Neuroradiology Fellow Research Articles

Improving Deep Learning Models for Pediatric Low-Grade Glioma Tumours Molecular Subtype Identification Using MRI-based 3D Probability Distributions of Tumour Location.

Purpose: Pediatric low-grade gliomas (pLGG) are the most common brain tumour in children, and the molecular diagnosis of pLGG enables targeted treatment. We use MRI-based Convolutional Neural Networks (CNNs) for molecular subtype identification of pLGG and augment the models using tumour location probability maps. Materials and Methods: MRI FLAIR sequences of 214 patients (110 male, mean age of 8.54 years, 143 BRAF fused and 71 BRAF V600E mutated pLGG tumours) from January 2000 to December 2018 were included in this retrospective REB-approved study. Tumour segmentations (volumes of interest-VOIs) were provided by a pediatric neuroradiology fellow and verified by a pediatric neuroradiologist. Patients were randomly split into development and test sets with an 80/20 ratio. The 3D binary VOI masks for each class in the development set were combined to derive the probability density functions of tumour location. Three pipelines for molecular diagnosis of pLGG were developed: location-based, CNN-based, and hybrid. The experiment was repeated 100 times each with different model initializations and data splits, and the Areas Under the Receiver Operating Characteristic Curve (AUROC) was calculated, and Student's t-test was conducted. Results: The location-based classifier achieved an AUROC of 77.9, 95% confidence interval (CI) (76.8, 79.0). CNN-based classifiers achieved an AUROC of 86.1, 95% CI (85.0, 87.3), while the tumour-location-guided CNNs outperformed the other classifiers with an average AUROC of 88.64, 95% CI (87.6, 89.7), which was statistically significant (P-value .0018). Conclusion: Incorporating tumour location probability maps into CNN models led to significant improvements for molecular subtype identification of pLGG.

NIMG-15. RAPID EARLY PROGRESSION AFTER SURGERY ON PRE-RADIATION PLANNING MRI FOR IDH-WILDTYPE GLIOBLASTOMA

Abstract BACKGROUND Use of pre-RT planning MRI for glioblastoma (GBM) is inconsistent and not required by many clinical protocols. We performed a detailed volumetric analysis of rapid early progression (REP) on pre-RT planning MRI’s, hypothesizing that REP is associated with worse survival and more accurately risk stratifies than extent of resection. METHODS Patients with IDHwt GBM who had resection (biopsy-only patients excluded) and pre-RT planning MRI ≥2 weeks after postop MRI were included. MRI’s were reviewed by a neuroradiology fellow and attending who measured in 3 dimensions new/increased contrast-enhancing disease between postop and pre-RT MRI. REP was defined as disease ≥1cm in 2 dimensions or ≥0.5cc. Extent of resection was scored 1: gross total (GTR, >95%); 2:near total (NTR, 90-95%); 3:subtotal (STR). The primary endpoint was survival (OS). Cox proportional hazards model analyzed association with age, resection extent, MGMT+ status, short course RT (a surrogate for performance status), and REP. RESULTS The final cohort included 77 patients (median age 63), with 9 (12%) receiving short course RT. Surgical extent was GTR in 42 (55%), NTR in 16 (21%), and STR in 19 (25%). REP occurred in 41 patients (53%), with a rate of 29% after GTR, 75% after NTR, and 90% after STR. Age (p=0.24, HR 1.02/yr), short course RT (p=0.64, HR 1.2), and surgical extent (p=0.49, HR 1.1) were not significantly associated with OS on univariate analysis. MGMT+ and REP were associated with survival on both univariate and multivariate (MGMT+ p=0.002, HR 0.4; REP p=0.005, HR 2.3) analyses. Median OS was 1.3 years vs. 1.9 years with vs without REP. CONCLUSIONS REP occurs frequently, adds a “growth rate” variable, and is a better prognostic factor than extent of resection. These data have implications for clinical trial design and support pre-RT planning MRI’s as standard of care.

Facial trauma education in radiology: using surgeon feedback as the benchmark for success.

Interpreting CT studies of facial trauma is challenging, and there are often substantial differences in the characterization of complex facial trauma between radiologists and surgeons. We designed a collaborative multidisciplinary project to reconcile differences in facial fracture interpretation through an educational intervention. The effectiveness of this intervention was evaluated through surgeon feedback on radiology reports. Radiology residents, neuroradiology fellows, and neuroradiology attendings were recruited as participants at a single tertiary care academic center. Otolaryngology residents were recruited as evaluators. Participants completed surveys and provided preliminary reports for example cases of facial trauma before and after attending an educational session. Evaluators performed a blinded review of these preliminary reports based on ideal reports developed by surgical and neuroradiology attendings. 26 participants (20 residents, 1 neuroradiology fellow, 5 neuroradiology attendings) completed the study. Six otolaryngology residents participated as evaluators. To assess interrater reliability, three evaluators graded a shared set of 15 reports and demonstrated substantial agreement with a Kendall's W of 0.71. Participants demonstrated significant improvement in overall report accuracy, clarity, and organization. In subunit analysis, there were significant improvements in reporting Le Fort, nasoseptal, and nasoorbitoethmoid fractures. No significant improvements occurred in the reporting of upper face, zygomaticomaxillary complex, or mandibular fractures. In contrast, survey analysis demonstrated significantly improved confidence in interpreting trauma involving all facial subunits. Compared with survey results, surgeon assessment of radiology reports better demonstrated areas of improvement after an educational intervention. A multidisciplinary approach to assessing educational efforts may better evaluate the practical effectiveness of educational interventions.

Geniculate ganglion diverticulum: a potential imaging marker in patients with idiopathic intracranial hypertension

PurposeThe diagnosis of idiopathic intracranial hypertension (IIH) is often challenging in patients who do not present with classic symptoms. Brain MRI can play a pivotal role, as several imaging findings, such as an empty sella appearance (ESA), have been shown to be associated with IIH. Yet, none of the MRI signs have been shown to have a high sensitivity and specificity. In this study, we tested the hypothesis that presence of a geniculate ganglion diverticulum (GGD) is a potential imaging marker for the detection of IIH. Materials and methodsThis is an IRB-approved, single-institution, retrospective, observational study. Brain MRI examinations of patients referred to Radiology by Otology clinic over a period of 10 years were reviewed. 244 MRI exams fulfilling inclusion and exclusion criteria were independently screened for the presence of GGD and ESA by two Neuroradiology fellows. Electronic medical records (EMR) of patients in this study were reviewed for presence of clinical manifestations of IIH. Receiver operator characteristic (ROC) curves were generated to estimate the accuracy of each covariate in diagnosing IIH. The area under each ROC curve (AUC) was calculated to identify an accurate prognostic covariate. Statistical analysis was done using R programming language V 4.2.2. ResultsGGD was identified in MRI exams of 51 patients. A 2:1 propensity score (PS) matching for age, gender, and Body Mass Index (BMI) was used to select non-GGD control group for comparison with the GGD group. There was strong agreement between the 2 reviewers (kappa = 0.81, agreement = 95 %). Twelve patients in this study were diagnosed with IIH. There was a high incidence of GGD (OR = 12.19, 95 % CI (2.56, 58.10)) and ESA (OR = 4.97, 95 % CI (1.47, 16.74)) in IIH patients. The AUC observed in GGD for predicting IIH was 0.771 (0.655–0.888), specificity = 0.709 (0.638–0.780), and sensitivity = 0.833 (0.583–1). The AUC observed for ESA in predicting IIH was 0.682 (0.532–0.831), specificity = 0.780 (0.709–0.844), and sensitivity = 0.583 (0.333–0.833). ConclusionGGD is potentially a novel imaging marker of IIH with sensitivity higher than and specificity comparable to that of ESA. Clinical relevance statementPresence of GGD should raise the possibility of IIH.

Implementing a rapid cord compression Magnetic Resonance Imaging protocol in the emergency department: Lessons learned.

(1) Evaluate efficacy of an abbreviated total spine protocol in triaging emergency department (ED) patients through retrospective evaluation. (2) Describe patient outcomes following implementation of a rapid cord compression protocol. (1) All contrast-enhanced total spine magnetic resonance imaging studies (MRIs) performed on ED patients (n = 75) between 10/1-12/31/2022 for evaluation of cord compression were included. Two readers with 6 and 5years of experience blindly reviewed the abbreviated protocol (comprised of sagittal T2w and axial T2w sequences) assessing presence of cord compression or severe spinal canal stenosis. Ground truth was consensus by a neuroradiology fellow and 2 attendings. (2) The implemented rapid protocol included sagittal T1w, sagittal T2w Dixon and axial T2w images. All ED patients (n = 85) who were imaged using the rapid protocol from 5/1-8/31/2023 were included. Patient outcomes and call-back rates were determined through chart review. (1) Sensitivity and specificity for severe spinal canal stenosis and/or cord compression was 1.0 and 0.92, respectively, for reader 1 and 0.78 and 0.85, respectively, for reader 2. Negative predictive value was 1.0 and 0.97 for readers 1 and 2, respectively. (2) The implemented rapid cord compression protocol resulted in 60% reduction in imaging time at 1.5T. The call-back rate for additional sequences was 7%. In patients who underwent surgery, no additional MRI images were acquired in 82% of cases (9/11). Implementing an abbreviated non-contrast total spine protocol in the ED results in a low call-back rate with acquired MRI images proving sufficient for both triage and treatment planning in most patients.

Analysis of Requirements within Neuroradiology Job Advertisements in a Specific Timeframe.

The neuroradiology job market is constantly changing along with the skill sets needed by fellowship graduates to participate successfully in the job market. This study aimed to establish a baseline of employer requirements in the neuroradiology job market within the study timeframe. The American Society of Neuroradiology and American College of Radiology job boards were queried for neuroradiology positions between August 12, 2022, and December 31, 2022. The positions and requirements were categorized into academic versus private practice, general diagnostic radiology, full-subspecialized neuroradiology, hybrid remote/onsite, outpatient, inpatient/emergency, general interventional radiology procedures, and neuroradiology procedures. Exclusion criteria included neurointerventional only, remote-only, pediatric-only, no preference for neuroradiology, and duplicate posts within and between the job boards. Of 1777 total job posts, 179 were neuroradiology-specific and the remainder were general. Of the 179 neuroradiology-specific jobs, 55 neuroradiology jobs were academic and 124 jobs were private practice. A higher proportion of private practice jobs required general diagnostic interpretations (83% versus 26%), a higher proportion of academic jobs required neuroradiology procedures (56% versus 31%), and a higher proportion of private practice jobs required general interventional radiology procedures (22% versus 0%). Thirty-nine percent of all neuroradiology-specific onsite jobs required neuroradiology procedures, and 15% required general interventional radiology procedures. Because there was a sizable difference between general radiology and procedure requirements between academic and private practice positions, tailoring fellowship training for career aspirations of neuroradiology fellows should be considered to adapt to the skills needed for the evolving job market. In the queried timeframe, 61% of neuroradiology-specific onsite jobs did not have a specific procedure skill requirement for job applicants. This article serves as a single snapshot of the job market and its requirements for neuroradiologists, to aid in planning training to meet the needs of employers.

Five-Year Serial Brain MRI Analysis of Military Members Exposed to Chronic Sub-Concussive Overpressures.

The Canadian Special Operations Forces Command conducts explosives operations and training which exposes members to explosive charges at close proximity. This 5-year longitudinal trial was conducted in follow-up to our initial trial which examined military breachers with MRI and EEG pre and post blast exposure. To examine brain MRI findings in military personnel exposed to multiple repeated blast exposures. Five-year longitudinal prospective trial. Ninety-two males aged 23-42 with an average of 9.4 years of blast exposure. 3 T brain MRI/T1-weighted 3D with reconstruction in three planes, T2-weighted, T2-weighted fluid attenuated inversion recovery (FLAIR) 3D with reconstruction in three planes, T2-weighted gradient spin echo (GRE), saturation weighted images, DWI and ADC maps, diffusion tensor imaging. All MRI scans were interpreted by the two neuroradiologists and one neuroradiology Fellow in a blinded fashion using a customized neuroradiology reporting form. Matching parametric statistics represented the number of participants whose brain parameters improved or deteriorated over time. Odds ratio (OR) and 95% confidence intervals (CI) were computed using log regression modeling to determine volume loss, white matter lesions, hemosiderosis, gliosis, cystic changes and enlarged Virchow Robin (VR) spaces. A Kappa (κ) statistic with a 95% CI was calculated to determine rater variability between readers. A significant deterioration was observed in volume loss (OR = 1.083, 95% CI 0.678-1.731, permutation test), white matter changes (OR: 0.754, 95% CI 0.442-1.284, permutation test), and enlargement of VR spaces (OR: 0.775, 95% CI 0.513-1.171). Interrater reliability was low: κ = 0.283, 0.156, and 0.557 for volume loss, white matter changes, and enlargement of VR spaces, respectively. There were significant changes in brain volume, white matter lesions, and enlargement of VR spaces. 2 TECHNICAL EFFICACY: Stage 2.

Workflow interruptions in an era of instant messaging: A detailed analysis

IntroductionThe complex practice environment and responsibilities incumbent on diagnostic radiologists creates a workflow susceptible to disruption. While interruptions have been shown to contribute to medical errors in the healthcare delivery environment, the exact impact on highly subspecialized services such as diagnostic radiology is less certain. One potential source of workflow disruption is the use of a departmental instant messaging system (Webex), to facilitate communications between radiology faculty, residents, fellows, and technologists. A retrospective review was conducted to quantify the frequency of interruption experienced by our neuroradiology fellows. Materials and methodsData logs were gathered comprising all instant messages sent and received within the designated group chats from July 5–December 31, 2021, during weekday shifts staffed by neuroradiology fellows. Interruptions per shift were calculated based on month, week, and day of the week. Results14,424 messages were sent across 289 total shifts. The 6 fellows assigned to the main neuroradiology reading room sent 3258 messages and received 10,260 messages from technologists and other staff. There was an average of 50 interruptions per shift when examined by month (range 48–53), and 52 interruptions per shift when examined by day of the week (range 40–60). ConclusionNeuroradiology fellows experience frequent interruptions from the departmental instant messaging system. These disruptions, when considered in conjunction with other non-interpretative tasks, may have negative implications for workflow efficiency, requiring iterative process improvements when incorporating new technology into the practice environment of diagnostic radiology.

My Short-Term Neurointerventional Training in Thailand Supported by a Women's Observership Program of the World Federation of Interventional and Therapeutic Neuroradiology.

This short report describes my personal experience of 1-month neurointerventional training at Ramathibodi Hospital in Thailand, supported by a women's observership grant from the World Federation of Interventional and Therapeutic Neuroradiology, in which many interventional neuroradiology (INR) fellows from various regions of Thailand also participated. The training program allowed me to experience numerous neurointerventional cases and to acquire skills on how to function as a member of the INR team. This experience prompts me to contemplate the significance of team-based medicine and the role of women in the field of neurosurgery and INR.

Validation of a deep learning model for traumatic brain injury detection and NIRIS grading on non-contrast CT: a multi-reader study with promising results and opportunities for improvement.

This study aimed to assess and externally validate the performance of a deep learning (DL) model for the interpretation of non-contrast computed tomography (NCCT) scans of patients with suspicion of traumatic brain injury (TBI). This retrospective and multi-reader study included patients with TBI suspicion who were transported to the emergency department and underwent NCCT scans. Eight reviewers, with varying levels of training and experience (two neuroradiology attendings, two neuroradiology fellows, two neuroradiology residents, one neurosurgery attending, and one neurosurgery resident), independently evaluated NCCT head scans. The same scans were evaluated using the version 5.0 of the DL model icobrain tbi. The establishment of the ground truth involved a thorough assessment of all accessible clinical and laboratory data, as well as follow-up imaging studies, including NCCT and magnetic resonance imaging, as a consensus amongst the study reviewers. The outcomes of interest included neuroimaging radiological interpretation system (NIRIS) scores, the presence of midline shift, mass effect, hemorrhagic lesions, hydrocephalus, and severe hydrocephalus, as well as measurements of midline shift and volumes of hemorrhagic lesions. Comparisons using weighted Cohen's kappa coefficient were made. The McNemar test was used to compare the diagnostic performance. Bland-Altman plots were used to compare measurements. One hundred patients were included, with the DL model successfully categorizing 77 scans. The median age for the total group was 48, with the omitted group having a median age of 44.5 and the included group having a median age of 48. The DL model demonstrated moderate agreement with the ground truth, trainees, and attendings. With the DL model's assistance, trainees' agreement with the ground truth improved. The DL model showed high specificity (0.88) and positive predictive value (0.96) in classifying NIRIS scores as 0-2 or 3-4. Trainees and attendings had the highest accuracy (0.95). The DL model's performance in classifying various TBI CT imaging common data elements was comparable to that of trainees and attendings. The average difference for the DL model in quantifying the volume of hemorrhagic lesions was 6.0mL with a wide 95% confidence interval (CI) of - 68.32 to 80.22, and for midline shift, the average difference was 1.4mm with a 95% CI of - 3.4 to 6.2. While the DL model outperformed trainees in some aspects, attendings' assessments remained superior in most instances. Using the DL model as an assistive tool benefited trainees, improving their NIRIS score agreement with the ground truth. Although the DL model showed high potential in classifying some TBI CT imaging common data elements, further refinement and optimization are necessary to enhance its clinical utility.

Diagnostic Accuracy and Reliability of Noncontrast Computed Tomography Markers for Acute Hematoma Expansion among Radiologists.

Noncontrast Computed Tomography (NCCT) features are promising markers for acute hematoma expansion (HE) in patients with intracerebral hemorrhage (ICH). It remains unclear whether accurate identification of these markers is also reliable in raters with different levels of experience. Patients with acute spontaneous ICH admitted at four tertiary centers in Germany and Italy were retrospectively included from January 2017 to June 2020. In total, nine NCCT markers were rated by one radiology resident, one radiology fellow, and one neuroradiology fellow with different levels experience in ICH imaging. Interrater reliabilities of the resident and radiology fellow were evaluated by calculated Cohen's kappa (κ) statistics in reference to the neuroradiology fellow who was referred as the gold standard. Gold-standard ratings were evaluated by calculated interrater κ statistics. Global interrater reliabilities were evaluated by calculated Fleiss kappa statistics across all three readers. A comparison of receiver operating characteristics (ROCs) was used to evaluate differences in the diagnostic accuracy for predicting acute hematoma expansion (HE) among the raters. Substantial-to-almost-perfect interrater concordance was found for the resident with interrater Cohen's kappa from 0.70 (95% CI 0.65-0.81) to 0.96 (95% CI 0.94-0.98). The interrater Cohen's kappa for the radiology fellow was moderate to almost perfect and ranged from 0.58 (95% CI 0.52-0.65) to 94 (95% CI 92-0.97). The intrarater gold-standard Cohen's kappa was almost perfect and ranged from 0.79 (95% CI 0.78-0.90) to 0.98 (95% CI 0.78-0.90). The global interrater Fleiss kappa ranged from 0.62 (95%CI 0.57-0.66) to 0.93 (95%CI 0.89-0.97). The diagnostic accuracy for the prediction of acute hematoma expansion (HE) was different for the island sign and fluid sign, with p-values < 0.05. The NCCT markers had a substantial-to-almost-perfect interrater agreement among raters with different levels of experience. Differences in the diagnostic accuracy for the prediction of acute HE were found in two out of nine NCCT markers. The study highlights the promising utility of NCCT markers for acute HE prediction.

An Educational Graphical User Interface to Construct Convolutional Neural Networks for Teaching Artificial Intelligence in Radiology.

Deep learning techniques using convolutional neural networks (CNNs) have been successfully developed for various medical image analysis tasks. However, the skills to understand and develop deep learning models are not usually taught during radiology training, which constitutes a barrier for radiologists looking to integrate machine learning (ML) into their research or clinical practice. In this work, we developed and evaluated an educational graphical user interface (GUI) to construct CNNs for teaching deep learning concepts to radiology trainees. The GUI was developed in Python using the PyQt and PyTorch frameworks. The functionality of the GUI was demonstrated through a binary classification task on a dataset of MR images of the brain. The usability of the GUI was assessed through 45-min user testing sessions with 5 neuroradiologists and neuroradiology fellows, assessing mean task completion times, the System Usability Scale (SUS), and a qualitative questionnaire as metrics. Task completion times were compared against a ML expert who performed the same tasks. After a 20-min introduction to CNNs and a walkthrough of the GUI, users were able to perform all assigned tasks successfully. There was no significant difference in task completion time compared to a ML expert. The educational GUI achieved a score of 82.5 on the SUS, suggesting that the system is highly usable. Users indicated that the GUI seems useful as an educational tool to teach ML topics to radiology trainees. An educational GUI allows interactive teaching in ML that can be incorporated into radiology training.

Effectiveness of a conceptual three-dimensionally printed model of the middle ear in teaching complex neuroanatomy to radiology trainees

Introduction Middle ear anatomy is difficult for learners because of its intricate and complex anatomy. Historically its anatomy has been taught with dissections and figures. 3D printed models have grown in popularity for their ability to represent complex structures. This study sought to assess the efficacy of a conceptual 3D printed middle ear model in radiology trainee education.Methods An uncontrolled before-after trial was performed in which radiology trainees participated in small group teaching sessions using a 3D printed conceptual middle ear model. Participant knowledge was assessed with identical pre- and 1-week post-intervention knowledge assessments and surveys.Results A total of 26 participants completed the study. The mean pre-intervention test score for participants (out of 20) was 6 ± 3.4, which increased to 11.7 ± 3.5 (p-value < 0.02) following interaction with the model. Second year radiology residents had the largest improvement in score, 9.0 ± 4.2, while fourth year radiology residents had the least, 2.8 ± 2.6. The small increase in post-intervention scores for the neuroradiology fellows was not found to be statistically significant (p-value 0.07). Subgroup analysis of post-intervention knowledge found no statistical difference among participants of different years of training. The survey showed increased understanding and desire for incorporation into curriculum.Discussion: Interaction with the 3D printed model was found to improve anatomical knowledge in radiology residents but not neuroradiology fellows, whose improvement was not statistically significant. All participants, regardless of their years of training, were found to have knowledge equivalent to that of a fellow following their training.

Pre-operative imaging and post-operative appearance of standard paediatric neurosurgical approaches: a training guide for neuroradiologists.

A short-cut narrative review was conducted according to the SANRA guidelines to identify studies describing normal and abnormal postoperative radiological features of the most common paediatric neurosurgical procedures. Rather than focusing on the original pathology addressed by neurosurgical means, this review explored three main areas of operative neurosurgery: ventricular access, supratentorial & infratentorial craniotomies, and posterior fossa/craniocervical junction decompression. A total of twenty-three landmark papers were included for review based on their relevance to address the research question and serve as a practical guide for paediatric neuroradiology trainees and fellows. Accurate in text referencing of the ClinicalTrials.gov identifier, and weblink, has also been provided for all trials discussed in the results section. All the above is complemented by relevant iconography meant to describe a wide range of postoperative changes and early complications. Finally, the review is enriched by a discussion touching upon haemostatic agents, intentionally retained foreign bodies and the future of machine learning for neuroradiology reporting. Overall, the information presented in a systematic fashion will not only help trainees and fellows to deepen these topics and expand their knowledge in preparation for written and oral boards, but will also represent a useful resource for everyone including trained neuroradiologists and neurosurgeons themselves.

Am I Ready to Be an Independent Neuroradiologist? Objective Trends in Neuroradiology Fellows' Performance during the Fellowship Year.

Aside from basic Accreditation Council for Graduate Medical Education guidelines, few metrics are in place to monitor fellows' progress. The purpose of this study was to determine objective trends in neuroradiology fellowship training on-call performance during an academic year. We retrospectively reviewed the number of cross-sectional neuroimaging studies dictated with complete reports by neuroradiology fellows during independent call. Monthly trends in total call cases, report turnaround times, relationships between volume and report turnaround times, and words addended to preliminary reports by attending neuroradiologists were evaluated with regression models. Monthly variation in frequencies of call-discrepancy macros were assessed via χ2 tests. Changes in frequencies of specific macro use between fellowship semesters were assessed via serial 2-sample tests of proportions. From 2012 to 2017, for 29 fellows, monthly median report turnaround times significantly decreased during the academic year: July (first month) = 79 minutes (95% CI, 71-86 minutes) and June (12th month) = 55 minutes (95% CI, 52-60 minutes; P value = .023). Monthly report turnaround times were inversely correlated with total volumes for CT (r = -0.70, F = 9.639, P value = .011) but not MR imaging. Words addended to preliminary reports, a surrogate measurement of report clarity, slightly improved and discrepancy rates decreased during the last 6 months of fellowship. A nadir for report turnaround times, discrepancy errors, and words addended to reports was seen in December and January. Progress through fellowship correlates with a decline in report turnaround times and discrepancy rates for cross-sectional neuroimaging call studies and slight improvement in indirect quantitative measurement of report clarity. These metrics can be tracked throughout the academic year, and the midyear would be a logical time point for programs to assess objective progress of fellows and address any deficiencies.

The In-between

In this narrative medicine essay, a geriatrician memory of a neuroradiology fellow’s words of calm when breaking the news that her 8-year-old daughter had a brain tumor often bubbles up years later when she breaks bad news to patients with COVID who are without the support of friends or family.

Identification and Characterization of Leptomeningeal Metastases Using SPINE, A Web-Based Collaborative Platform.

Leptomeningeal metastases (LMs) carry a poor prognosis. Existing LM scoring systems show limited reproducibility. We assessed the contribution of education level on the reproducibility of LM scoring using structured planning and implementationof new experiments (SPINE), a novel web-based platform. Stringent radiological definitions of LM and a customized interactive scoring system were implemented in SPINE. Five patients with brain LM and 3 patients with spine, but no brain LM, were selected. Each patient's baseline post-contrast T1-weighted brain MRI was analyzed by three attending neuroradiologists, two neuroradiology fellows, and two radiology residents. Raters identified and characterized all LMs based on: (1) location (cerebrum, cerebellum, brainstem, ventricle, and/or cranial nerves); (2) shape (nodular and/or linear/curvilinear); (3) size (≥ or <5mm in two orthogonal diameters); (4) spatial extension (focal or diffuse). Inter-rater agreement and association of LM with patient survival were investigated. On average, 6.5 LMs per case were detected. Forty-nine percent of LMs were cerebral, 77.7% were nodular, 86.6% were focal, and 66% were <5 × 5 mm. Agreement on the total number of LMs and the above-mentioned common LM characteristics was higher between attendings (intra-class correlation [ICC] = 0.8-0.94) than fellows (ICC = 0.6-0.82) or residents (ICC = 0.43-0.73). Agreement on ventricular, cranial nerve, and nodular + linear LM was low even between attendings. The number of brainstem LMs showed significant correlation with survival. Structured education using SPINE may improve consistency in LM reporting. Future work should address the impact of the presented approach on the reproducibility of longitudinal analyses directly relevant to the assessment of treatment-response.

Comparison of Diffusion and Perfusion Techniques in Differentiating High Versus Low Grade Pediatric Brain Tumors

Background/Objective: Brain tumors are the most common solid cancer in children and cause significant mortality and morbidity. We compare the effectiveness of different parameters in predicting tumor grade between dynamic contrast enhancement (DCE), intravoxel incoherent motion (IVIM), dynamic susceptibility contrast (DSC) perfusion and diffusion weighted imaging (DWI). &#x0D; &#x0D; Methods: A retrospective blinded review of pediatric brain tumors with DCE, IVIM, DWI, and DSC was performed. Parametric maps were registered to T2 weighted images. Volumetric regions of interest (ROI) were manually segmented from solid tumor components for each patient by a neuroradiologist (CH), neuroradiology fellow (BG), and medical student (EC). Resulting mean values for parameters from DCE (Ktrans, Kep, Ve, Vp,), IVIM (D, D*, f), DSC (rCBV) and DWI (ADC) were compared using Student’s t-test for high- and low-grade tumor groups based on WHO grading from pathology. For significant parameters, receiver operating characteristic (ROC) analysis with area under curve (AUC) was performed. &#x0D; &#x0D; Results: 20 subjects were included with 9 low grade and 11 high grade tumors. Significant differences between low versus high grade were demonstrated for D (10−3 mm2/s) (1.4±0.4 vs 0.9±0.2, p=0.01), f (0.04±0.02 vs 0.07±0.02, p=0.02), ADC (10−3 mm2/s) (1.4±0.4 vs 0.9±0.3, p=0.009) and rCBV (2.2±0.9 vs 4.7±2.1, p=0.003). No significant difference was found for D* or any DCE parameter. AUC from ROC was similar for all significant parameters [D (0.81, p=0.003); f (0.80, p=0.003); ADC (0.83, p=0.001); rCBV (0.83, p=0.0005)]. &#x0D; &#x0D; Conclusion: D and f parameters from IVIM can significantly differentiate high versus low grade pediatric brain tumors similar to ADC and rCBV. Conversely, no DCE parameter was significant. &#x0D; &#x0D; Scientific Implications: The results will assist the selection of MRI sequences that best predict tumor grade, as well as guide tumor biopsy for the most aggressive tumor portions. Further study of these techniques may correlate with molecular profiling and predict outcome.

Subspecialty-Level Deep Gray Matter Differential Diagnoses with Deep Learning and Bayesian Networks on Clinical Brain MRI: A Pilot Study.

To develop and validate a system that could perform automated diagnosis of common and rare neurologic diseases involving deep gray matter on clinical brain MRI studies. In this retrospective study, multimodal brain MRI scans from 212 patients (mean age, 55 years ± 17 [standard deviation]; 113 women) with 35 neurologic diseases and normal brain MRI scans obtained between January 2008 and January 2018 were included (110 patients in the training set, 102 patients in the test set). MRI scans from 178 patients (mean age, 48 years ± 17; 106 women) were used to supplement training of the neural networks. Three-dimensional convolutional neural networks and atlas-based image processing were used for extraction of 11 imaging features. Expert-derived Bayesian networks incorporating domain knowledge were used for differential diagnosis generation. The performance of the artificial intelligence (AI) system was assessed by comparing diagnostic accuracy with that of radiologists of varying levels of specialization by using the generalized estimating equation with robust variance estimator for the top three differential diagnoses (T3DDx) and the correct top diagnosis (TDx), as well as with receiver operating characteristic analyses. In the held-out test set, the imaging pipeline detected 11 key features on brain MRI scans with 89% accuracy (sensitivity, 81%; specificity, 95%) relative to academic neuroradiologists. The Bayesian network, integrating imaging features with clinical information, had an accuracy of 85% for T3DDx and 64% for TDx, which was better than that of radiology residents (n = 4; 56% for T3DDx, 36% for TDx; P < .001 for both) and general radiologists (n = 2; 53% for T3DDx, 31% for TDx; P < .001 for both). The accuracy of the Bayesian network was better than that of neuroradiology fellows (n = 2) for T3DDx (72%; P = .003) but not for TDx (59%; P = .19) and was not different from that of academic neuroradiologists (n = 2; 84% T3DDx, 65% TDx; P > .09 for both). A hybrid AI system was developed that simultaneously provides a quantitative assessment of disease burden, explainable intermediate imaging features, and a probabilistic differential diagnosis that performed at the level of academic neuroradiologists. This type of approach has the potential to improve clinical decision making for common and rare diseases.Supplemental material is available for this article.© RSNA, 2020.

Validation of the focal cerebral arteriopathy severity score (FCASS) in a Swiss cohort: Correlation with infarct volume and outcome.

Focal cerebral arteriopathy (FCA), a major cause of childhood arterial ischemic stroke (AIS), can progress and lead to increased infarct size and/or recurrent stroke. Evaluating treatment options depends on the ability to quantify reliably the degree of stenosis in FCA. We validated the recently introduced FCA severity score (FCASS) in an independent cohort from the Swiss Neuro-Paediatric Stroke Registry (SNPSR). We included children with FCA who had MR or CT angiography and a Pediatric Stroke Outcome Measure (PSOM) at 6-months and 2-years post-stroke. A paediatric neuroradiologist applied the FCASS and the modified pediatric Alberta Stroke Program Early Computed Tomography Score (ASPECTS), a measure of infarct volume, to all available imaging. Two senior paediatric stroke neurologists and a neuroradiology fellow independently assigned FCASS scores to test interrater reliability. Pairwise correlations between FCASS, pedASPECTS, and PSOM were examined. Thirty-two children [median (IQR) age=5.9 (1.8, 9.6), 19 males] were included. The median maximum FCASS score at any time was 9 (IQR 6, 12; range 3, 16). Larger infarct volume scores correlated with both higher maximum FCASS scores and worse post-stroke outcomes, although we found no direct correlation between FCASS and outcomes. Stroke neurologists tended to assign lower FCASS scores than the neuroradiologist, but interrater reliability was predominantly good. In this independent validation cohort, higher maximum FCASS correlated with greater infarct volume scores that also correlated with worse neurological outcomes. Scoring by non-imaging specialists seems to be valuable, although differences are present.