Can large language models pass official high-grade exams of the European Society of Neuroradiology courses? A direct comparison between OpenAI chatGPT 3.5, OpenAI GPT4 and Google Bard.

We compared different LLMs, notably chatGPT, GPT4, and Google Bard and we tested whether their performance differs in subspeciality domains, in executing examinations from four different courses of the European Society of Neuroradiology (ESNR) notably anatomy/embryology, neuro-oncology, head and neck and pediatrics. Written exams of ESNR were used as input data, related to anatomy/embryology (30 questions), neuro-oncology (50 questions), head and neck (50 questions), and pediatrics (50 questions). All exams together, and each exam separately were introduced to the three LLMs: chatGPT 3.5, GPT4, and Google Bard. Statistical analyses included a group-wise Friedman test followed by a pair-wise Wilcoxon test with multiple comparison corrections. Overall, there was a significant difference between the 3 LLMs (p < 0.0001), with GPT4 having the highest accuracy (70%), followed by chatGPT 3.5 (54%) and Google Bard (36%). The pair-wise comparison showed significant differences between chatGPT vs GPT 4 (p < 0.0001), chatGPT vs Bard (p < 0. 0023), and GPT4 vs Bard (p < 0.0001). Analyses per subspecialty showed the highest difference between the best LLM (GPT4, 70%) versus the worst LLM (Google Bard, 24%) in the head and neck exam, while the difference was least pronounced in neuro-oncology (GPT4, 62% vs Google Bard, 48%). We observed significant differences in the performance of the three different LLMs in the running of official exams organized by ESNR. Overall GPT 4 performed best, and Google Bard performed worst. This difference varied depending on subspeciality and was most pronounced in head and neck subspeciality.

Diagnostic and interventional neuroradiology training in the UK: a national trainee survey.

AIM: Training structure in neuroradiology can be variable, nationally and internationally. Globally, there is a trend towards standardised training pathways, curricula and targeted competencies. Currently, there is limited understanding of the structure of neuroradiology training in the UK. This survey aims to: [1] identify different contemporary models of neuroradiology training in the UK, [2] compare UK trainees' commitments against national and international standards, and [3] understand whether career expectations match the predicted future demands of neuroradiologists. MATERIALS AND METHODS: A survey was developed after consultation with BSNR and UKNG representatives. The eligibility criteria included current neuroradiology trainees in the UK with at least 3 months of experience or had recently completed neuroradiology training, but less than 18 months had elapsed since achieving a certificate of completion of training. RESULTS: A total of 50 trainees responded to the survey; 26 (52%) diagnostic neuroradiologists (DNRs) and 24 (48%) interventional neuroradiologists (INRs) with an overall mean age of 33 years. The mean duration of training at the time of survey was 18 months. The survey details trainee demographics, experience at work, research and teaching commitments and future goals. CONCLUSION: Most respondents are satisfied with their training and 90% want to remain in the UK after completion of training. There is room for improvement but the future of training and working in neuroradiology seems promising internationally, with ever-evolving techniques and developments. ADVANCES IN KNOWLEDGE: Advances in knowledge: This study evaluates neuroradiology training in the UK to enhance the training of future neuroradiologists, and safeguard the future of the speciality.

Early intracranial haemorrhages in acute promyelocytic leukaemia: analysis of neuroradiological and clinico-biological parameters.

Acute promyelocytic leukaemia (APL) represents a modern success of precision medicine. However, fatalities occurring within the first 30 days of induction treatment, in particular intracranial haemorrhage (ICH), remain the main causes of death. We studied the clinico-biological characteristics of 13 patients with APL who experienced ICH. Compared to 85 patients without this complication, patients with ICH were older and more frequently had high-risk APL. Moreover, positivity for the 'swirl' sign at neuroradiological imaging and hydrocephalus were predictors of a fatal outcome, together with lower fibrinogen, prolonged international normalized ratio (INR) and higher lactate dehydrogenase levels.

One hundred years of neuroradiology in Acta Radiologica: a Swedish perspective.

This year, Acta Radiologica marks its 100th anniversary. In this article, the authors present some of the highlights from the close collaboration between Swedish neuroradiology and Acta Radiologica over the last 100 years.

Adoption of the Transradial Approach for Neurointerventions: A National Survey of Current Practitioners.

OBJECTIVES: The transradial approach (TRA) is technically feasible for both diagnostic and therapeutic neurointerventions. It improves patient comfort and is not associated with increased procedural complications when compared to the transfemoral approach (TFA). To date, no studies have looked at barriers to adoption of TRA in the neurointerventionalist community. This study aims to obtain neurointerventionalist perspectives on their adoption of TRA. MATERIALS AND METHODS: Online survey distributed to neurointerventionalists. RESULTS: A total of 55 neurointerventionalists, 52 of whom utilized TRA, responded to our survey. Overall, participants were not concerned about TRA's technical feasibility for diagnostic or therapeutic neurointerventions or about procedural complications. Most of our cohort adopted TRA due to its increased patient comfort and to reduce access site complications. In-institution interventionalists were strongly perceived to be the most effective method of teaching TRA when compared to other methods. Catheters and equipment issues were reported by about 30% of our cohort as a barrier to TRA adoption. CONCLUSIONS: The neurointerventionalist community largely perceives TRA to be technically feasible and was not concerned about its procedural complications. In-person institutionalists are strongly perceived to be the most effective method of teaching the approach. A significant barrier to adoption seems to be related to catheters and equipment issues.

Artificial Intelligence and Deep Learning in Neuroradiology: Exploring the New Frontier.

There have been many recently published studies exploring machine learning (ML) and deep learning applications within neuroradiology. The improvement in performance of these techniques has resulted in an ever-increasing number of commercially available tools for the neuroradiologist. In this narrative review, recent publications exploring ML in neuroradiology are assessed with a focus on several key clinical domains. In particular, major advances are reviewed in the context of: (1) intracranial hemorrhage detection, (2) stroke imaging, (3) intracranial aneurysm screening, (4) multiple sclerosis imaging, (5) neuro-oncology, (6) head and tumor imaging, and (7) spine imaging.

Standards for European training requirements in interventional neuroradiology : Guidelines by the Division of Neuroradiology/Section of Radiology European Union of Medical Specialists (UEMS), in cooperation with the Division of Interventional Radiology/UEMS, the European Society of Neuroradiology (ESNR), and the European Society of Minimally Invasive Neurological Therapy (ESMINT).

This document sets out standards for training in Interventional Neuroradiology (INR) in Europe. These standards have been developed by a working group of the European Society of Neuroradiology (ESNR) and the European Society of Minimally Invasive Neurological Therapy (ESMINT) on the initiative and under the umbrella of the Division of Neuroradiology/Section of Radiology of the European Union of Medical Specialists (UEMS).

COVID-19: A primer for Neuroradiologists.

The potential for central nervous system (CNS) involvement in coronavirus disease 2019 (COVID-19) is a matter of grave concern and there is a relevant body of evidence in the basic sciences to support this possibility. A neuroradiologist should be aware of the potential mechanisms involved in the neuropathogenesis of this virus, as we begin to see cases with abnormal brain scans emerging from all parts of the world.

Puumala hantavirus: an imaging review.

Puumala virus (PUUV) is the most common hantavirus in Europe. It is known to cause nephropathia epidemica, which is considered a mild type of hemorrhagic fever with renal syndrome. However, it does not only involve the kidneys and is rarely accompanied by symptomatic hemorrhage. We review the imaging abnormalities caused by PUUV infection, from head to pelvis, emphasizing the broad spectrum of possible findings and bringing further support to a previously suggested denomination "Hantavirus disease" that would encompass all clinical manifestations. Although non-specific, knowledge of radiological appearances is useful to support clinically suspected PUUV infection, before confirmation by serology.

An experience-based review of HIFU in functional interventional neuroradiology: transcranial MRgFUS thalamotomy for treatment of tremor.

Tremor is a common and very disabling symptom in patients with essential tremor and Parkinson's disease. In the recent years, transcranial ablation of thalamic nuclei using magnetic resonance guided high-intensity focused ultrasound has emerged as a minimally invasive treatment for tremor. The aim of this review is to discuss, in the light of our single-center experience, the technique, current applications, results, and future perspectives of this novel technology.

Imaging findings in hypophysitis: a review.

Hypophysitis (HP) is a rare acute or chronic inflammatory condition of the pituitary gland. The greatest challenge in the management of HP is establishing a diagnosis through clinical criteria and non-invasive methods and predicting the patients' clinical outcome. The aim of this review is to describe the neuroradiological findings of this rare disease, providing some information regarding the possible differential diagnosis in order to avoid unnecessary surgery. Gadolinium-enhanced pituitary magnetic resonance imaging (MRI) is considered the neuroradiological investigation of choice. The features suggestive for HP include an enlarged triangular- or dumbbell-shaped gland with a thickened and not obviously deviated stalk, further supported by the absence of posterior pituitary bright spot on T1weighted images, particularly in patients presenting with diabetes insipidus. Contrast enhancement pattern is quite variable; dural enhancement has been reported in some cases after intravenous contrast administration. The characterization of the unusual sellar mass is not straightforward and generally results in a wide differential. HP should be primarily differentiated from pituitary adenomas (including pituitary apoplexy), from pituitary metastases, and from other sellar and parasellar tumors, e.g., craniopharyngiomas, germinomas, gliomas, lymphomas, meningiomas, pituicytomas, chordomas, teratomas, dermoids and epidermoids, Rathke's cleft cysts, and abscesses. In patients suspected for secondary forms related to systemic pathology, additional imaging is helpful in identifying other involved sites. Neuroradiologists need to know MRI appearance of this rare disease, as well as its typical symptoms and serological markers. A strict collaboration with endocrinologists and neurosurgeons is mandatory in order to reach a definitive diagnosis, allowing to promptly initiating an appropriate treatment.

Comparison between postmortem computed tomography and autopsy in the detection of traumatic head injuries.

INTRODUCTION: The aim of this study was to assess the agreement between postmortem computed tomography (PMCT) and autopsy in detecting traumatic head injuries. MATERIALS AND METHODS: Consecutive cases of death that underwent both unenhanced PMCT and conventional autopsy were collected from our institution database during a period of 3 years and reviewed retrospectively. PMCT images were reviewed for the presence of fractures (cranial vault, skull base, facial bones and atlas/axis) and intracranial hemorrhage. Kappa values were calculated to determine the agreement between PMCT and autopsy reports. RESULTS: 73 cases were included, of which 44 (60%) had head trauma. Agreement between PMCT and autopsy was almost perfect (κ = 0.95) for fractures and substantial (κ = 0.75) for intracranial hemorrhage. PMCT was superior to autopsy in detecting facial bone and upper cervical spine fractures, and intraventricular hemorrhage. However, in some cases thin extra-axial blood collections were missed on PMCT. CONCLUSIONS: The agreement between PMCT and autopsy in detecting traumatic head injuries was good. Using a combination of both techniques increases the quality of postmortem evaluation because more lesions are detected.

Presurgical differentiation between malignant haemangiopericytoma and angiomatous meningioma by a radiomics approach based on texture analysis.

PURPOSE: To assess whether a machine-learning model based on texture analysis (TA) could yield a more accurate diagnosis in differentiating malignant haemangiopericytoma (HPC) from angiomatous meningioma (AM). MATERIALS AND METHODS: Sixty-seven pathologically confirmed cases, including 24 malignant HPCs and 43 AMs between May 2013 and September 2017 were retrospectively reviewed. In each case, 498 radiomic features, including 12 clinical features and 486 texture features from MRI sequences (T2-FLAIR, DWI and enhanced T1WI), were extracted. Three neuroradiologists independently made diagnoses by vision. Four Support Vector Machine (SVM) classifiers were built, one based on clinical features and three based on texture features from three MRI sequences after feature selection. The diagnostic abilities of these classifiers and three neuroradiologists were evaluated by receiver operating characteristic (ROC) analysis. RESULTS: Malignant HPCs were found to have larger sizes, slighter degrees of peritumoural oedema compared with AMs (P<0.05), and more serpentine-like vessels. The AUC of the enhanced T1WI-based classifier was 0.90, significantly higher than that of T2-FLAIR-based or DWI-based classifiers (0.77 and 0.73). The AUC of the SVM classifier based on clinical features was 0.66, slightly but not significantly lower than the performances of 3 neuroradiologists (AUC=0.69, 0.70 and 0.73). CONCLUSION: Machine-learning models based on clinical features alone could not provide a better diagnostic performance than that of radiologists. The SVM classifier built by texture features extracted from enhanced T1WI is a promising tool to differentiate malignant HPC from AM before surgery.

Isolated neonatal bilateral vocal cord paralysis revealing a unilateral medullary defect: a case report.

BACKGROUND: Congenital bilateral vocal cord paralysis is a rare occurrence. Approximately half the cases are associated with a major comorbidity, usually neurological, neuromuscular or malformative. CASE PRESENTATION: In a male newborn, respiratory distress syndrome and stridor were observed immediately following birth. The cause was bilateral vocal cord paralysis in the adducted position. Neuroradiological investigation revealed a unilateral discontinuity between the upper pons and the right medulla oblongata. Hypoplasia of the right posterior hemiarches of C1-C2 and the right exo-occipital bone was observed, as was a small clivus. MR angiography showed the absence of the distal right vertebral artery, with hypoplasia and parietal irregularities of the proximal segments. Respiratory autonomy was not obtained despite endoscopic laser cordotomy, corticosteroid therapy and nasal continuous positive airway pressure. The infant died at the age of 4 weeks after treatment was limited to comfort care. CONCLUSIONS: A medullary lesion is an exceptional cause of congenital bilateral vocal cord paralysis. The strictly unilateral neurological and vascular defect and the absence of associated intracranial or extracranial malformation make this clinical case unique and suggest a disruptive mechanism. This case also highlights the help provided by advanced neuroimaging techniques, i.e. fibre tracking using diffusion tensor imaging, in the decision-making process.

MR Neurography of Brachial Plexus at 3.0 T with Robust Fat and Blood Suppression.

Purpose To develop and evaluate magnetic resonance (MR) neurography of the brachial plexus with robust fat and blood suppression for increased conspicuity of nerves at 3.0 T in clinically feasible acquisition times. Materials and Methods This prospective study was HIPAA compliant, with institutional review board approval and written informed consent. A low-refocusing-flip-angle three-dimensional (3D) turbo spin-echo (TSE) sequence was modified to acquire both in-phase and out-of-phase echoes, required for chemical shift (Dixon) reconstruction, in the same repetition by using partial echoes combined with modified homodyne reconstruction with phase preservation. This multiecho TSE modified Dixon (mDixon) sequence was optimized by using simulations and phantom studies and in three healthy volunteers. The sequence was tested in five healthy volunteers and was evaluated in 10 patients who had been referred for brachial plexopathy at 3.0 T. The images were evaluated against the current standard of care, images acquired with a 3D TSE short inversion time inversion recovery (STIR) sequence, qualitatively by using the Wilcoxon signed-rank test and quantitatively by using the Friedman two-way analysis of variance, with P < .05 considered to indicate a statistically significant difference. Results Multiecho TSE-mDixon involving partial-echo and homodyne reconstruction with phase preservation achieved uniform fat suppression in half the imaging time compared with multiacquisition TSE-mDixon. Compared with 3D TSE STIR, fat suppression, venous suppression, and nerve visualization were significantly improved (P < .05), while arterial suppression was better but not significantly so (P = .06), with increased apparent signal-to-noise ratio in the dorsal nerve root ganglion and C6 nerve (P < .001) with the multiecho TSE-mDixon sequence. Conclusion The multiecho 3D TSE-mDixon sequence provides robust fat and blood suppression, resulting in increased conspicuity of the nerves, in clinically feasible imaging times and can be used for MR neurography of the brachial plexus at 3.0 T. © RSNA, 2016 Online supplemental material is available for this article.

Radiology research in mainland China in the past 10 years: a survey of original articles published in Radiology and European Radiology.

PURPOSE: To evaluate the features and trends of Radiology research in Mainland China through bibliometric analysis of the original articles published in Radiology and European Radiology (ER) between 2006 and 2015. MATERIALS AND METHODS: We reviewed the original articles published in Radiology and ER between 2006 and 2015. The following information was abstracted: imaging subspecialty, imaging technique(s) used, research type, sample size, study design, statistical analysis, study results, funding declarations, international collaborations, number of authors, department and province of the first author. All variables were examined longitudinally over time. RESULTS: Radiology research in Mainland China saw a substantial increase in original research articles published, especially in the last 5 years (P < 0.001). Within Mainland China's Radiology research, neuroradiology, vascular/interventional Radiology, and abdominal Radiology were the most productive fields; MR imaging was the most used modality, and a distinct geographic provenience was observed for articles published in Radiology and ER. CONCLUSION: Radiology research in Mainland China has seen substantial growth in the past 5 years with neuroradiology, vascular/interventional Radiology, and abdominal Radiology as the most productive fields. MR imaging is the most used modality. Article provenience shows a distinct geographical pattern. KEY POINTS: • Radiology research in Mainland China saw a substantial increase. • Neuroradiology, vascular/interventional Radiology, and abdominal Radiology are the most productive fields. • MRI is the most used modality in Mainland China's Radiology research. • Guangdong, Shanghai, and Beijing are the most productive provinces.