysms arise due to repeated arterial wall dissection. Inflammatory changes along the vasa vasorum is responsible for the evolution of Type 3 aneurysms. Type 4 aneurysms is based on hemodynamic changes within a giant aneurysm. The proposed classification of PTIAs demonstrated a good correlation between the clinical presentation and the evolution of the different subtypes and helped to guide management i.e. endovascular treament options. FD was performed in 31 of the 42 treated cases with good results, ranging from aneurysm size stabilization to complete collapse of the aneurysm.

Unlike the morphology-based classification scheme of PTIAs by Lawton et al. (2005) , this new classification scheme confers the advantage of including the parameter, etiopathologies of PTIA subtypes, which importantly impact an aneurysm's clinical presentation, natural history and patient management. As our case series reviewed endovascular treatment methods only, its major shortcoming is the absence of data on neurosurgical management which might have a role to play in selected cases.

The new classification of PTIAs, based on their morphology, location and possible etiology, facilitates better understanding of the clinical presentation, prognosis and endovascular treatment approach of these aneurysms. FD was found to be a safe and effective treatment method across all PTIA subtypes.

Severe bleeding after blunt maxillofacial trauma is a rare but lifethreatening event. We present a case in which transarterial embolization (TAE) of the left ECA and the right sphenopalatine artery (SPA) using a novel embolization agent PHIL 25 (25% precipitating hydrophobic injectable liquid; MicroVention, Tustin, California) was performed to effectively stop an intractable life-threatening bleeding following blunt maxillofacial trauma.

A 20-year-old male was brought to the emergency department after a 20 m (65 ft) fall. He presented with exsanguinating bleeding from the left ear not amenable to conventional treatment options, resulting in severe hypotension despite agressive fluid resuscitation. CTA identified the left internal maxillary artery (IMA) and the right SPA as the sole sources of major blood loss. As the bleed origins were considered not amenable to surgical repair due to extensive soft tissue damage and swelling, a decision to perform TAE was made. DSA showed a laceration of the left ECA with ensuing 3 × 4 cm pseudoaneurysm continuing into the proximal part of left IMA, and a right SPA laceration with two small accompanying pseudoaneurysms. The two culprit arteries were then superselectively catheterized and embolized. ECA was embolized under temporary proximal balloon occlusion using platinum coils and PHIL 25, while the SPA was embolized without prior balloon vessel occlusion using PHIL 25 only. Embolization of both lesions was followed by complete cessation of the ear bleeding. Postembolization imaging showed total exclusion of the lacerated vessels, complete patency of all proximal vessels, no collateral pathways to the pseudoaneurysm and no other origins of bleeding.

Coils and gelatin foam (Gelfoam) have historically been the EAs most frequently used for trauma-related ECA TAE. During the past two decades, however, novel liquid EAs have been introduced, including Onyx (Medtronic, Dublin, Ireland) and PHIL. Both polimerize independently of coagulopathy and offer a well-controlled application, reducing the possibility of an extracranial-intracranial injection. PHIL seems superior to Onyx as it is supplied ready-to-use (no shaking is necessary), requires lower volume to achieve the same extent of embolization, is faster to fully precipitate, produces fewer artefacts in postinterventional imaging, and is not dark coloured which diminishes the possibility of skin discoloration. To our knowledge, this is the first case report of PHIL use for trauma-related TAE in the ECA territory. Further clinical evaluation is thus warranted to fully evaluate its feasibility, safety and efficacy in this particular setting.

Neuroradiology S75 (2020) 62 :S1-107 (Suppl 1) S 

The experience and evidence of the direct thrombus aspiration using a distal access catheter as a type of mechanical thrombectomy technique for managing thromboembolic complications during endovascular aneurysm treatments is still limited and little literature is available regarding this topic. We present a case of thromboembolic occlusion of the left middle cerebral artery during the treatment of two left internal carotid artery aneurysms, which was successfully treated by direct thrombus aspiration using a Sofia (5F) distal access catheter (MicroVention Europe, Saint-Germain-en-Laye, France).

A 54-year-old woman was admitted to our hospital for the elective endovascular treatment of two intracranial left internal carotid artery aneurysms. She was prepared for the intervention using the dual oral antiplatelet therapy since the placement of flow-diverting stent was planned. During the intervention, on working-projection preprocedural angiograms, the thromboembolic occlusion of the M1 and M2 segment of the left middle cerebral artery was perceived. This complication was successfully managed using already positioned distal access catheter Sofia (5F) in the left internal carotid artery intended for the support of opening the flow-diverting stent. The distal access catheter was placed in the M1 segment of the left middle cerebral artery, with the tip of the catheter in contact with the clot. The thrombus was manually aspirated for 2 minutes and Thrombolysis in Cerebral Infarction scale 3 flow was restored. Afterwards, the flow diverting stent Derivo (Acandis GmbH, Pforzheim, Germany) was placed over the necks of the two left internal carotid artery aneurysms and the flow-diversion effect was observed on the final angiograms. Control CT scan showed no signs of pathologic densities and the patient was discharged from the hospital without any neurological symptoms.

This case showed that the use of Sofia (5F) distal access catheter already positioned for aneurysm treatment is a safe and effective method to perform zero-delay direct thrombus aspiration in thromboembolic complications that may have occurred during endovascular aneurysm treatments.

Taha Yusuf Kuzan 1 , Feyyaz Baltacioglu 1 Introduction Intrasaccular flow disruption is a novel option in the treatment of intracranial wide-necked bifurcation aneurysms. In this study, we aimed to evaluate the efficacy, safety, long-term results, and the effect of the aneurysm axis on the treatment response in patients treated with a Woven EndoBridge (WEB) device in our center due to intracranial wide-neck bifurcation aneurysm.

After obtaining approval of our study from the local ethics committee, clinical information, DSA images of the patients treated with WEB device in our unit between January 2014 and February 2018 were evaluated retrospectively. Clinical information and demographic data of the cases were obtained from the hospital information management system. The treatment response with the WEB device was classified as complete occlusion, adequate occlusion, partial occlusion, neck remnant, and aneurysm remnant. Complete occlusion, adequate occlusion, partial occlusion, and neck remnant were defined as acceptable aneurysm occlusion. The relationship between the aneurysm axis and the parent artery was examined in predicting the WEB device treatment response.

Twenty-five aneurysms of 22 patients, 7 of whom were within the scope of the WEB-IT study, were included in the study. 18 of the cases were female and 4 were male. The age of the patients ranged between 30 and 84, and the mean age was 56.4±11.4. Additional endovascular treatment was applied to one of the target aneurysms treated with the WEB device in the same session as the WEB, and 4 at an additional session during the follow-up. Proceduralperioperative complications were observed in 3 of 25 aneurysms treated with the WEB device. No mortality was seen due to the aneurysm during the procedure and at 2 years of follow-up. Acute aneurysm occlusion was observed in 5 (20%) of 25 aneurysms in the perioperative period. In the follow-up, an acceptable level of aneurysm occlusion was observed in 16 (66.7%) of 24 aneurysms at the 6th month, 17 (73.9%) of the 23 aneurysms at the 12th month, and 20 (87%) of the 23 aneurysms at the 24th month. No statistically significant relationship was determined between the aneurysm axis and parent arteries in determining the treatment response.

WEB devices have their own technical difficulties and rare complications. However, WEB devices are an effective and reliable endovascular treatment method, with low complication and mortality rates, for the treatment of widenecked bifurcation aneurysms that are complex and difficult to treat with a high aneurysm occlusion rate. 

Patients aged ≥90 represent a growing stroke population characterized by a higher frailty. Although endovascular treatment (ET) is a cornerstone of the management of acute ischemic stroke (AIS) related to large vessel occlusion, the benefit of reperfusion among nonagenarians remains uncertain, since these patients have been excluded or underrepresented in past thrombectomy trials. The aim of this study was to investigate the feasibility and safety of ET in nonagenarians with AIS.

A retrospective analysis of clinical and imaging data from all patients aged 90 or older presenting with AIS who underwent ET, between January 2015 and February 2020, was performed. Baseline characteristics, procedural and functional outcomes were analyzed and compared with younger patients (70-89 years old). Successful recanalization was defined as a modified Thrombolysis in Cerebral Infarction score of 2b/3 and good functional outcome was defined as a modified Rankin Scale score of 0 to 2 or equal to prestroke value at 90 days.

Of 1030 patients with AIS submitted to ET, 5.05% (n=52) were ≥90 years old, and 39.13% (n=403) were 70-89 years old. Median National Institutes of Health Stroke Scale (NIHSS) score on admission was higher in the nonagenarians (20 versus 18, P=0.047). Successful revascularization (71.15% versus 78.91%, P=0.204) and good functional outcome (28.85% versus 30.59%, P=0.797) were achieved more often in those <90 years old, but differences did not reach statistical significance. Also, there was no significant difference in periprocedural and postprocedural complications, length of stay and mortality. Logistic regression analysis showed that baseline NIHSS (P<0.001), baseline Alberta Stroke Program Early Computed Tomography Score (P<0.001), number of passes (P=0.032), and procedure duration (P=0.016) were significant predictive factors, but being ≥90 years old, successful revascularization and use of tissue plasminogen activator were not. In those ≥90 years old, initial NIHSS (P=0.024) was identified as an independent predictor for good functional outcome.

Our findings suggest that ET in nonagenarians with acute ischemic stroke is safe and effective and outcomes are not different compared to younger patients. Decision making for thrombectomy in these patients should consider age as one factor, rather than an exclusion criterion, with special regard to initial NIHSS.

Harol Sotelo Benavides 1 , Pablo Pfister 1 , Jairo Hernandez Pinzon 1 , Mercedes Serra 1 , Hernan Chaves 1 , Claudia Cejas 1

Diagnostic imaging methods have been used to determine qualitative and quantitative parameters of cerebral thrombus in the setting of stroke. The aim of our work is to establish if there is a correlation between intrinsic thrombus features on computed tomography (CT) with reperfusion and functional outcome in patients with acute ischemic stroke who underwent mechanical thrombectomy (MT).

We retrospectively searched for consecutive patients in our database between January 2015 and December 2019. We included patients with an age over 18 years, clinical suspicion of stroke and treated with MT. We excluded patients who did not have a CT previous to MT and patients with non-middle cerebral artery (MCA) stroke. Imaging features evaluated included thrombus location (M1, M2 and/or M3 segments), thrombus density (in Hounsfield units, HU) and thrombus length (<5, 5 to 10 and >10 mm). Outcome was evaluated with the modified Treatment in Cerebral Ischemia score (mTICI) and modified Rankin Scale (mRS) at 90 days. Successful reperfusion was defined as mTICI scores 2b or higher. Good functional outcome (FO) was defined as an mRS between 0 and 2.

Result 12 subjects were included, 8 had successful reperfusion and 9 had good FO. Subjects with good FO were slightly younger (mean: 72 y/o) than subjects with bad FO (mean: 77 y/o). Mean thrombus density was 55 HU in successful reperfusion subjects and 66 HU in unsuccessful reperfusion subjects. Mean thrombus density was 59.5 HU in good FO subjects and 56.3 HU in bad FO subjects. In subjects with successful reperfusion, thrombus length was < 5 mm in 62,5%, 5 to 10 mm in 12,5% and > 10 mm in 25%, and in subjects with unsuccessful reperfusion, thrombus length was < 5 mm in 50%, 5 to 10 mm in 25% and > 10 mm in 25%. Thrombus was located on M1 in 75% of subjects and on M2 in 25% of subjects in both successful and unsuccessful reperfusion groups.

Thrombus characteristics on CT have been positively correlated with outcome after reperfusion in patients with stroke. However, we found no statistically significant differences in reperfusion (measured by mTICI) and/or functional (measured by mRS at 90 days) outcome. Further work should be done to clarify these contrasting results.

In this series of patients with MCA stroke, we found no statistically significant difference in thrombus characteristics on CT and clinico-radiological outcome after MT.

Thijs van der Zijden 1 , Caro Franck 1,2 , Maurits Voormolen 1,2 , Olivier D'archambeau 1 , Frans Van Den Bergh 1 , Maarten Niekel 1 , Annelies Mondelaers 1,2

International directives and legislations impose structured data registration about administered radiation doses to patients undergoing radiological procedures. Collection of these data enables not only radiation dose registrations in individual patients, but also establishment of local, regional or national diagnostic reference levels (DRL) for benchmarking individual hospitals or institutions. In contrast to conventional radiology and computer tomography (CT), data about radiation doses in the interventional neuroradiology (INR) angiosuite are scarce. Lack of standardization with differences in complexity among procedures results in limited data about DRLs. Therefore, this study aimed at evaluating radiation doses in INR in our hospital.

Within this retrospective single-center study, three procedures, i.e. diagnostic cerebral angiography (DCA), cerebral aneurysm embolization (CAE) and endovascular stroke treatment (EVT), were examined. Dose area product (DAP) values were extracted from dose management software for 30 patients per procedure (90 patients in total) to establish local DRLs (75th percentile). Distinction was made between procedures with and without flat-detector CT (FD-CT). Brain doses of 2D acquisitions were estimated according to Sanchez et al. [1] . Brain doses of 3D acquisitions were calculated using Monte Carlo simulations in an adult mathematical phantom. Values were compared with typical brain doses resulting from multi-detector CT (MD-CT) angiography and perfusion examinations.

Neuroradiology (2020) S78 62 :S1-107 (Suppl 1) S Result Significant differences in exposures exist between procedures with and without FD-CT. When 3D imaging was used, brain doses were on average 3 times higher for diagnostic, and 1.8 times higher for EVT. For all CAE procedures FD-CT was used. Statistics of delivered DAP and brain doses for each procedure are given in Table 1 . Based on the 75th percentile of our data, local DRLs were 44, 104 and 96 Gy.cm 2 for diagnostic, CAE and EVT respectively (Figure 1 ). Mean brain dose for a single 3D DSA acquisition used in DCA and CAE was 11 mGy. In MD-CT, angiography of the Circle of Willis results in a brain dose of 20 mGy. For EVT, FD-CT parenchymal blood volume (PBV) acquisitions result in an average brain dose of 17 mGy. Brain doses of MD-CT cerebral perfusions are on average 3 times higher (50 mGy).

Use of FD-CT in INR should be considered carefully, because of its significant impact on delivered patient radiation doses. However, brain doses in MD-CT are 2-3 times higher compared to FD-CT.

[1] Sanchez RM, et al. Brain radiation doses to patients in an interventional neuroradiology laboratory. AJNR. 2014;35 (7):1276-1280. doi:10.3174/ ajnr.A3884

To identify possible relevant factors for tremor relapse after treatment with tcMRgFUS.

We identified all cases of tremor recurrence among the patients treated in the period February 2018 and January 2020. We analyzed the demographic (gender, age) and clinical (etiology, disease duration) characteristics, compared to patients who did not present recurrences. For the analysis of imaging and procedural factors we selected a control group matched for clinical and demographic characteristics. Among the procedural factors we analyzed: SDR, number of active elements, skull surface area, number of sonications, maximum duration of sonication, maximun energy (J) and power (W), and accumulated thermal dose at the end of the treatment. For the imaging evaluation, we measured the size of the thalamotomy lesions with follow-up at 24hrs and 1,3,6 and 12 months. For the evaluation of the DRTt, a tractographic reconstruction was performed before and after (1 month) treatment; DTI parameters (FA, ADC, RD and AD) were also recorded.

Concerning the clinical and demographic characteristics, we did not find statistically significant differences in gender, age and disease duration. 73% of patients with tremor relaps were Parkinson's disease patients. We did not find statistically significant differences in procedural and imaging factors between the two groups. In the tractographic evaluation we found an eccentric position of the DRTt (with respect to the thalamotomy lesion) in patients with tremor relapse.

The most relevant determining factors for tremor relapse after MRgFUS thalamotomy appears to be the tremor from Parkinson's disease and inaccurate thalamic targeting. No differences emerged regarding the size of the thalamotomy lesion.

Temporal bone paragangliomas are divided according to their location and relationship with the jugular bulb into tympanic/tympanomastoid and tympanojugular. The decision to surgically intervene depends on various factors and a thorough clinical evaluation complemented by imaging is of key importance. Angiography findings are invaluable in order to assess the vascular supply of the tumor, and endovascular pre-surgical embolization plays a relevant role to reduce blood losses intra-operatively, especially in tympanojugular paragangliomas. Our aim is to characterize the angiographic findings of temporal bone paragangliomas and to study possible relationships with the follow-up outcome.

We reviewed the clinical and imagiological findings of patients with the diagnosis of temporal bone paraganglioma followed at our center between 2011 and 2019. Data regarding demographics, treatment modality and follow-up Neuroradiology S79 (2020) 62 :S1-107 (Suppl 1) S was collected. Paragangliomas were classified according to the Modified Fisch classification. Pre-surgical embolization images were reviewed.

Twenty-five patients were identified, with a mean age at diagnosis of 60 ± 14.1 years old. According to the Modified Fisch classification, paragangliomas were class A (n=1), class B (n=6) and class C (n=16). Intracranial tumor extension (class D) was observed in 9 patients, with intradural extension in one-third of these cases. Twenty patients were treated surgically, 14 of which performed pre-surgical angiography. The number of arterial feeders varied between 1 and 5, arising mainly from the neuromeningeal trunk of the ascending pharyngeal artery and from the posterior auricular artery. Mean percentage of tumor devascularization (embolization percentage) was 80% (± 17.6). Embolization percentage showed a strong inverse correlation with the number of arterial feeders (r=-0.75; p=0.002), with no relevant relation with the size of the lesion. Complete surgical excision at follow-up was achieved in 90% of the patients. Residual/relapsing paragangliomas were larger at presentation and showed more arterial feeders (both p<0.05), but the embolization percentage did not differ from the non-relapsing ones.

Our case-series suggests that the number of arterial branches feeding the tumor and the initial size of the lesion are more associated with recurrence than the embolization percentage obtained pre-surgically.

Paragangliomas of the temporal bone are slow growing but locally destructive lesions and the treatment goal is to obtain complete surgical excision. The number of arterial feeders and the initial size of the tumor may be important when predicting residual lesion or relapse after surgery.

To present the case of a patient suffering from Essential Tremor subjected to bilateral thalamotomy by tcMRgFUS for bilateral treatment of tremor of the upper limbs.

Method 63-year-old male patient suffering from Essential Tremor from about 8 years with bilateral symptoms of the upper limbs and resistant to drug therapy. The patient underwent 2 thalamotomy sessions, using tcMRgFUS, with treatment of the left Vim and -2 years later -the right Vim. Periodic clinical and instrumental MRI checks were carried out, to assess the intensity of the tremor, the quality of life and the evolution of the thalamotomic lesions.

The first treatment (left Vim) determined an immediate control of the tremor of the upper right limb, stable and persistent up to 2 years after the treatment, without significant adverse effects. The follow-up MRI documented a progressive and regular evolution of the edema and the lesion. The second procedural thalamotomy (right Vim) was facilitated by the presence of the contralateral thalamotomy lesion, used as an anatomical landmark. Even in this case, the control of the tremor of the left upper limb was complete and immediate. The clinical evaluation documented a stable and bilateral reduction of tremor (approximately 78.6%), at 24 hours, 1 month and 6 months.

Discussion & Conclusion our preliminary experience suggests the possibility of using the tcMRgFUS method for the bilateral treatment of tremor refractory to medical therapy. However, further studies are needed to evaluate any long-term complications related to bilateral thalamotomy, as well as the possibility of extending treatment to PD patients. To evaluate the validity of the use of direct targeting by tractography of the dentato-rubro-thalamic tract (DRTt) compared to indirect targeting using atlasbased coordinates. We retrospectively evaluated 12 patients submitted to MRgFUS thalamotomy in which target shifts were performed during treatment for non-optimal clinical response. In each treatment we recorded the initial indirect coordinates (LL, AP, CC) and the coordinates where the clinical response (tremor reduction) was obtained. The coordinates of the DRTt were measured, after tractographic reconstructions, at the level of the thalamus. The mean indirect coordinates were 14.2 RL, 7.1 AP, 1.8CC. The coordinates of the clinical response were 13.8 RL, 6.8 AP, 1.5 CC. The average coordinates measured on the tractographic reconstructions were 13.7 RL, 6.9 AP, 1.4 CC. The target coordinates obtained by tractographic reconstruction of the DRTt showed greater agreement with the coordinates of the clinical response, compared to the indirect coordinates. These preliminary results should be corroborated by future prospective studies to validate the use of intraprocedural direct targeting during MRgFUS thalamotomy.

It is often seen low hemoglobin (Hb) level in patients admitted with acute stroke. It is known that anemia has a negative impact in multiple ischemic mechanisms, and has recently been identified as an independent risk factor for ischemic stroke, causing worse functional outcomes and increasing mortality. The author aimed to analyze the effects of low hemoglobin levels on admission of stroke patient undergoing mechanical thrombectomy (MT).

62 :S1-107 (Suppl 1) S

In this retrospective study, we collected data from 83 adult patients with anterior circulation strokes who were submitted to MT between January 2019 and December 2019. We compared the functional outcome [modified Rankin Scale] and the reduction of NIHSS score at hospital discharge in patients with different levels of Hb on hospital admission. The Hb levels were analyzed as a categorical variable divided in three categories: ≥12,5 g/dL; between 12,5 g/dL and 10 g/dL; ≤10 g/dL. We also divided the stroke severity in three groups based on NIHSS pre-procedure: Severe (NIHSS ≥16); Moderate (NIHSS 9-15); Mild (NIHSS 1-8). It was considered a good functional outcome a modified Rankin Scale (mRS) between 0-2.

Result and Discussion 44 of all patients were female (53%). The median age of the subjects was 79 (13) years old. The majority of subjects suffered a severe stroke (75%). Approximately one-third (34,9%) of all patients had Hb levels equal or below 12,5 g/dL (24,1% had Hb 12,5-10 g/dL; 10,8% had Hb ≤10 g/dL). There was a statistically significant improvement on the NIHSS pre-procedure and on discharge in all hemoglobin groups, except patients with Hb ≤10 g/dL [Hb ≥12,5 g/dLp value <0,001; Hb 12,5-10 g/dLp value 0.02; Hb ≤10 g/dLp value 0.396]. It seems to be an apparent correlation between low hemoglobin levels and lower improvement on NIHSS values, however we did not verify a statistically significant difference between the groups. This aspect could be better analyze with a larger sample. Concerning the functional outcome and mortality, there was not a statistically significant difference between the groups.

With this study we suggest that there is a questionable beneficial effect of MT in patients with Hb <10 g/dL, at least applying the same selection criteria, namely time of onsetarterial puncture. Although it has been observed a statistically significant improvement on NIHSS after MT in stroke patients with Hb >10 g/dL, we didn't verified a significant difference in functional outcome or mortality.

Mechanical thrombectomies (MT) have been one of the central management strategies for selected acute ischemic stroke patients. While data supports the efficacy of MTs in patients between the ages of 18-65, MT data is limited in patients older than 80.

Retrospective analysis was conducted on patients from January 2018 to December 2019. The patients were subdivided by age into an octogenarian group (OG; ages 80 or above) and a non-octogenarian group (NOG; ages 18-79). Primary outcomes were measured with NIHSS (NIH stroke scale) scores difference between 24 hours post-MT and at arrival. Additional outcome measures included post-MT thrombolysis in cerebral infarction (TICI) scores 2B/3, mortality, and symptomatic intracranial hemorrhage (sICH) within 24 hours. Hypertension, diabetes, atrial fibrillation, prior TIA/CVA history, alcohol and smoking history, location of the occlusion, IV tPA administration, ASPECT scores (pre and post thrombectomy), door to groin puncture time, and door to recanalization time were analyzed. A two-sample t-test was used for statistical analysis.

Result 156 NOG and 46 OG patients were included in this study. The median age of NOG and OG groups were 63 and 86, respectively. OG patients were more likely to be female (69.6%) than NOG patients (females=48.73%).

Hypertension was the single most common risk factor in both groups, with 69.62% in NOG and 78.26% in OG. Statistically significant improvement in NIHSS score (mean ± SD [median]) after MT was seen in both OG (from 14.0. ± 6.8[14] to 12.5 ± 8.8[10]) (p<0.0001) and NOG (16.2 ± 6.1[16]) to 12.0 ± 8.1[11.5]) (p=0.0078). However, the degree of NIHSS improvement of OG (mean ± SD = -4.24 ± 6.22) was larger compared to those of NOG (-1.44 ± 8.79) (p=0.017). All-cause mortality rates of OG (39.3%) was higher than NOG (19%). There were no differences in symptomatic ICH rates, TICI 2b/ 3 rates, IV tPA administration rates, median door to groin puncture (NOG: 1:26 vs. OG: 1:17) and median door to recanalization (NOG: 2:17 vs. OG: 2:15) between the two groups. MCA occlusion was the most common presentation in both groups (NOG: 68.4% v OG: 87.0%).

Octogenarians showed a better degree of NIHSS improvement in 24 hours post-MT which may indicate that MT can be an effective treatment strategy in acute strokes in all age groups. However, the post-MT mortality rate was higher in octogenarians, thus warranting additional investigation.

The endoscopic endonasal transsphenoidal approach for pituitary tumors and sellar lesions has always been a challenge due to complex anatomy of this area and the important neurovascular structures around the sella. Imaging techniques, especially computed tomography (CT) images, provide the surgeon with information about the main anatomical structures and variants which knowledge is essential for planning the surgery, avoiding complications or failure of the endoscopic surgery. Our goal was to propose an easy mnemonic rule that encompass all these structures to systematically apply when reporting.

By considering the different anatomic structures and variants that radiologists should evaluate when reporting a preoperative CT scan of patients before an endoscopic endonasal transsphenoidal approach, we constructed the acronym SPHENOID with the initials of major items to be considered.

Radiological evaluation for planning endonasal transsphenoidal surgery should always consider the following aspects, and their initials form the word SPHENOID (figure 1): Septa in the sphenoid sinuses; Pneumatization extensions and recesses; Hamberger classification of pneumatization in the sagittal plane; Endonasal structures, either pathologic or variants conditioning the Neuroradiology S81 (2020) 62 :S1-107 (Suppl 1) S nasal phase of the surgery; Neurovascular structures to be considered; Onodi cells to avoid accidental intracranial penetration; Invasion of cavernous sinus as an essential limitation to complete resection and; Dehiscences that could increase the risk of vascular or neural damage.

Mnemonic are learning tools to completely remember important information. They should be considered not only a learning resource, but also a working tool. By systematically applying the "SPHENOID rule" radiologists can easily remember all the aspects that a report of a preoperative CT scan should include in patients before endoscopic endonasal transsphenoidal approach surgery.

Using the proposed "SPHENOID" mnemonic facilitates the systematic assessment of the anatomic structures and landmarks to take into account in the endonasal transsphenoidal approach.

Vascular anomalies are the most common cause of pediatric soft-tissue masses. The current classification of these anomalies considers two categories: vascular tumors and vascular malformations. Although their diagnosis can be clinical, magnetic resonance (MR) imaging is often needed for better characterization.

The authors present two case reports of a lymphatic-venous malformation and a congenital hemangioma of the oral cavity in pediatric age. After bibliographic review, we discuss it in accordance with the ISSVA classification (2018) and focus on their clinical characteristics and MR imaging findings.

In the first case, we present a 2-month-old baby admitted for "multiple masses under the tongue" since birth and increasing in size over time. The physical exam and MR imaging findings (figure 1) were compatible with the diagnosis of a multiloculated macrocystic lymphatic-venous malformation. Then, we present a case of an eleven year-old infant with a painless mass of his tongue. The lesion has remained constant in size with time, growing with the child. Intraoral examination (figure 2) and MR imaging findings (figure 3) confirmed a mass in the upper right surface of the tongue and a diagnosis of a noninvoluting congenital hemangioma was made.

Vascular anomalies often involve the head, neck and oral cavity. Within vascular malformations, venous and lymphatic malformations are the most common. They are low flow vascular malformations that usually appear hyperintense on T2-weighted and STIR, and hypo-isointense on T1-weighted MR images. Minimal enhancement may occur after contrast. The treatment may include percutaneous sclerotherapy or surgery. Congenital hemangiomas develop in the fetal period and are fully developed at birth. The non-involuting subtype grows proportional with the child without regression. On MR images, they appear hypo-isointense on T1-weighted, and hyperintense on T2weighted and STIR. Homogeneous enhancement may be seen after gadolinium administration. Surgery or laser therapy can be done when indicated.

Vascular anomalies are the most common cause of pediatric soft-tissue masses and they should be considered in the differential diagnosis. MR imaging is important for the correct diagnosis and characterization, which is crucial for follow-up and treatment.

Nasal masses in children evoke a wide variety of conditions. They can arise from any element of the nose, often presenting as nasal dorsum or ala tumefactions or as obstructing lesions located within the nasal cavity. Alternatively, they can represent lesion extension from adjacent structures, such as the paranasal sinus, the maxilla and the nasopharynx. Clinical presentation is frequently non-specific. While some lesions may be readily apparent from birth, others may be perceived later on during the workup of nasal obstruction or epistaxis. Imaging studies are frequently key in the diagnosis and treatment planning of theses masses.

We provide a pictorial review of nasal masses in children and depict the key and occasionally distinctive imaging findings that can guide the neuroradiologist towards an accurate diagnosis.

In our series we describe pediatric nasal masses from several etiological categories. Causes of nasal masses in children can be grouped into non-neoplastic and neoplastic lesions. Non-neoplastic lesions include congenital and developmental disorders, such as congenital nasolacrimal duct mucocele, the unusual hairy polyp and the rare developmental midline masses, comprising dermoid/epidermoid cysts, cephalocele and nasal neuroglia heterotopia. Comparing to adults, inflammatory and infectious conditions such as mucocele, nasal or antrochoanal polyps and sinonasal polyposis are much rarer in children, and are frequently associated with cystic fibrosis or primary ciliary dyskinesia. Miscellaneous non-neoplastic disorders may comprehend traumatic lesions and iatrogenic causes, such as presence of surgical material. Foreign bodies must also be considered, as young children commonly insert these in their nasal cavities. Regarding the neoplastic group, benign vascular tumours such as infantile hemangioma and juvenile angiofibroma can show characteristic imaging findings. Rarer benign tumours like infantile myofibroma and myxoid chondroma can present as nasal masses. Different benign neoplasms and neoplasm-like lesions can arise from the osseous structures, including fibrous dysplasia and ossifying fibroma. Malignant lesions may also occur, such as sarcomatous tumours.

A wide variety of neoplastic and non-neoplastic conditions may present as nasal masses in children. While direct visualization might be sufficient in some cases, neuroimaging frequently has a crucial role in the management of these children. Knowledge of the most common causes and awareness for the key and distinctive imaging features can guide the neuroradiologist in the differential diagnosis. Further characterization of the lesion, determining its exact location, extension and, importantly, whether intracranial connection exists is critical for treatment planning and avoidance of a potentially devastating biopsy. 

To review the literature regarding the key MRI features of perineural tumor spread (PNS) in head and neck cancers, with emphasis on the relevant neuroanatomy; -to systematize an imaging diagnostic flowchart of the PNS, exemplified on the cases encountered in the institution database; -to increase awareness of the importance to diagnose timely the PNS in head and neck cancers.

Background PNS represents a distinct pattern of metastatic dissemination along large nerves which can be radiologically imaged, whereas perineural invasion (PNI) is the corresponding microscopic entity associated with small nerves infiltration. The trigeminal and facial nerves are the most commonly affected, although any of the cranial nerves (CNs) can be involved. Adenoid cystic carcinoma and squamous cell carcinoma are more frequently associated with PNS with a significant impact on the prognosis and on the applied treatment.

MRI is the modality of choice when evaluating PNS, thus helping in making a treatment decision without histopathologic confirmation of nerve involvement. Therefore proper use of suitable sequences is crucial. Being mostly clinically underdiagnosed and often radiologically overlooked, detection of PNS implies a well-structured diagnostic approach scheme ( fig.1 ) useful for the evaluation of cases from our department. The radiological approach is based on knowledge of the CNs anatomy, of the check-points for each neural route, nerves interconnections, and radiographic features, integrated into the clinical context.

Although the exact pathologic definition of PNS remains ambiguous, it obviously leads to high morbidity and mortality rates; hence it is of vital importance that radiologists search for this tumor extension in every imaging study of patients with known head and neck cancer including the asymptomatic cases, as well as in patients with any symptoms related to CNs pathology with an unexplained origin. A good comprehension of CNs innervated territory and anatomical course represents the "GPS" for detecting the PNS.

Neuroradiology S83 (2020) 62 :S1-107 (Suppl 1) S

Neuro-imaging studies are among the most frequently performed examinations in daily radiology practice, on which osseous skull lesions may be detected as incidental imaging findings or less commonly in symptomatic patients presenting with painful or palpable lesions.

Characterization starts with analysis of non-imaging parameters such as patient age, location, multifocality, symptoms and history. Langerhans' cell histiocytosis typically affects young patients, may be either solitary or multiple and may involve many organs. Metastasis and Paget's disease affect older patients and may be multifocal. Meningioma is more frequent in female patients.

Pseudo-lesions such as Pacchioni granulations and hyperostosis frontalis interna have a typical location and imaging pattern. Lesions can be subdivided into sclerotic or lytic lesions. MRI may reveal extra-osseous soft tissue extension and adds information by analysis of signal characteristics (e.g. high T2-signal in venous malformations). Usual sclerotic lesions include osteoma, fibrous dysplasia and meningioma-en-plaque each with their characteristic presentation. Usual lytic lesions may be caused by metastases, multiple myeloma or venous malformation. Less frequent sclerotic lesions include Paget's disease, sclerotic bone metastasis and sclerotic bone dysplasia. Less frequent lytic lesions include an eosinophilic granuloma and an epidermoid cyst.

Skull lesions comprise a large variety of entities. Although imaging features may overlap, correct characterization and relevant differential diagnosis depends on a combination of patient details, previous medical history and precise analysis of imaging features. This strategy may allow a confident lesion characterization or at least differentiation between benign don't touch lesions and potential malignant lesions. This may avoid unnecessary invasive diagnosis in many scenario's. Neuroradiology (2020) S84 62 :S1-107 (Suppl 1) S

Eosinophilic granuloma/Langerhans cell histiocytosis showing bevelled edges due to uneven destruction of the internal versus external table 

Sensorineural Hearing Loss (SNHL) is a major cause of childhood disability worldwide, affecting 6 in 1000 children. Cochlear implantation is a proven option for severe to profound sensorineural hearing loss. National screening programs and improved diagnostic, radiological and audiological facilities have increased the number of candidates for this procedure. Understanding the basic workings of the device along with the surgical anatomy emphasizes the importance of structured reporting. Through the use of our WAVES mnemonic to formulate a structured report, we aim to improve pre-operative diagnostics, prognostics and help surgical planning.

The absolute requirements for cochlear implantation are the presence of a cochlea (either normal or malformed) and of a cochlear nerve. The standard surgical approach involves posterior tympanotomy with access to the middle ear cavity by opening the facial recess allowing visualisation of the round window niche. An electrode array is then inserted into the scala tympani of the basal turn through the cochleostomy and used to stimulate the cochlear nerve. Thus, assessment of the mastoid process, middle ear, cochlea, and internal auditory canal can help determine suitability for and safety during cochlear implantation.

Result 1) Surgical Anatomy (Figure 1 ) 2) Cochlear Implant Device (Figure 2) 3) WAVES -Check List (Figure 3 ) 4) Mosaic of checklist pathology (Figure 4 )

Use of the WAVES mnemonic will add value to temporal bone CT and internal auditory meatus (IAM) MRI reports. In particular when reporting the pre-implant ear it will help determine candidacy and alert the surgeon to anatomical variants that could alter the operative approach. Neuroradiology S85 (2020) 62 :S1-107 (Suppl 1) S 

To evaluate variants of sphenoid sinus and to outline surgically risky anatomic variants vulnerable to injury. To identify the relationship between neurovascular structures towards sphenoidal pneumatization, as well as the occurrence of Sternberg's canal.

This study incorporated 104 computed tomography (CT) scan of paranasal sinus (PNS), done from January 2016 to December 2017 in single tertiary health centre. All softcopy images of paranasal sinuses were evaluated for pneumatization of anterior clinoid process (ACP), greater wing of sphenoid (GWS) and pterygoid plate (PP). The structures were examined, which include protrusion or dehiscence of internal carotid artery (ICA), optic nerve (ON), maxillary nerve (MN), and vidian nerve (VN); variants of sphenoidal septum and presence of Sternberg's canal.

The authors found that ACP pneumatization has significant relationship with protrusion of ICA and ON into sphenoid sinus. Presence of a single septum and multiple septa has a significant relationship with septal insertion to ICA canal. No prevalence of Sternberg's canal was established in this study.

CT scan imaging of the paranasal sinus is widely used as an important adjunct for radiological assessment of the type of sphenoid sinus pneumatization and also anatomical morphology of related neurovascular structures prior to transphenoidal surgery. This to highlight important variants, such as protrusions or dehiscence of key structures, presence of sphenoidal septum with risky insertion as well as the rare prevalence of Sternberg's canal.

Introduction Three-dimensional reconstructed imaging has become integral to radiological interpretation and surgical planning. Despite the inherent challenges of producing 3D imaging from non CT datasets clinical expectations remain high. Automated bone segmentation from MRI would have a profound impact clinically, particularly in the craniofacial skeleton where complex anatomy is coupled with radiosensitive organs. Techniques such as gradient-echo Black Bone (GRE-BB) and short echo time (UTE, ZTE) have shown potential in this quest. The objectives of this study were to ascertain: (1) whether the high-contrast of zero echo time (ZTE) could drive segmentation of high-resolution GRE-BB data to enhance 3D-output, and (2) if these techniques could be extrapolated to ZTE driven segmentation of a routinely used non bone-specific sequence (FIESTA-C).

Following ethical approval and informed consent, 11 adult volunteers underwent 3T MRI examination with sequential acquisition of ZTE, GRE-BB and FIESTA-C imaging. Craniofacial bone segmentation was performed using a fully automated segmentation algorithm. Segmentation was completed individually for GRE-BB and a modified version of the algorithm was subsequently implemented, wherein the bone mask yielded by ZTE segmentation was used to initialise segmentation of GRE-BB. The techniques were subsequently applied to FIESTA-C datasets. Since it was not possible to ethically obtain comparable CT imaging in healthy volunteers, the resulting 3D reconstructions were evaluated for areas of unexpected bony defects and discrepancies.

The acquisition times for the three sequences were 1:16, 4:48 & 3:28 minutes for ZTE, BB and FIESTA-C respectively. The automated segmentation algorithm yielded acceptable 3D outputs for all GRE-BB datasets. These were enhanced with the modified algorithm using ZTE as a driver, with improvements in areas of air/bone interface and dense muscular attachments. Comparable results were obtained with ZTE+FIESTA-C (Figure 1 ). Since only structures certain to be of no clinical interest were removed and the intensity information was preserved in the segmentation process, this permitted manipulation of the 3D rendering to achieve the best possible result.

Evaluating the diagnostic performance of conventional Magnetic Resonance (MR) sequences and diffusion-weighted imaging in differentiating malign and benign salivary gland lesions. 

Histopathologic assesment revealed 20 benign and 5 malign lesions. Benign group consisted of seven pleomorphic adenomas (two of them showed seeding after surgery), seven whartin tumors, two infantile parotid hemangioma, one hemorrhagic cyst, and three inflammatory conditions. Malignant lesions included two mucoepidermoid carcinomas, one adenoid cystic carcinoma, and two lymphomas. We found no significant difference in T1, T2 signals and signal intensity after contrast injection. Also mean ADC values did not show any significant difference among groups (p>0,005). Morphologic parameters; ill defined margin, deep lobe affection, perineural spread, lymphadenopathy and extension to adjacent structures were found to be significantly associated with malignant lesions (p<0,005).

Signal intensity, heterogeneous contrast enhancement, irregular borders, ADC values and time signal intensity curve in dynamic series were reported to be helpful in differential diagnosis in previous studies. Although deep lobe involvement has been significant in previous studies, it is not specific for malignancy because some benign conditions may also be localized in the deep lobe. Fine needle aspiration biopsy in the diagnosis is often unable to give definitive results. Due to difficulties in daily practice dynamic MRI is not always applicable for all patients in head and neck imaging. Also, some of these lesions are detected incidentally. Thus, conventional MRI findings become more important. The margin feature of lesions and the spread to adjacent tissues were the most valuable parameters in this distinction. The lack of dynamic series, the small sample of malignant lesions were the limitations of the study. Our study concluded that; the morphological features of the lesions seem more reliable than quantitative parameters such as signal intensity and ADC measurements in differentiation of salivary gland lesions. Introduction Thyroid-associated ophthalmopathy (TAO) is the most common orbital disease. At computerized tomography (CT) or magnetic resonance imaging (MRI), a muscle increase involving the superior, medial, and inferior rectus is quite typical of TAO. However, a number of conditions affecting orbital tissue can present with the commitment of the extraocular muscles (EOM). The most common inflammatory condition affecting EOM and that mimics TAO is idiopathic orbital inflammation (IOI). Nonetheless, other orbital diseases should be considered in the differential diagnosis. The purpose of this study was to identify the main conditions presented with enlargement and signal abnormality at MRI of EOM, highlighting how to distinguish each one.

A pre-existing database of all patients who had undergone orbit MRI at a public hospital unit from May 2019 to June 2020 was selected. The patients who presented enlargement and signal abnormality of the EOM or both were chosen, according to image analysis. Clinical and histopathological information of these patients were investigated.

EOM involvement was identified in 43 patients. The most frequent condition was TAO, observed in 19 patients (44%). Nine patients (23%) were considered affected with IOI by the ophthalmology team.

Two patients (5%) have a history of radiotherapy, one to treat an orbital rhabdomyosarcoma, and the other an orbital lymphoma. Metastasis of breast tumor in the orbit was seen in two patients (5%). One case of the following conditions, each one representing 2% of the total, was observed: carotid-cavernous fistula, trauma, amyloidosis, T-cell acute lymphoblastic leukemia, lymphoma, and tuberculosis. Five of the patients with EOM involvement remain under investigation.

Discussion IOI, also known as orbital inflammatory pseudotumor, represents an inflammatory condition and a cause of EOM involvement. It is the third most common orbital disease after TAO and lymphoproliferative diseases. Metastasis, lymphoma, and low-flow carotid-cavernous fistula may mimic IOI. Immunoglobulin G4 (IgG4)-related disease is a systemic disease that commonly involves the head and neck. Orbital tumor-like conditions such as IOI and lymphoid hyperplasia are now considered as a part of the IgG4-related disease spectrum. Other diseases, some of them rares, such as amyloidosis, should also be considered in the differential diagnosis of EOM involvement.

EOM enlargement is observed in a heterogeneous group of conditions. Imaging plays an important role in elucidating any underlying etiology behind EOM involvement.

Neuroradiology (2020) S88 62 :S1-107 (Suppl 1) S Introduction Septo-optic dysplasia plus is a rare congenital syndrome characterized by the classic triad of optic nerve hypoplasia, hypothalamic-hypophyseal dysfunction, and midline abnormalities, with associated different cortical malformations. Clinical manifestations include optic nerve disease, endocrine dysfunction and global developmental delay. All published studies in other countries have shown a prevalence of 1.9-5.4 per 100,000 live births with no gender predilection.

We present the case of a infant admitted to our hospital for further diagnostic evaluation due to persisting associated ophthalmic (nystagmus, bilateral severe visual impairment with optic nerve hypoplasia), neurologic (hypotonia, delayed milestones) and endocrine (diabetes insipidus) manifestations. Family history was unremarkable apart from older maternal age. Child was underwent to MRI examination within all coordinated planned diagnostic procedures and clinical examinations including genetic testing. The initial (performed at the age of 5 months) and follow up MRI examination (performed at the age of 13 months) were reported with following findings: bilateral hypoplasia of optic nerves including optic chiasm; anterior pituitary hypoplasia including absent pituitary stalk; ectopic posterior pituitary along median eminence; absence of rostrum with hypoplastic splenium of corpus callosum; dysplastic but existing septum pellucidum; thick and dysplastic both fornices; thin bridge as fusion of brain tissue at septo-preoptic area; presence of thick dysplastic cortex vs polymicrogyria in frontoparietal and temporooccipital regions; subcortical heterotopia along left parietooccipital sulcus. Due to the presence of the multiple abnormalities of optic nerves, pituitary and other midline abnormalities, including septo-preoptic holoprosencephaly along with associated cortical malformations, a diagnosis of septo-optic dysplasia plus was made. Expected mutations as HESX1, SOX2, OTX2, which can explain clinical and imaging findings in this patient, were not identified in performed genetic testing, therefore, could be continued for other gene panels and discovering possible mutations.

The term SOD-plus was suggested by Miller in 2000. to differentiate those patients with global development delay and motor deficits which could not be explained by classical triad in findings. The term "SOD-complex" was also proposed due to very heterogeneous phenotype and genotype expression. MRI is required to make the diagnosis in patients with ophthalmologic, endocrine and neurologic clinical manifestations.

Pedro Henrique Pestana Barradas 1 , Manuel Joao Delgado de Brito 2 , Silvia Carvalho 3

Characterized by ocular, consciousness and gait alterations, Wernicke Encephalopathy(WE) is an acute neurological pathology due to Vitamin B1(thiamine) deficiency. Often associated with alcohol consumption, thiamine deficiency leading to WE can also occur in nonalcoholic(NA) patients and result from prolonged vomiting, systemic infections and chemotherapy. Given WE's potential morbidity and mortality, early diagnosis and treatment are imperative. Pediatric oncology cases of WE are rare. Magnetic Resonance Imaging(MRI) in WE can show symmetric signal change alterations typically involving the medial thalami, mammillary bodies, tectal plate and periaqueductal area. There are also atypical MRI findings, found more frequently in NA patients, with involvement of the cerebellum, cranial nerve nuclei, red nuclei, caudate nuclei, splenium and cerebral cortex.

We report a case of WE in a pediatric patient with Burkitt lymphoma, whose MRI showed lesions in typical locations, as well as in atypical ones.

A 16 year-old boy diagnosed with Burkitt lymphoma, and reporting recent weight loss, was admitted to undergo a cycle of chemotherapy with Neuroradiology S89 (2020) 62 :S1-107 (Suppl 1) S Rituximab, high-dose Cytarabine and Etoposide(R-CYVE). At the 3rd day of treatment he suffered nausea, anorexia and apathy. After the 4th and final day of treatment he showed no signs of improvement and now had pancytopenia requiring transfusion support. Over the next 10 days his clinical status deteriorated, as he exhibited fluctuations in consciousness, fever and altered speech, eventually needing intubation. The clinical picture could suggest Reversible Posterior Encephalopathy Syndrome, sometimes associated with cytarabine. Lumbar puncture and CT were conducted but showed no significant changes. Subsequent brain MRI showed multiple T2/FLAIR hyperintense lesions involving the medial thalami, periphery of the third ventricle, tectal plate and the periaqueductal gray matter, suggesting the diagnosis of WE. Additionally, abnormal signal intensity was seen in the medulla, vermis and frontal lobe cortex. No enhancement was shown after gadolinium. Thiamine was started for 10 days, with radiological improvement on follow-up MRI on the 6th day of treatment, showing significant regression of all lesions. Clinical improvement was seen after the 10th and final day of thiamine, with the patient being discharged 5 days later, still showing periodic confusion and marked weakness, especially in the lower limbs.

Neuroimaging findings are important to suggest the diagnosis of WE, which can be missed in NA patients, namely in cancer patients. This case highlights also the occurrence of atypical locations for the disease, like the cerebellum and cerebral cortex, in the group of NA patients. 

Our objective is to show a case of Macrocephalic Vanishing White Matter Disease. We will also make a literature review and show some typical cases from the file of our tertiary hospital, comparing the imaging findings with our unusual case, highlighting the possible causes for that appearance.

Vanishing white matter disease (VWM), also known as myelinopathy centralis diffusa, is an inherited leukoencephalopathy that typically manifests as cerebellar ataxia following minor stress conditions, such as mild head trauma, infection or extreme fright. While childhood onset is the most common presentation, it may affect people of all ages, including adolescents and adults. It is characterized by severe neurological deterioration following episodes of acute symptoms that include cerebellar ataxia, spasticity, epilepsy, optic atrophy migraine and psychiatric symptoms, triggered by minor events, such as mild head trauma or fever. These events may be recurrent and worsen previous neurological disabilities, eventually leading to rapid decline and eventually lethargy or coma. VWM shows autosomal recessive inheritance and is related to mutations in any of the five EIF2B genes that encode subunits of the eukaryotic translational initiation factor 2B (EIF2B). This multisubunit protein plays a pivotal role in initiation and control of mRNA translation and therefore is critical to overall protein production regulation in the cell. Imaging abnormalities are always present and may precede clinical symptoms. Typical MRI findings are diffuse and symmetrical white matter hyperintensity on T2-weighted and FLAIR images with involvement of the arcuate "U" fibers, which may affect both supratentorial, cerebellar and brainstem regions. Contrast enhancement and restricted diffusion are not common. Radiating stripes of remaining normal white matter may be seen permeating the rarefied white matter. Cerebellar atrophy may be present in late stages. Disease progression manifests as white matter rarefaction and cavitation, ultimately leading to replacement of white matter by CSF-like signal intensity in all MRI sequences. In our case, a girl with consanguineous parents had a normal psychomotor development initially and with 18 months she started a psychomotor regression, with motor and cognitive impairments, also accompanied by seizures (febrile). Progressively she developed macrocrania, which we believe may be caused by some cerebrospinal fluid absorption disturbance, that we will discuss with details in this work.

It is important for the neuroradiologist to know the typical imaging findings of VWM disease, also being able to think of it in the differential diagnosis of cases that aren't that typical. Neuroradiology (2020) S90 62 :S1-107 (Suppl 1) S

Homocystinuria is a rare congenital disorder of methionine metabolism characterized by elevated plasma concentration homocysteine and urine concentration of its oxidation product, homocysteine, most commonly secondary to deficiency of cystathionine beta-synthase (CBS). Although not specific, the most common imaging findings include venous thrombosis and multiple cortical and subcortical cerebral infarctions. Diffuse bilateral white matter changes are rarely reported in the literature. 

We hypothesize venous thrombosis in the context of homocystinuria to have triggered the transient diffuse white matter changes through a chronic congestive mechanism due to deficient venous drainage, seldom reported in this disease. However, in addition, this white matter appearance could be related to abnormal accumulation of homocysteine and its metabolites. This unusual imaging presentation of homocystinuria illustrates the importance of considering this entity and other amino-acidopathies in the differential diagnosis of white matter diseases in the pediatric population, as well as associated to venous thrombotic phenomena.

Michael Paddock 1,2 , Andrew Martin 3 , Christopher S Johns 4 , Jessica Smith 5 , Ashok Raghavan 6 , Daniel Ja Connolly 3,6 , Amaka Offiah 2,6

To assess whether head CT with 3D reconstruction can replace skull radiographs (SXR) in the imaging investigation of suspected physical abuse (SPA)/ abusive head trauma (AHT).

PACS was interrogated for antemortem skeletal surveys performed for SPA, patients younger than 2 years, SXR and CT performed within 4 days of each other. Paired SXR and CT were independently reviewed. One reviewer analysed CT without and (3 months later) with 3D reconstructions. SXR and CT expert consensus review formed the gold standard. Observer reliability was calculated.

A total of 104 SXR/CT examination pairs were identified, mean age 6.75 months (range 4 days to 2 years); 21 (20%) had skull fractures; two fractures on CT were missed on SXR. There were no fractures on SXR that were not seen on CT. For SXR and CT, respectively: PPV reviewer 1, 95% confidence interval (CI) 48-82% and 85-100%; reviewer 2, 67-98% and 82-100%; and NPV reviewer 1, 95%, CI 88-98% and 96-100%; reviewer 2, 88-97% and 88-98%. Inter-and intraobserver reliability were respectively the following: SXR, excellent (kappa = 0.831) and good (kappa = 0.694); CT, excellent (kappa = 0.831) and perfect (kappa = 1). All results were statistically significant (p < 0.001).

CT has greater diagnostic accuracy than SXR in detecting skull fractures which is increased on concurrent review of 3D reconstructions and should be performed in every case of SPA/AHT. SXR does not add further diagnostic information and can be omitted from the skeletal survey when CT with 3D reconstruction is going to be, or has been, performed.

Head CT with 3D reconstruction is more sensitive and specific for the diagnosis of skull fractures. Skull radiographs can be safely omitted from the initial skeletal survey performed for suspected physical abuse when head CT with 3D reconstruction is going to be, or has been, performed.

Vigabatrin (VGB) is an antiepileptic drug recommended as an effective firstline therapy for infantile spasms in tuberous sclerosis (TSC) and the secondline therapy for other etiologies. Infantile spasms are a very specific type of seizures, typically begining between 3 and 8 months of age. We present symmetrical involvement of basal ganglia, thalami, brainstem and dentate nuclei in children treated with VGB. Magnetic resonance imaging (MRI) exhibited abnormal T2-hyperintensities and/or restricted diffusion on diffusion weighted imaging sequence (DWI).

Neuroradiology S91 (2020) 62 :S1-107 (Suppl 1) S

We retrospectively analyzed MR examinations of 77 children with infantile spasms. Abnormalities of signal intensity were found in 19 patients (9 TSC patients, 2 pts with Down syndrome, 1 with Sturge-Weber syndrome and 7 with infantile spasm other etiology). The study group consists of 9 male and 10 female, aged from 3 to 29 months. All these patients were treated with vigabatrin. During therapy follow-up examinations were performed in 14 patients. MR examinations were performed on a 1.5T scanner. The standard brain protocol contained DWI sequence with b-values = 0,800 or 1000. The ADC maps were automatically derived.

Result 18 patients (18 of 19; 95%) demonstrated MR signal abnormalities within the globi pallidi, 13 (68%) in the brainstem, 12 (63%) in the thalami, and 7 (37%) in the dentate nuclei. Four patients out of 19 (21%) presented with abnormalities in all of the aforementioned anatomical regions. In eight cases (42%) the abnormal signal was clearly visible in the DWI sequence but it was not noticeable in the T2-W images. Complete or partial but significant resolution of changes was seen in all follow-up studies.

MR signal abnormalities in children treated with vigabatrin are reversible, benign findings. DWI has proved to be a useful and objective parameter in assessing those changes that are not clearly visible on T2WI. Asymptomatic T2-hyperintensity and restricted diffusion remitted even without discontinuation of VGB.

To determine whether there is a need to perform skull radiographs (SXR) in the assessment of suspected inflicted injury where CT head with 3D reconstructions are planned or have been performed. To help direct future guidelines.

PACS was interrogated between 01/01/2017 and 01/01/2020 for antemortem skeletal surveys performed for suspected inflicted injury. The age cut off used at time of examination was 2 years old. The skull radiograph reports where reviewed for the presence of fracture. Following this, the corresponding CT head reports were reviewed for fracture and any additional findings.

Result 39 skeletal surveys for suspected inflicted injury were identified. 30 distinct SXR/CT head examination pairs were identified. Of this subgroup, the mean age at time of examination was 6 months (range 1 month to 20 months). 8 patients (26%) had documented skull fractures. One fracture identified on 3D CT head was not seen on SXR. None of the fractures identified on SXR were overlooked on CT. In our centre the CT Head preceded or occurred within 30mins of the skeletal survey in 90 % of cases.

3D CT head has a greater diagnostic yield than skull radiographs in identifying skull fractures. In children being assessed for suspected inflicted injury, who have had or will have a CT head performed, SXR can be safely excluded from the imaging protocol. L-2 hydroxyglutaric aciduria (L2HGA) is a rare autosomal recessive inherited disease, which presents with developmental delay, and macrocephaly. L2-HGA is characterized by an elevated level of 2 hydroxyglutaric acid (2-HGA), especially in the urine. The disease has characteristic findings on Magnetic Resonance Imaging (MRI) such as bilateral and symmetrical centripetal involvement in subcortical U-fibers and basal ganglia. Even in the late stages, the deep white matter of the periventricular region, corpus callosum, internal capsule, brainstem, and cerebellar white matter is spared. Clinical features begin with learning difficulty and developmental delay. Cerebellar findings become evident in later years. Patients with L2-HGA have a predisposition to cerebral neoplasms. This matter may be because of high amounts of 2-HG, which have been found in brain neoplasms that have IDH1 mutations. The purpose of this presentation is to describe the typical MRI findings of disease and to draw attention to neoplasm development and to emphasize the importance of follow-up.

For five years (between January 2014 and January 2019), seven patients diagnosed as L2-HGA by MRI were included in the study. All the patients were referred to MRI with suspicion of metabolic disease and were diagnosed with laboratory tests after imaging.

The records of seven patients were obtained as a result of the screening. Four of these patients were male. The parents of six patients were 1st-degree cousins. The mean age was 11.57±5.62. Four patients had seizures, one had ataxia, and one had a headache. The mean followup period of the patients was 3.43±1.62 years, and during this period, the patients had an average of 3.14±2.34 MRI. At the time of diagnosis, subcortical U-fibers, globus pallidus, external capsule, and dentate nucleus were involved in all patients. Other areas of involvement were shared at table 1. Subcortical U-fibers involvement increases with age and shows an anterior-posterior gradient. The ventricular dilatation increased linearly with atrophy. One patient developed anaplastic oligodendroglioma during followup and therefore underwent surgery.

L2-HGA is a rare neurometabolic disorder, and it is mainly diagnosed with massive 2-HGA in the urine. Characteristic MRI findings should strongly Neuroradiology (2020) S92 62 :S1-107 (Suppl 1) S suggest the diagnosis. Since malignant brain tumors may develop during the disease, follow-up is essential. According to the literature, it is an extremely rare disease, with an estimated incidence of 1 in 50.000-100.000 births, but with no sex predominance. In Portugal, with a total population of approximately 10 million inhabitants, there is record of under 20 cases diagnosed. Clinical features include: epilepsy (70-75%), distinctive facial characteristics, moderate to severe cognitive impairment, and multiple congenital malformations. Notwithstanding almost all patients have abnormal imagiological findings, there is no specific and pathognomonic pattern to MWS. Globally, changes may include abnormalities of the hippocampus, ventricular system, white matter as well as basal ganglia; agenesis of the corpus callosum, cortical and/or cerebellar development impairment.

The authors report two clinical cases of children diagnosed with MWS before the first year of life, which are currently aged six and eight (patients 1 and 2). The main objective is not only to express their clinical presentation and evolution, but also to analyze and compare the correlation between imaging and clinical findings.

The first patient, presented at 3 months of age with epileptic seizures, spasticity and hypertonia of the lower limbs associated with a facial dysmorphic syndrome. Neuroimaging investigation revealed microcephaly with dysmorphic and enlargement of the basal ganglia and thalamus. Today, with 6 years old, he has axial hypotonia, mild behavior disorder, but controlled epileptic seizures with a favorable outcome and no need for hospitalization. Patient two was diagnosed at 9 months of age, as he showed distinctive facial characteristics and global development delay. Imagiological findings included complete corpus callosum agenesis, colpocephaly, incomplete folding of the hippocampus, and changes in the normal morphology of the remaining cerebral commissures. At this moment, with 8 years old, he presents language impairment and epilepsy, that required hospital admission for clinical stabilization. Moreover, congenital abnormalities of the heart, kidneys and genitals were identified.

MWS is a very rare disease and its diagnosis is based on clinic, imagiological and genetic findings. With the increased knowledge of this condition and improvement of the diagnostic tools (namely genetic assessment), a higher number of diagnosed cases will be expected. Due to the multiple forms of presentation and evolution, the neuroradiologist has to have a high index of suspicion, despite non-specific imagiological findings, in order to help establishing a probable diagnosis.

Neuroradiology S93 (2020) 62 :S1-107 (Suppl 1) S Introduction Leukemia is the most common childhood malignancy and Lymphomas (Hodgkin and non-Hodgkin, HL and NHL) are the third most common malignant neoplasm in pediatric patients, both diseases have a wide spectrum of manifestations, but CNS is one of the most common places that can be affected by the primary disease, relapse and treatment complications. Methods We review our data base studies of pediatric CNS imaging, focussed on CT/ MRI (brain, head&neck, spine) during the last 10 years in patients with CNS related clinical symptoms (at diagnosis, during treatment or suspected relapse), and confirmed Lymphoma or Leukemia histopathology.

CNS findings in Leukemia patients have a wide range of appearances: hemorrhage or infarction, abnormal enhancement of the meninges or nerve roots, retinal hemorrhage. Extramedullary leukemia (granulocitic sarcoma or chloroma) can affect orbits, CNS and spine and is most common in AML. CNS is a frequent site of relapse in leukemia patients, being more common in ALL than in AML, and extramedullary involvement is more common at relapse than presentation. This incidence has being reduced in the past years by CNS prophylaxis treatment, but this can increased the possibility of treatment complications (dural venous thrombosis, drug induced encephalopathy…). HL is more frequent in older children and adolescents, clinical hallmarks are cervical or supraclavicular lymphadenopathy and B symptoms, meanwhile extranodal manifestations with CNS primary involvement are more common in NHL. Between the types of NHL that commonly occur in children (Burkitt lymphoma (BL), lymphoblastic lymphoma (LL), diffuse large B-cell lymphoma (DLBCL) and anaplastic large cell lymphoma(ALCL)); BL is the most frequent subtype with extranodal sites involved, including SNC, specially in patients with bulky disease, in LL CNS can also be seen, but is rare in DLBCL and ALCL. In HL an NHL, relapse is more common in the primary site of disease but can be seen in any localization. Treatment complications also can involve CNS as in leukemia patients.

Haemopoietic and Lymphoproliferative disorders are the first and third most common malignancies in paediatric population, we must know the more frequent manifestations of these diseases and keep in mind that CNS is one of the target organs that can be affected in all the stages of the disease (diagnosis, relapse or treatment complications), and some histopathological subtypes are specially related with this specific manifestations. High index of suspicions of branchial cleft fistula is necessary which was often not been suspected at first presentation of neck abscess or even recurrent neck abscess in childhood. This study is aimed to correlate the imaging studies and surgical findings in patients with first time or recurrent neck abscesses.

We retrospectively reviewed paediatric patients who had cross sectional imaging done either CT and/or MRI in our institution from 2010-2019 with diagnosis of branchial fistula. Clinical presentation, laryngoscope findings and imaging features were evaluated. We also obtained the intra-operative findings and histopathological results for those who underwent surgery.

Ten (n=10) paediatric patients, six male and five female with age ranged from 11 months to 14 years old were included in this study. First, second and third/fourth type of branchial fistula was reported in 2, 2 and 6 respectively. Five cases reported as fistula on imaging with other showed sinus or collection only. Eight patients underwent examination under anaesthesia with laryngoscope and operation with 2 patients with first branchial anomaly were treated conservatively. Intra-operative direct laryngoscope with or without methylene dye administration demonstrated patent fistula in 4 out of 5 patients (80%) with one of the case thought to be fistula was actually a sinus as the methylene dye was negative. Six patients have histopathological correlation of fistula as 2 patients only had the fistula cauterized and no tissue sample was sent for evaluation. The imaging and intraoperative correlation will be presented in pictorial review format.

Cross section imaging such as CT and MRI are good in demonstrating associated fistula which is concordant with the intraoperative and histopathological findings.

Thomas Rogers 1 , Richa Sinha 1 , Samuel Gregson 1 , John Cain 1 Neuroradiology (2020) S94 Introduction Hypoxic ischaemic encephalopathy (HIE) is an important cause of neonatal mortality and morbidity. Prompt initiation of therapeutic cooling improves outcome. Early and accurate assessment of brain injury is important for prognostication and guiding care decisions. MRI is an excellent predictor of outcome. Furthermore, DWI is the most reliable MRI sequence to detect mildmoderate HIE injury during the first week after a hypoxic-ischaemic event. After approximately 7 days DWI imaging findings become falsely negative due to pseudonormalisation. Therefore the recommended optimal time frame, for MRI examination in neonates ranges from 3-7 days. We performed a retrospective analysis of MRI brains performed at our institution in term neonates with suspected HIE to determine the proportion performed at less than 8 days and if timing of scan had an impact on incidence of abnormal findings.

Interrogation of trust PACS system to identify all MRI examinations, for suspected HIE in neonates. 45 cases dating back 8 years were identified. All MRI scans were reported by neuroradiologists. Data on neonates age at time of scan and the scan findings was collected.

Infant age at MRI ranged from 4 to 24 days, mean 9.7 days. 73% of MRI for HIE occurred in infants aged 8 days or more. In patients who had MRI at 8 days or more, 12% showed diffusion restriction changes consistent with HIE versus 50% of patients who were scanned at less than 8 days. The chi-square statistic is 7.3052, p-value 0.006 (<0.05).

Significantly fewer neonates scanned after 8 days of life demonstrated HIE associated diffusion abnormalities. This raises the possibility that mild HIE cases could be missed in some of these later examinations due to pseudonormalisation of DWI. Some examinations were requested later than expected or did not specify a particular time the examination was required. Other requests specified they wanted the examination at or around 10 days. This suggests clinicians may not be aware of temporal variations in MRI features of HIE or they may not fully understand these and the optimal timeframe for scanning.

This study found that 73% of MRI brains performed for suspected HIE in term neonates were performed after 8 days when diffusion abnormalities can be masked. We identified a lack of standardised requesting policy causing a delay in scanning neonates and suggest requesting imaging at time of therapeutic cooling, earlier in the management pathway, of suspected HIE to enable prompt scanning. Cortical laminar necrosis is associated with hypoxia, metabolic disturbances, drugs and infections. It appears as cortical gyriform hyperintensities with a laminar pattern on T1 weighted images in magnetic resonance imaging (MRI) without signs of hemorrhage or calcification on T2 weighted images or computed tomography (CT). A 1 year-old boy, born term with uneventful antenatal and postnatal history, presented with head enlargement. The initial computed tomography (CT) and magnetic resonance imaging (MRI) brain showed obstructive hydrocephalus secondary to aqueductal stenosis. Subsequent CT and MRI brain were repeated 2 years later as the child presented with delayed developmental milestones and convulsions. It revealed cortical and subcortical gyriform hyperdensities / hyperintensities at bifrontal and left occipitoparietal region, with vein of Galen aneurysmal dilatation; the latter complicated with early subacute left basal ganglia bleed with intraventricular extension and hydrocephalus. In MRI brain, T1 weighted image shows corresponding subtle T1W hypointense signal within the grey matter of bifrontal and left occipitoparietal region, associated with T2W hyperintense signal and swelling of these affected areas. No blooming artefact demonstrable in gradient recalled echo (GRE). These findings are suggestive of cortical Neuroradiology S95 (2020) 62 :S1-107 (Suppl 1) S

Cerebral Cavernous Malformation (CCM) affects patients of any age with 35% of cases found in children. CCMs can be occult on CT and even on MRI. In this pictorial educational review of cases from a tertiary paediatric neurosciences centre we explore the challenges of diagnosing CCMs in the paediatric population, key imaging features and subsequent management options.

The presentation and clinical course of CCM is variable, up to 50% of cases are asymptomatic. Mean age in children at the time of diagnosis is 10 years. Children can present with focal seizures, intracranial haemorrhage (ICH) and focal neurological deficits (FNDs) without radiological evidence of recent haemorrhage. Vascular malformations, including CCMs, should be excluded in children who present with spontaneous ICH. The annual risk of haemorrhage in children with CCM is 1.6% per patient year, climbing to 8% in symptomatic patients. ICH from CCMs are typically small and result in mild functional impairment. Risk factors for symptomatic haemorrhage include: brainstem location (accounting for 11-15% of CCMs in children); haemorrhagic clustering and an associated DVA.

CT findings in CCM are subtle and nonspecific. MRI is the investigation of choice, key imaging features being; the dark haemosiderin ring of the classic 'popcorn' lesion seen on T2-weighted imaging, intrinsic high T1 signal from methaemoglobin and the susceptibility of iron within the haemosiderin which can be exploited to increase lesion conspicuity. Susceptibility weighted imaging is therefore the most sensitive sequence in the detection of CCM due haemosiderin's 'blooming' artefact and should be performed in children undergoing MRI following acute ICH identified on CT imaging and those with seizure and/or FNDs.

An individualised approach to patient management is advocated. Management options including microsurgery and stereotactic radiosurgery (specifically gamma knife radiation) should be stratified dependent on; lesion location, symptom burden and history of haemorrhage. Frank discussions between patients and caregivers are encouraged to explore the risks of treatment, including the implications of genetic testing to the patient and their family unit.

CCM in children is an uncommon but important diagnosis, which can be overlooked with standard imaging techniques. Understanding common presentations ensures patients have access to appropriate imaging sequences and management options. Fig. 1 Selected neuroimaging in a 15-year-old male who was presented to the Emergency Department (ED) with acute onset headache, confusion (GCS 13) and bilateral leg weakness, left greater than right. An initial head CT performed out-of-hours for suspected acute intracranial hemorrhage was reported as normal. MRI was performed later the same day for persistent unexplained symptoms. (a) Review of the axial DWI (b0), which can be thought of as a 'free' T2* sequence, demonstrated a low signal intensity focus in the right frontal lobe. (b) Axial SWI mIP demonstrated a low signal intensity 'popcorn' shaped lesion with prominent 'blooming' at the same location in the right frontal lobe as (a) with the lesion appearing larger and more obvious. Appearances were consistent with a CCM. When susceptibility or 'blooming' artefact is detected on DWI, it should prompt formal SWI to be obtained. In our center, SWI is included as a routine pediatric brain sequence. Review of the patient's previous imaging revealed that a head CT had been performed at 6 years of age. He was presented to the ED at that time after being head-butted in school and subsequently became 'sleepy' with one episode of vomiting. (c) Axial head CT demonstrated a focus of hyperattenuation (arrow) within the right frontal lobe which, whilst described at the time, was not felt to represent acute ICH. (d) Upon review of the more recent head CT, a subtle focus of hyperattenuation was identified in the same location (arrow): both foci in (c) and (d) lie at the same location as the CCM demonstrated on MRI in (a) and (b). This case demonstrates that the CT findings of CCM can be subtle and may be overlooked: a subtle area of hyperattenuation on CT should not always be dismissed, particularly if there is previous cranial imaging with which to compare. The possibility of an underlying CCM should be ever present when reporting head CT imaging in children, particularly given the potential for ICH and significant FND. As demonstrated in (a) and (b), MRI facilitates the diagnosis.

Selected neuroimaging in a 16-year-old female who was referred for MRI for a bilateral inferior nasal defect and a suspected 'intracranial lesion'. There were no ocular findings on examination. Axial T2-WI (a) demonstrated a small central high signal lesion in the left frontal lobe surrounded by a low signal ring (arrow) with corresponding sagittal high T1-WI signal (b) (arrow).

(c) b0 DWI showed a corresponding focus of signal loss i.e. 'blooming' artefact in the same location as (a) and (b) (arrow). Review of the b0 DWI revealed a further area of low signal in the periventricular white matter adjacent to the occipital horn of the left lateral ventricle in (d) (arrow) but with no corresponding lesion or signal abnormality on the contemporaneous axial T2-WI in that location (e). The patient was recalled for SWI. The axial mIP sequence demonstrated several areas of 'blooming' artefact in both compartments of the brain (arrows) (f-h). Corresponding axial T2-WI through the infratentorium (i) revealed none of the low signal lesions identified in (h). Head CT (j) obtained at the time of the SWI mIP revealed a subtle focus of hyperattenuation in the left frontal lobe (arrow) at the same location as in (a) and (b). The imaging features in (a-j) are in keeping with multiple CCM. This case further reinforces the value of SWI in cases of multiple CCM which may not be visualized on other MRI sequences. Features were not consistent with an acute ICH because of the CT density (measured in Hounsfield Units, HU) and the absence of perilesional edema. Axial (b) T2-WI and (c) T1-WI demonstrated the characteristic lobular 'popcorn' appearance in the left frontal lobe with heterogenous signal intensity reflecting the varying stages of hemoglobin breakdown. There is central higher signal (due to methemoglobin) surrounded by a dark peripheral rim (low signal intensity due to hemosiderin deposition). (d) SWI mIP at the same location demonstrated prominent 'blooming' artefact in the left frontal lobe. The imaging features are typical of a CCM. Lesions which abut or involve the cerebral cortex can be a potential seizure focus (epileptogenic). Microsurgical resection was undertaken for repeated, medically uncontrolled seizures and follow-up MRI did not demonstrate a residual lesion. Postoperatively, his seizures had resolved, and he remained on antiepileptic medication. He was subsequently discharged from the neurosurgical services. Axial T2*-WI better demonstrated the surrounding peripheral low signal hemosiderin rim and also identified a further small dot-like CCM in the right frontal lobe anterior to the genu of the right corpus callosum. There was no corresponding lesion or signal abnormality identified on the contemporaneous axial T2-WI in (a). This further reinforces the value of SWI in the diagnosis of CCMs, particularly of smaller/tiny dot-like lesions which may be associated with giant CCMs and should be actively sought. The (c) coronal T1-WI volume acquisition further demonstrates the internal heterogeneity of the giant CCM and confirms its part-cystic nature in the lateral aspect of the lesion. (d) TOF MRA of the Circle of Willis (coronal reconstruction) demonstrates the absence of arterial feeders to the lesion which demonstrates T1-WI shine through. DTI was performed given the highly eloquent area in which the giant CCM is located: oblique coronal (e) and angled axial (f) views demonstrate the right corticospinal tract and its relationship with the giant CCM. Dynamic susceptibility contrast (DSC) perfusion imaging was also performed (g) and demonstrates central hypoperfusion within the lesion (bottom two axial images, through the level of the lesion). There was a family history of FCCM, and the patient tested positive for the KRIT1 mutation (CCM1, autosomal dominant inheritance). She proceeded to surgical resection for the giant right basal ganglia CCM which significantly improved the hemidystonia. Two years later, she underwent SRS for the smaller right frontal lobe lesion seen in (b). Follow-up MRI over the subsequent 6 years remained stable. T2-WI pre-SRS treatment planning imaging obtained in a 14-year-old male with a history of learning difficulties and autistic spectrum disorder.

The typical appearances of a CCM in the left thalamic region with central high T1 signal due to methemoglobin and a dark peripheral hemosiderin rim are better demonstrated in (b) than in (a). Note the circumferential fiducial markers. Follow-up MRI was obtained 5 years after (a) and (b) for a deterioration in clinical state with 'difficulty in holding up his head'. Selected (c) axial T2-WI, (d) coronal T2-WI and (e) sagittal T1 imaging revealed that the left thalamic CCM had reduced in size when compared with the pre-SRS imaging in (a) and (b). There was no evidence of ICH or further 'blooming' artefact on the axial GRE sequence (f). A further follow-up MRI was obtained 3 years after (c-f) and 8 years after SRS treatment (a-b) at 22 years of age for a recent deterioration in his behaviour. Selected (g) axial T2-WI, (h) coronal T2-WI, (i) sagittal T1 and (j) axial SWI imaging at the same levels as (c-f) demonstrated stable appearances with no change since the previous imaging. Note the presence of the endotracheal tube in (e) and (i).

Pre-and post-operative imaging in children with left-sided CCMs. Microsurgical resection is generally preferred if the lesion is superficially located (as in these cases), if a large ICH has occurred, or if the risk of damage to surrounding tissue from irradiation is too high. Selected imaging in a 14-year-old male who was presented to the ED with seizures. (a) Axial T2-WI, (b) post-contrast T1-WI volume acquisition sagittal reconstruction and (c) sagittal FLAIR demonstrated a left frontal parafalcine lesion toward the vertex with central heterogeneity and a low signal intensity hemosiderin rim consistent with a CCM. The patient experienced continued uncontrolled daily seizures. Post-operative follow-up imaging in (d-f) performed 10 months after (a-c), with the same sequences at the same level, demonstrated a non-enhancing surgical bed with simple central fluid content and post-surgical low signal intensity hemosiderin deposition. Note the postsurgical metallic artefacts in the frontal and parietal skull in (d-f). Post-operatively, his seizures had resolved, and he remained on antiepileptic medication. He was subsequently discharged from the neurosurgical services.

Selected axial MRI in a different 5-year-old male who was presented to the ED with seizures. Pre-operative axial (g) T1-WI, (h) T2-WI and (i) T2*-WI demonstrated a lesion in the left parietal lobe with central heterogeneity and a low signal intensity peripheral rim in (h) and (i) consistent with a CCM. Post-operative follow-up imaging in (j-l) performed nearly 2 years after (g-i) with the same sequences at the same level, also demonstrated a non-enhancing surgical bed with simple central fluid content and post-surgical low signal intensity hemosiderin deposition. Note the 'blooming' artefact in the left parietal skull from the surgical clip in (j-l). He was asymptomatic for the subsequent 9 years until he presented with seizure recurrence. Further MRI was stable and did not demonstrate acute ICH, or a residual or new lesion. He was commenced on antiepileptic medication and was subsequently discharged from the neurosurgical services.

Tomas Freddi 1 , Lucas Avila Lessa Garcia 1 , Felipe Diego Gomes Dantas 1 , Gustavo Andreis 1 , Bruno Augusto Telles 2 , Arthur Freitas Ferreira 1 , Nelson Paes Fortes Diniz Ferreira 1 1 Hospital do Coracao, Sao Paulo, BRAZIL, 2 Inc-Curitiba, BRAZIL Keywords: vigabatrin, DWI, toxicity, fornix Vigabatrin (VGB) is an antiepileptic drug which is used frequently in patients with infantile spasms, and as first-line treatment in those with tuberous sclerosis. VGB irreversibly inhibits GABA transaminase (GABA-T) and therefore, promotes higher concentrations of gamma-Amino-butyric acid (GABA), the most important inhibitory neurotransmitter in humans, in the neuronal synapses and presynaptic nerves (1). Side effects of VGB include VGBassociated brain abnormalities on magnetic resonance imaging (MRI) (VABAM) and retinopathy resulting in permanent peripheral visual field defects termed VGB-associated visual field loss (VAVFL) (2). The pathophysiology of the intracranial abnormalities is unknown, however the pharmacological effect of increasing GABA concentrations may lead to failure of the Na/K ATP pump, transient vacuolation or neuronal excitotoxicity due to the depolarizing effects of GABA in immature brain and possibly promote vacuolization and intramyelinic edema in the central white matter (3). VABAM are characterized by T2-FLAIR hyperintensity with diffusion restriction, usually located in the basal ganglia, dentate nuclei, dorsal brainstem and thalami (1-8), which reflect regional variation in GABA metabolism. In our retrospective analysis we identified forniceal signal abnormality on diffusion weighted images (DWI) in association to other structures. The fornix are paired thin white matter clusters that course beneath the septum pellucidum and the corpus callosum, connecting the hippocampal formations with the mammillary bodies. Being part of the Papez circuit, the fornices are essential in memory and emotion skills (9). The fornix can be impaired by different etiologies such as vascular, metabolic, tumoral and inflammatory lesions, but its involvement on imaging exams is often overlooked (9, 10, 11). In this case series the objective is to review the main locations of signal abnormality induced by vigabatrin on brain MRI, demonstrate the incidence of fornix involvement and to discuss if it can be a marker for VABAM.

Kate Chandler 1 , Berna Aygun 2 , Dipak Ram 3 , Mars Skae 4 , Amit Maniyar 5 , Amit Herwadkar 6 , Hannah Stockley 6 , Christopher Hilditch 6 , Julija Pavaine 5,7 SC Phocomelia syndrome is a rare autosomal-recessive condition presenting with four limb reduction deformities, pre and post-natal growth retardation, microcephaly, dysmorphism and intellectual disability. We present a 14-year-old boy who was investigated for severe skeletal dysplasia and growth retardation. He is the second child born to consanguineous parents. Their daughter was born with similar limb abnormalities and died aged 18 months following cardiac surgery. This boy has global developmental delay and learning difficulties but no history of epilepsy or stroke and no evident neurological deficits. He is microcephalic and Neuroradiology S97 (2020) 62 :S1-107 (Suppl 1) S dysmorphic with a narrow nose and hypoplastic nasal alae, small ears and micrognathia. His palate is high but there is no cleft. Four limbs are short, with severe fixed flexion deformities and pterygia at the elbows, hips and knee joints. His arms are more severely affected than his legs with bilateral radial club hand. Subsequent ophthalmic assessment identified optic nerve hypoplasia. Skeletal survey demonstrated a number of abnormalities including absent radius bilaterally ( Figure 1 ). MRI Brain revealed bilateral frontoparietal cystic encephalomalacia/gliosis, cerebellar calcifications, and left pericallosal artery aneurysm with evidence of previous subarachnoid and parenchymal bleeding ( Figure 2 ). CT angiography confirmed partially thrombosed 5 mm left pericallosal aneurysm and cerebellar calcifications (Figure 3 ). Cerebral angiogram was attempted but aborted as the patient was unable to obtain appropriate position for vascular access due to his severe lower limbs contractures. Aneurysm management was discussed at clinical meetings. Follow-up cerebral CT and MR angiography were scheduled, surgical options will be considered. Currently family does not accept any invasive treatment. Microarray and chromosome fragility studies were normal. Exome sequencing identified homozygous likely pathogenic variant in the ESCO2 gene, c.1489A>T p.(Ile497Phe). Both parents were shown to be heterozygous carriers. Chromosome studies confirmed premature centromere separation. Biallelic loss of function mutations in ESCO2 gene cause Roberts and SC phocomelia syndromes which present with a spectrum of skeletal and craniofacial malformations including limb reduction defects, cleft lip and palate, microcephaly and growth retardation. Other associated anomalies include cardiac defects, ophthalmological and renal abnormalities. Cerebrovascular disease is rarely described. Moya-Moya disease was identified in a child presenting with hemiparesis. Another patient was reported with intracranial haemorrhage due to multiple intracranial aneurysms. Our patient adds to the literature on intracranial vascular manifestations of SC Phocomelia syndrome. We recommend that cerebral MR angiography should be a part of routine surveillance in patients with this condition, especially if presenting with neurological symptoms.

Glioblastoma multiforme is the most common primary brain tumor, representing 20% of all intracranial tumours, while spinal glioblastoma is extremely rare, accounting for 0,2% of all glioblastomas, 1,4% of spinal gliomas and < 1% of all CNS malignancies.

We present a case of a 42 years-old Portuguese caucasian male presenting to our hospital with a subacute onset of lumbar back pain and progressive weakness of the lower limbs. After a lumbar MRI showing anterolisthesis and foraminal stenosis, he was submited to L4-L5 arthrodesis and foraminal decompression, without any improvement. Four months later, development of headaches, nausea, confusion and osbtipation motivated several episodes on the emergency room.

Posteriorly he was submited to another orthopedic procedure. On post-op day, it was documented limitation of abduction of the right eye, paraplegia, hyposthesia below the C5 sensory dermatome and then he underwent neuro-axis MRI.

Spine MRI showed an extensive dorsal intramedullary lesion between C7 and D10, with diffuse pial enhancement extending to the brainstem, placing the differential diagnosis between primary neoformative, metastatic or inflammatory lesion. Brain MRI showed involvement of temporal mesial and brainstem structures, suggesting limbic encephalitis and rhombencephalitis, most likely of paraneoplastic etiology. There was no indication for surgery, and corticoid and Ig treatment was performed, without any improvement.

Histopathological study was compatible with GBM and then he was proposed to radiotherapy. There were no other lesions.

Our main goal is to analyse the radiological characteristics of spinal glioblastoma and to report some atypical features of this case. MRI is the goldstandard imaging modality to diagnose spinal tumours; The majority of spinal glioblastomas appear typically as T1 isohypointense and T2 hyperintense and the thoracic region is the most common location. However, some radiological features have to be analyzed in order to make the differential diagnosis from other tumoral or inflammatory lesions.

Although metastatic spinal lesions are more commonly seen in brain glioblastomas, paraneoplastic brain lesions are not often reported in primary spinal glioblastomas. In this case, we strongly believe that a spinal lesion was settled before the onset of brain lesions, although there is no previous radiological study. Spinal glioblastoma is an infrequently entity, much rarer in adulthood and the presentation with the atypical brain manifestations made the etiological approach even more complex. Despite aggressive treatment, spinal glioblastoma usually progresses rapidly with a poor prognosis. The approach of metastatic disease remains palliative.

Neuroradiology (2020) S98 62 :S1-107 (Suppl 1) S Diffusion Tensor Imaging (DTI) is a sensitive tool in the evaluation of white matter integrity in the central nervous system. Our goal is to evaluate whether abnormal fractional anisotropy (FA) and apparent diffusion coefficient (ADC) values may be seen in patients with cervical spondylotic myelopathy (CSM), independently of radiological signs of myelopathy on conventional magnetic resonance imaging (MRI) sequences.

We performed a retrospective search on our institutional database for all patients who underwent a cervical spine MRI with reduced FOV DTI in 3.0T between November 2018 and December 2019. A total of 58 subjects were selected, 34 were excluded due to non-spondylotic myelopathy24 subjects were finally included: 15 with CSM and 9 without abnormal findings on conventional MRI sequences, which were used as healthy controls (HC). We analyzed the cross-sectional spinal cord area, the FA and ADC values. In CSM subjects, we evaluated the level with the most severe compression and compared it to the immediately superior segment without clear sign of stenosis in the same subject, and to HC. Subjects with CSM were further discriminated between those with or without radiological signs of myelopathy on conventional MRI sequences. Stata 15.0 (StataCorp, College Station, Texas) was used for statistical analysis. The main demographic characteristics were represented using central tendency and dispersion statistics. The Wilcoxon signed-rank test was used to determine statistically significant differences between groups.

We found statistically significant differences in ADC and FA between the maximum compression level and the immediately superior level in the same subject (p-values = 0.019 and 0.03 respectively). There were no statistically significant differences between HC and CSM subjects or between CSM subjects with or without radiological signs of myelopathy.

Our results show that DTI may identify spinal cord damage earlier than morphological changes on conventional MRI sequences. Comparing DTI values in the same subject is more reliable than comparing DTI values between different subjects.

Charcot-Marie-Tooth (CMT) disease is an inherited neurological disorder presenting as peripheral neuropathy. The diagnosis is typically made through neurological history, examination and neurophysiological studies. In the current study, we investigate the utility of magnetic resonance imaging (MRI) with diffusion tensor imaging (DTI) in determining the microstructural integrity of the sciatic and peroneal nerves and the severity of muscle atrophy.

Nitrous oxide (N2O), commonly known as laughing gas, has become a popular recreational drug. Subacute combined degeneration of the spinal cord is a known side effect caused by a nitrous oxide-induced functional B12 deficiency. Together with the increasing recreational popularity of this gas, an increase in the reported patients with neurological side effects is also seen. Here a case of a young woman with typical MRI findings of nitrous oxide-induced subacute combined degeneration is presented.

A 23-year old healthy woman presenting with a gradually progressing sleeping sensation, first in both legs and later in both hands, over a period of 3 days was referred to the neurological department. She reported the inhalation of 100-150 Neuroradiology S101 (2020) 62 :S1-107 (Suppl 1) S nitrous oxide gas canisters per week over the last 2-3 months. The neurological examination revealed severe sensory loss with a loss of fine touch in legs and hands with a sensibility border at C3/C4 and the complete loss of positioning and vibration sensation in both legs. She presented with sensory ataxia, a positive Romberg's test and an unstable broad gait. Blood tests revealed a low B12 level (74 pmol/L). A 1.5T Philips Ingenia MRI scan of the spinal cord was performed including sagittal and axial T2 weighted images.

MRI revealed bilateral longitudinal T2 hyperintensities along the entire length of the spinal cord except for the most caudal 1.6 cm. Axial images showed that the hyperintensities were in the posterior columns with possible involvement of the lateral columns in the cervical region.

This case illustrates typical MRI findings consistent with subacute combined degeneration caused by nitrous oxide abuse. The T2 hyperintensities are symmetrical and seen in a typical inverted V-sign or "rabbit ears" localised primarily in the dorsal columns of the spinal cord which is consistent with symptoms of sensory loss. The changes are associated with swelling of the myelin sheaths and spongy vacuolation with possible axonal damage and gliosis. Changes are typically most pronounced in the cervical and upper thoracic regions of the spinal cord, which is also seen in this patient. Both MRI findings and clinical symptoms of subacute combined degeneration have been reported as being reversible in varying degrees depending on severity and duration of symptoms and abuse. Early diagnosis is therefore crucial to prevent lasting neurological deficits in this primarily young group of patients. This patient improved at first follow-up visit; long-term outcome remains to be seen. To evaluate the lumbar nerve root alterations in patients with lumbar disc herniation sciatica using advanced multimodality MR sequences and the correlations with clinical and neurophysiological findings

We prospectively evaluated 45 patients suffering from unilateral lumbar radiculopathy due to discoradicular conflict. Al patients underwent MRI examinations using a standard MRI protocol and additional advanced MRI sequences (DWI, DTI, and T2 mapping sequences). Relative metrics of ADC, FA, and T2 relaxation times were recorded placing ROIs at the pre-, foraminal, and post-foraminal level, either at the affected side and the contralateral side, used as control. Al patients were also submitted to electromyography testing, recording the spontaneous activity, voluntary activity, F wave amplitude and latency, and motor evoked potentials (MEP) amplitude and latency, either at the level of the tibialis anterior and the gastrocnemius. Clinical features (disease duration, pain, sensitivity, strength, osteotendinous reflexes) were also recorded.

Among clinical features, we found a positive correlation of pain intensity with ADC values of the lumbar nerve roots. The presence of spontaneous activity was correlated with lower ADC values of the affected lumbar nerve root. F wave and MEP latency were correlated with decreased FA values at the foraminal level and increased values at the post-foraminal level. The same neurophysiological measures correlated positively with pre-foraminal T2 mapping values and negatively with post-foraminal T2 mapping values. Increased T2 mapping values at the foraminal level were correlated with disease duration.

Evaluation of lumbar nerve roots using advanced MRI sequences may provide useful clinical information in patients with lumbar radiculopathy, potentially indicating active inflammation/myelinic damage (DTI, T2 mapping) and axonal damage/chronicity (DWI Neuroradiology S107 (2020) 62 :S1-107 (Suppl 1) S

REPORT laminar necrosis secondary to chronic venous hypertension, signifying chronic venous brain ischaemia as the underlying relatable cause. 4-P14 PAEDIATRIC CEREBRAL CAVERNOUS MALFORMATIONS: AN EDUCATIONAL PICTORIAL REVIEW

Automated 3D segmentation of the craniofacial skeleton is enhanced through the incorporation of a modified segmentation algorithm utilising ZTE. These techniques are transferrable to FIESTA-C imaging which offers reduced acquisition time and therefore improved clinical utility. Nine patients with CMT and nine age matched healthy controls were recruited. 3-Tesla MRI with in/out phase and DTI sequences were performed on the dominant limb of patients. DTI values for the sciatic and peroneal nerves were obtained. The degree of muscle atrophy was determined based on the Goutallier classification [ Figure 1 ]. DTI values and muscle atrophy grades were compared between CMT and control groups.

In the CMT group, the fractional anisotropy (FA) values were significantly reduced and the radial diffusivity (RD) values significantly increased in the sciatic and peroneal nerves when compared to controls (p<0.05) [ Figure 2 ]. The muscle atrophy grade was also significantly higher (p<0.001) in the CMT group [ Figure 3 ] and there were significant correlations between the degree of neuropathy and muscle atrophy grade for both sciatic (p<0.001) and peroneal nerves (p=0.001).

MRI-DTI was able to detect peripheral neuropathy and determine the degree of muscle atrophy in CMT. This modality can serve as a useful screening tool for asymptomatic at-risk patients in order to aid early detection. 

Dural ectasia is defined as the widening of the spinal dural sac, most frequently seen at the lumbosacral region and can result in posterior vertebral scalloping. It is associated with herniation of nerve root sleeves and can be related to certain diseases such as Marfan syndrome, neurofibromatosis type I and Ehlers-Danlos syndrome. However, there are a few reports of ankylosing spondylitis (AS) patients with symptomatic dural ectasia, a rare manifestation of long-standing AS. The mechanism is not yet fully understood, but it appears to be related to chronic changes regarding decreased elasticity of the dural sac due to the inflammatory process of AS. Neurological complications of AS are rare and manifestation with radiculopathy is uncommon. It is postulated that AS neurological symptoms are due to the compression of nerve tissue adherent to the inflamed dura. MR imaging of dural ectasia usually reveals characteristic expansion of the lumbar spinal canal with scalloping of the laminae and spinous processes and plays an essential role on the diagnosis, excluding other intradural pathologies.

A 78-year-old male patient, diagnosed with AS at the age of 16, was referenced to neurosurgery consultation with a 2-year history of paresthesias in his left leg. Neurological examination revealed commitment of thermo-algic sensation at L5-territory and ipsilateral distal hypopallesthesia. CT scan showed flowing ossification of the paraspinous ligaments, fusion of the facet joints, diffuse syndesmophyte ankylosis with squaring of the vertebral bodies, producing a "bamboo spine" appearance typical of AS and could also be noted expansion of the spinal canal with remodeling and erosion of the posterior elements of the vertebrae including the pedicles, laminae and spinous processes from L1 to L5. Lumbar MRI revealed posterior dural ectasia from L1 to L5 with scalloping of the posterior elements. Nerve roots of the cauda equina were deviated into the dural ectasia region. Conservative treatment and annual follow-up were adopted.

Although rarely, neurological complications such as radiculopathy or even cauda equine syndrome can occur in AS patients with long-standing spinal fusion and dural ectasia. It is of paramount importance to be acquainted with the typical imaging findings of dural ectasia in order to establish a correct diagnosis and to provide proper management. Keywords: disc pathology, Intervertebral disc calcification, pediatric

Pediatric idiopathic intervertebral disc calcification (PIIVDC) is an uncommon condition typically affecting children aged between 5 to 12 years. It is characterized by calcification of an intervertebral disc, most commonly at the lower cervical level, which can progress to inflammation or extrusion. Seldomly, PIIVDC can be associated with a calcified disc herniation into the spinal canal to an extent that causes neurological compromise. Most cases are asymptomatic and are diagnosed incidentally during routine screening or clinical evaluation of neck pain, torticollis, fever or radiculopathy.

A 12-year-old male presented to the emergency unit due to recurrent back pain with poor response to anti-inflammatory drugs. Thoracic spine radiography and computed tomography demonstrated calcifications in the T3-T4 and T7-T8 intervertebral discs. Laboratory findings, including calcium metabolism and hormones, were normal. Due to persistent symptoms, a thoracic spine Magnetic Resonance Imaging (MRI) was performed nine days later, which revealed a left postero-lateral epidural mass from T4 to T10 causing noncompressive right deviation of the spinal cord. An acute hematoma was suspected due to a hypointense signal on T2-and an isointense signal on T1-weighted sequences. One week later, follow-up MRI revealed a dimensional increase of the mass and depicted its continuity with the disc through the left lateral epidural space; the signal remained low on T2-weighted sequences, suggesting marked calcification, with a slight rim enhancement after contrast administration. Spontaneous calcification on herniated and migrated disc material was considered. Since there were no neurological deficits, a conservative treatment was followed. Follow-up MRI at 1-month showed sub-total involution of both the posterior epidural calcified mass, and of the corresponding calcified disc space.

To the best of our knowledge, this is the first description of a pediatric idiopathic calcification on a migrated disc fragment. The prevalence of disc herniation in PIIVDC is superior to 30% in symptomatic patients, however, neurological deficits are rarely reported. Over time, the calcified lesion spontaneously involutes, characterizing this pathology. In symptomatic cases, there is usually resolution of symptoms in less than 6 months and recurrence rarely occurs. Therefore, the best approach to this condition is conservative treatment and an invasive treatment should be reserved only for cases with persistent neurological deficits.

This case illustrates an unusual presentation of a known entity and confirms that a correct and early diagnosis is essential in order to avoid unnecessary invasive testing and surgical intervention.