Background: Interest in artificial intelligence (AI) and machine learning (ML) has been growing in neuroradiology, but there is limited knowledge on how this interest has manifested into research and the field’s trends, challenges, and future directions. Methods: The American Journal of Neuroradiology was queried for original research articles published since inception (Jan. 1, 1980) to Sept. 19, 2022 that contained any of the following key terms: “machine learning”, “artificial intelligence”, or “radiomics”. Articles were screened, categorized into Statistical Modelling (Type 1), AI/ML Development (Type 2), or End-user Application (Type 3) and then bibliometrically analyzed. Results: A total of 124 articles were identified with 85% being non-integration focused (Type 1 n = 41, Type 2 n = 65) and the remaining (n = 18) being Type 3. The total number of articles published grew two-fold in the last five years, with Type 2 articles mainly driving this growth. While most (66%) Type 2 articles were led by a radiologist with 55% possessing a postgraduate degree, a minority of Type 2 articles addressed bias (15%) and explainability (20%). Conclusions: The results of this study highlight areas for improvement but also strengths that stakeholders can consider when promoting the shift towards integrating practical AI/ML solutions in neuroradiology.

Background: The aim of our educational exhibit is to review the anatomy and pathology encountered and often overlooked of the excretory lacrimal apparatus from the lacrimal sac to the nasal fossa. Methods: We will provide an anatomical review of the various structures easily identifiable on CT and MRI and suggestions of the best imaging protocols to be used. Results: The lacrimal apparatus includes the various structures related to the production and flow of tears. In this educational exhibit we will focus on the excretory apparatus from the lacrimal sac to the nasal fossa. We will present various pathologies affecting the excretory lacrimal apparatus with attention to the specific features of each condition to facilitate an appropriate differential diagnosis. We will emphasize specific anatomical/imaging findings to help the diagnosis and propose a standardized reporting system for the Neuroradiologist and useful to the ENT surgeon. Conclusions: This educational exhibit offers a unique opportunity to review the anatomy and pathology of sometimes overlooked or forgotten structures which are however always included in our CT and MRI studies.

Background: Radiofrequency ablation (RFA) is a minimally-invasive procedure that has been used to treat temporal lobe epilepsy (TLE), however its long-term efficacy is unknown. We aim to characterize the long-term outcomes of patients from the original series by Parrent and Blume (1999). Methods: Consecutive patients who underwent stereotactic RFA for TLE were retrospectively reviewed. Demographics, procedural details, and seizure outcomes until last follow-up were abstracted. Seizure-freedom after initial RFA treatment was estimated with Kaplan-Meier analysis. Results: 27 patients underwent RFA from 1994 to 2002. There were 14 female (52%) patients. 24 (89%) had mesial temporal sclerosis. Mean age at time of RFA was 33.1 years (range 12-45 years). 17 (63%) patients underwent left-sided RFA. 15 (56%) patients had further interventions: 4 (15%) underwent only repeat RFA, 1 (4%) had repeat RFA and anterior temporal lobectomy (ATL), and 10 (37%) underwent subsequent ATL only. Mean follow-up was 9.0 years (range 0.5-22.7 years). At last follow-up, 16 (59%) patients were seizure-free: 5 (19%) received one RFA treatment and 11 (41%) underwent additional procedures. Conclusions: Based on the original series describing the technique, stereotactic RFA for TLE is a safe, minimally-invasive procedure. The role of stereotactic RFA in the treatment of TLE remains to be determined.

Background: Bi-directional brain interfacing (closed loop DBS) is a modern focus of neuroengineering research. Most current clinical systems are open loop, allowing one way communication from the IPG battery to the brain. Bi-directional systems allow both stimulation and recording of neural activity (local field potential, LFP). The system algorithm can measure known pathologic LFPs to guide change in stimulation. However, recording LFPs from the brain encounters electrical artifact from the heart. Reducing artifact is imperative to accurate measurement of neural activity. Artifact will cause the bi-directional system to miscalculate stimulation parameters. This project evaluated reduction of artifact by moving the IPG further away from the heart in a device implanted into the skull. Methods: LFP data from ongoing clinical trials was collected and analysed for artifact using open source code. Anatomic targets include STN, PPN, CMT, and PAG. Results: Cardiac artifact is reduced in skull mounted DBS as shown by power spectral density of LFPs in each region. Conclusions: This project shows the importance of surgical placement of DBS sensing devices to reduce cardiac artifact in bi-directional brain interfacing. This has important engineering and surgical design implications for safety and performance as the field of closed loop DBS transitions from research to clinical settings.

Background: Success of deep brain stimulation (DBS) in Parkinson’s disease (PD) relies on time-consuming trial-and-error testing of stimulation settings. Here, we prospectively compared an fMRI-based stimulation optimization algorithm with >1 year of standard-of-care (SoC) programming in a double-blind, crossover, non-inferiority trial. Methods: Twenty-seven PD-DBS patients were prospectively enrolled for fMRI using a 30-sec DBS-ON/OFF cycling paradigm. Optimal settings were identified using our published classification algorithm. Subjects then underwent >1 year of SoC programming. Clinical improvement was assessed, after an overnight medication washout period, under SoC and fMRI-determined stimulation conditions. A predefined non-inferiority margin was -5 points on the Unified Parkinson’s Disease Rating Scale (UPDRSIII). Results: UPDRSIII improved from 45.3 (SD=14.6) at baseline to 24.9 (SD=10.9) and 24.1 (SD=10.9) during SoC and fMRI-determined stimulation, respectively. The mean difference in scores was 0.8 (SD=8.5; 95% CI -4.5 to 6.2). The non-inferiority margin was not contained within the 95% confidence interval, establishing non-inferiority (p=0.013). Conclusions: This is the first prospective evaluation of an algorithm able to suggest stimulation parameters solely from the fMRI response to stimulation. It suggests equivalent outcomes may be achieved in 3 hours of fMRI scanning immediately after surgery compared to SoC requiring 6 or more in-person clinic visits throughout >1 year.

Background: Stereotactic laser amygdalohippocampotomy (SLAH) has recently been shown to be comparable to traditional temporal lobectomy procedures. The ideal extent and volume of laser ablations remains an area of investigation Methods: 65 patients treated with SLAH for MTS were considered in this retrospective study. Manual segmentations of ablations were created using post-procedure T1-MRI scans. Ablations were assessed in relation to whether they crossed the coronal plane of the superior lateral mesencephalic sulcus (LMS), the extent to which ablation crossed this landmark, and extent of ablation of the uncus. Analysis of was done with binary categorization of 12-month Engel classification score. Results: Distance of ablation posterior to the coronal plane of the LMS was not associated with better surgical outcome (Engel class 1: 6.32 ± 4.16 mm; Engel class 2-4: 7.93 ± 3.75mm; (p = 0.099)). Ratio of ablations extending posterior to the LMS was 0.82 (SD = .39) in Engel 1 patients, and 0.90 (SD = 0.3) in Engel 2-4 patients; (p = 0.370). Volume of ablation showed little correlation with outcome (Engel class 1: 6064 ± 2128 mm3; Engel class 2-4: 5828 ± 3031 mm3; (p=0.239)). Ablation of the uncus showed a strong association with better surgical outcome (Engel class 1: 0.71(SD = 0.31); Engel 2-4: 0.37 (SD = 0.36); p <0.001). Conclusions: Contrary to current practice, extension of ablation posterior to the LMS did not demonstrate improved outcome.

Background: Chordomas are rare, malignant bone tumors that present significant challenges in management and treatment due to their complex anatomical locations and propensity for recurrence. Advancements in artificial intelligence (AI) and machine learning (ML) show promise in improving chordoma management. Methods: A comprehensive literature search was conducted following PRISMA guidelines across multiple databases, including MEDLINE, Cochrane, Embase, Scopus, and Web of Science. The search targeted articles related to AI and ML applications in clinical tasks associated with chordoma management. The selection process involved systematic screening, data extraction, and assessment of inter-rater variability. Results: The search yielded 1,006 records, with 18 included for analysis. Convolutional neural networks (CNNs) excelled in tumor volume estimation, with the state-of-the-art model achieving a Dice similarity score of 74.2%, sensitivity of 79.4%, and positive predictive value of 74.3%. Clustering algorithms were effective in prognostic evaluations. Bayesian models and logistic regression demonstrated robustness in diagnostics. Support vector machines (SVMs) were noted for their diagnostic precision. Conclusions: AI and ML algorithms, particularly CNNs, clustering algorithms, Bayesian models, logistic regression, and SVMs, show promise in improving chordoma management through enhanced imaging, diagnostics, and prognostics. Future research should focus on larger, externally validated datasets and explore underutilized techniques like multi-modal data integration.

Background: Primary central nervous system lymphoma (PCNSL) is highly sensitive to corticosteroid induced cell arrest, apoptosis and shrinkage. However, the precise impact of preoperative corticosteroid on accuracy of PCNSL diagnosis using tissue obtained from open or stereotactic biopsies remains debated. Methods: We conducted a systematic review and meta-analysis to determine the effect of preoperative corticosteroids on non-diagnostic biopsy rates for PCNSL in immunocompetent adults. Subgroup analyses explored whether non-diagnostic rates varied based on biopsy type. Results: Nineteen studies, comprising 1226 patients (55% male; mean age: 60.3 years), of which 679 (55.4%) received corticosteroids prior to biopsy were included. Overall, patients pretreated with corticosteroids were two times more likely to have a non-diagnostic biopsy compared to patients that were corticosteroid-naïve prior to biopsy (RR = 2.1 [95% CI: 1.1-4.1]). In the subgroup analysis limited to stereotactic biopsies, patient pretreated with corticosteroids were three times more likely to have a non-diagnostic biopsy (RR = 3.0 [95% CI: 1.2-7.5]). Whereas, in the open biopsy subgroup, there was no significant difference in non-diagnostic rates. Conclusions: Corticosteroids should be withheld, if clinically safe, prior to stereotactic biopsies in cases of suspected PCNSL. If corticosteroids are administered preoperatively, an open biopsy should be considered instead of stereotactic biopsy.

Background: Surgical resection of vestibular schwannoma (VS) is often curative if gross total resection is achieved, however, it is a delicate procedure with high risk to the facial nerve. With retrosigmoid approach for resection, the head is positioned to maximize lateral head rotation and neck flexion in order to optimize the surgical field. However, this may inadvertently occlude cerebral venous drainage, elevating intracranial pressure (ICP) and increasing intraoperative bleeding. Methods: Here, we review relevant literature regarding the effects of head rotation and neck flexion on internal jugular vein (IJV) occlusion and ICP, and highlight the notion that head rotation and flexion may occlude the ipsilateral IJV, increasing ICP. Subsequently, we propose a novel technique using continuous, real-time monitoring of jugular bulb pressure (JBP) to detect obstructions in jugular venous flow and guide optimal head positioning prior to VS resection. Results: As proof of concept, we present a case in which JBP monitoring was employed to optimize head positioning prior to a VS resection, which shows a significant reduction in JBP compared to traditional positioning. Conclusions: This innovative approach offers promise in enhancing the safety and efficacy of intracranial surgery for VS and potentially other neurosurgical procedures.

Background: Intracranial epidermoid cysts (IEC) are benign congenital intracranial lesions that rarely undergo malignant transformation. We report a case of IEC evolving into squamous cell carcinoma (SCC) 1-year post-resection. Further, we conducted a systematic review on cases of early malignant transformations of IECs. Methods: MEDLINE, EMBASE, and Scopus were searched from inception until December 2023 for studies reporting malignant transformations of IECs within 2 years of diagnosis. Results: A 48-year-old female underwent surgical resection of a cerebellopontine angle (CPA) IEC in May 2022. She re-presented in July 2023 with headaches, nausea, vomiting, right facial weakness, and rapid cyst progression. Repeat surgical resection revealed a high-grade SCC. Our systematic review identified 19 (10 females, 9 males) additional IEC cases undergoing malignant transformation within 2 years. The mean age at presentation was 57.6 years, most common location was CPA (n=13, 68.4%) and mean time between IEC to malignant transformation was 10.6 months. Eighteen (94.7%) cases transformed to SCC, of which 2 had leptomeningeal carcinomatosis, and 1 transformed to glioblastoma. Conclusions: While malignant transformations of IECs are rare, regular postoperative follow-up is crucial for early malignancy detection and treatment initiation. Further study is warranted to evaluate factors contributing to accelerated malignant progression of IECs.

Background: High grade gliomas (HGG) are incurable, aggressive brain malignances that carry poor prognoses. Significant scientific advances have uncovered many of the features of these diseases; however, it remains unclear how mutations and transcriptional changes drive glioma growth and progression. Methods: We used a Nestin-Cre mouse model in combination with an extrinsic chemical mutagen (N-ethyl-N-nitrosurea, ENU), to model HGG. We combined our mouse model with live animal in vivo magnetic resonance imaging to track tumor growth over time, and sample discrete lesions during premalignant, early stage tumor, and end stage tumor phases. Results: We show that the somatic mutations, copy number changes, and transcriptional profiles of tumors vary depending on the stage of growth, and that the Raf/Ras pathway is key for tumor growth with a recurring Braf mutation occurring in early stage lesions. Gene set enrichment analysis (GSEA) shows that end stage tumors have increased immunogenic/inflammatory activity, and increased signaling through Raf/Ras. Conclusions: The combination of genetic and nongenetic insults results in activating mutations in early lesions, which continue to be biologically active and underlie key differences between early and end stage tumors. Overall, this work sheds light on important differences between early and late stage tumors.

Background: The “weekend effect” is the finding that patients presenting for medical care outside of regular working hours tend to have worse outcomes. There is a paucity of literature in the neuro-oncology space exploring this effect. We investigated the extent of resection and complication rates in patients undergoing after-hours high grade glioma resection. Methods: A retrospective review was conducted on patients with high-grade gliomas requiring emergent surgery between January 2021 to March 2023. After-hours was defined as surgical resection on the weekend and/or evening. These patients were matched to patients undergoing resection during regular working hours. Results: A total of 38 patients were included in this study (19 after-hours, 19 regular hours). There was no significant difference in age, sex, tumor grade, and tumor size between the two groups (all p>0.05). There was no significant difference in the extent of resection between the groups (p=0.7442). There was no significant difference in complications rates, reoperation rates, and death at 6 months (all p>0.05). Estimated blood loss was significantly higher in the regular hours group (p=0.0278). There was no significant difference in the total operative time (p=0.0643) and length of stay (p=0.0601). Conclusions: After-hours high grade glioma surgery is not associated with increased morbidity or mortality.

Background: Meningiomas are common brain neoplasms that can significantly influence health-related quality of life (HRQOL), yet the factors influencing HRQOL in adult patients remain unclear. We aimed to bridge this knowledge gap by determining these key factors. Methods: We conducted a systematic review, searching EMBASE, MEDLINE, CINAHL, Scopus, and PsycINFO up to March 2023. We included original, peer-reviewed studies focusing on adult patients (>18 years) with current or past meningioma at any stage of treatment that measured HRQOL or its proxies in relation to tumour-, treatment-, and patient-related factors. Two independent reviewers screened abstracts and full texts, selecting studies with acceptable risk of bias for data extraction and narrative synthesis. Results: Of N=2942 identified studies, N=30 were included. Key factors found to influence HRQOL in adult meningioma patients include surgery, radiotherapy, neurological function, functional status, comorbidities, sleep quality, psychological impairment, age, and employment. Factors related to tumour characteristics yielded inconsistent findings. Heterogeneity and inconsistencies in HRQOL measurement across studies hindered definitive conclusions about the impact of factors on HRQOL. Conclusions: Our review emphasizes the need for standardized, disease-specific HRQOL assessments in meningioma patients. More consistent, large-scale, prospective research is essential to comprehensively understand and improve HRQOL, and thereby ensure tailored care for this population.

Background: For treatment of high-grade gliomas (HGGs), subtotal resection (STR) may be preferred to minimize injury to eloquent areas. We aimed to characterize neurologic deficits developed in STR patients within the first month post-operatively and to establish a potential threshold for a safe volume of residual tumor to avoid neurological worsening. Methods: This is a single institution retrospective chart review, with 146 charts reviewed and 78 patients deemed eligible. Preoperative deficits and postoperative neurological deficits presenting prior to 1 month after surgery were captured. Imaging features such as tumour volume, edema, and other pertinent imaging characteristics were collected from preoperative and postoperative imaging. Results: Most patients that developed a postoperative deficit presented with motor deficits (55.1%), while only 1.3% of patients developed new or worsening tremor after surgery. On average, in patients with a new deficit, 26.5% of tumor was resected, and all patients had more than 19% of residual tumor. Conclusions: Postoperative neurologic deficits may develop after a subtotal resection when an average of 73.5% of tumor remains. The proposed threshold for tumor resection is greater than 26.5% to minimize the potential of neurologic worsening 1 month postoperatively.

Background: Reporting extent of resection (EOR) in pituitary adenoma (PA) surgery via endoscopic endonasal approaches (EEA) is not standardized. The use of 3-dimensional volumetric analysis is proposed for measurement of tumor volumes and EOR. Their relationship with visual outcomes is explored. Methods: A retrospective analysis of PA patients presenting with visual disturbances and treated surgically via EEA by a single surgeon between 2006 and 2021. The main outcome was visual function at 12 months post-operatively. Results: 142 patients were included. Majority were male, with mean age of 57.1 years. Most (58.2%) presented with bitemporal hemianopsia. The mean tumor size was 11.3 cm3. The mean EOR was 84.5% (range 21.5-99.8%), with a mean post-operative tumor volume of 1.9 cm3. Visual function improved in 92.2%. Re-resection for visual deterioration was performed in 5.7% of patients, (mean time 2.4 years). No clinical, pathologic, or imaging factors were significantly associated with visual outcome. A significant association was found between EOR and re-resection (mean EOR 66.7% vs 85.6%, p=0.002). Conclusions: For patients with PA presenting with visual deficits, treatment with EEA led to improvement in visual function in the majority of patients, without the need for gross total resection. EOR was significantly associated with the need for re-resection.

Background: Pituicytoma is a rare, low grade tumour typically of the sellar region. Here we present a pituicytoma located in the third ventricle. Methods: 56 year old female presented with amenorrhea, hyperprolactinemia, and progressive bitemporal hemianopsia. MRI revealed a suprasellar mass located within the third ventricle and appearing separate from the pituitary. A supraciliary and translamina terminalis surgical approach to tumour resection was completed without complication. Post-operatively, she developed transient DI which resolved by post-opertaive day 3 and she was discharged hoem without any neurological deficits. Pathology revealed pituicytoma, WHO grade I. Results: Pituicytomas are rare tumours arising from neuroepithelial cells of the pituitary. The majority of cases are pure sellar or sellar with suprasellar extension, or at least have some connection to the pituitary. In many cases, imaging findings are synonymous to pituitary adenomas. We present a unique case in which the tumour was suprasellar but appeared separate from the pituitary. Surgical intervention is the most highly predictive factor of recurrence, as gross totoal reseciton can be curable. Conclusions: Here we present a unique location of pituictyoma. Due to the exceedingly rare nature of pituicytoma, unique presentations and management help to provide better understanding of the breadth of this disease presentation.

Background: Synthetic data has garnered heightened attention in contemporary research due to confidentiality barriers and its capacity to simulate variables challenging to obtain. This study aimed to evaluate the reliability and validity of synthetic data in the context of neuro-oncology research, comparing findings from two published studies with results from synthetic datasets. Methods: Two published neuro-oncology studies focusing on prognostic factors such as serum albumin and systemic inflammation scores were selected, and their methodologies were replicated using MDClone Platform to generate five synthetic datasets for each. We used Chi-Square test to assess inter-variability between synthetic datasets. Survival outcomes were evaluated using Kaplan-Meier and t-test was used to determine statistical significance. Results: Findings from synthetic data consistently matched outcomes from both original articles, with serum albumin and systemc inflammation scores correlating with survival prognosis in glioblastoma and metastasis patients (p<0.05) Reported findings, demographic trends and survival outcomes showed significant similarity (P > 0.05) with synthetic datasets. Conclusions: Synthetic data consistently reproduced the statistical attributes of real patient data. Integrating synthetic data into clinical research offers excellent potential for providing accurate predictive insights without compromising patient privacy. In neuro-oncology, where patient follow-up pose challenges, the adoption of synthetic datasets can be transformative.

Background: Brain metastases indicate an advanced tumour stage for many cancers. We sought to investigate the incidence change of tissue-sampled brain metastases and its relation to staging challenges during the COVID-19 pandemic in Newfoundland and Labrador. Methods: We reviewed all brain metastasis cases from 2015-2022 requiring first-time tissue sampling according to pathology reports from the St. John’s Health Sciences Centre. Incidence rates were calculated using yearly population data by regional health authorities and standardized using the 2011 Canadian standard population. Results: We included 173 cases. The average annual age-standardized incidence rate of brain metastases requiring tissue sampling per 100,000 increased from 2.5 (95% CI: 2.0-3.1) pre-COVID-19 to 4.1 (95% CI: 3.3-5.0) post-COVID-19. Brain metastases from lung primaries accounted for 69% of this increase. While incidence declined to near-baseline in the Eastern provincial population by 2022 (3.3; 95% CI: 1.5-5.1), incidence rose into 2022 in the Western population (8.6; 95% CI: 3.9-13.2). Conclusions: These data suggest a delayed presentation of malignancies during the COVID-19 pandemic and underscore the importance of prioritized staging during times of strain on healthcare systems. Regional, temporal trends suggest regions distant from tertiary care centres could face challenges in resolving cases with delayed presentation post-COVID-19.

Background: Elevated BMI has been proposed as a risk factor for the development of meningioma. The relationship between body mass index (BMI) and disease control in high-grade meningioma has not yet been examined. A retrospective cohort study was performed to assess the relationship between high-grade meningioma recurrence and BMI. Methods: This is a retrospective cohort study of patients with Grade 2 or Grade 3 meningioma at a single tertiary care center between 2008 and 2017. We collected clinical data including age, sex, BMI, location, Simpson grade, brain invasion, and radiation treatments. Disease control was monitored on followup MRI scans. We stratified patients by BMI greater than or less than 25. Results: A total of 45 patients were included. Recurrence was observed in 15 patients (33.3%). There were 32 (71.1%) patients with BMI > 25, and 13 (28.9%) patients with normal BMI. Patients with elevated BMI had higher risk of recurrence (p=0.04). Multivariate analysis identified BMI as an independent predictor of recurrence. Conclusions: Our results suggest that overweight patients with a Grade 2 or Grade 3 meningioma are at higher risk of recurrence than patients with normal BMI. The explanation for this association unknown. Further research is suggested to confirm and better characterize this association.

Background: Considering regional and temporal trends, we sought to explore the incidence of primary malignant brain tumours in Newfoundland and Labrador. Methods: We reviewed all primary, malignant brain tumour cases from 2015-2022 confirmed by St. John’s Health Sciences Centre pathology reports. Incidence rates were standardized using the 2011 Canadian standard population. Results: We included 362 cases. The average annual age-standardized incidence rate of primary, malignant brain tumours per 100,000 was 7.0 (95% CI: 6.3-7.7), lower than the national average (7.93; 95% CI: 7.78-8.08). The incidence of glioblastoma (5.1; 95% CI: 4.5-5.7) was significantly higher than the national average (4.05; 95% CI: 3.95-4.16). Temporal trends revealed that oligodendroglioma incidence spiked from 0.5 (95% CI: 0.2-0.7) in 2015-2019 to 1.5 (95% CI: 0.4-2.6) in 2020 before returning to baseline in 2022. Regional trends indicated a lower incidence of malignant tumours in Labrador-Grenfell (5.1; 95% CI: 2.5-7.6), compared to 6.9 (95% CI: 6.2-7.6) averaged elsewhere. Conclusions: Higher rates of glioblastoma in Newfoundland and Labrador could have a genetic or multi-factorial cause. The increased occurrence of oligodendroglioma during the COVID-19 pandemic necessitates broader investigation, potentially linked to delays in patient care during this period. Regional trends could suggest less access to care in rural populations and underestimated incidence.