The Impact of MRI-Based Advanced Neuroimaging on Neurooncologists' Clinical Decision-Making in Patients With Posttreatment High-Grade Glioma: A Prospective Survey-Based Study.

BACKGROUND. Advanced MRI-based neuroimaging techniques, such as perfusion and spectroscopy, have been increasingly incorporated into routine follow-up protocols in patients treated for high-grade glioma (HGG), to help differentiate tumor progression from treatment effect. However, these techniques' influence on clinical management remains poorly understood. OBJECTIVE. The purpose of this article was to evaluate the impact of MRI-based advanced neuroimaging on clinical decision-making in patients with HGG after treatment. METHODS. This prospective study, performed at a comprehensive cancer center from March 1, 2017, to October 31, 2020, included adult patients treated by chemoradiation for WHO grade 4 diffuse glioma who underwent MRI-based advanced neuroimaging (comprising multiple perfusion imaging sequences and spectroscopy) to further evaluate findings on conventional MRI equivocal for tumor progression versus treatment effect. The ordering neurooncologists completed surveys before and after each advanced neuroimaging session. The percent of episodes of care with a change between the intended and actual management plan on the surveys conducted before and after advanced neuroimaging, respectively, was computed and compared with a published percent using the Wald test for independent samples proportions. RESULTS. The study included 63 patients (mean age, 54.6 ± 12.9 [SD] years; 36 women, 27 men) who underwent 70 advanced neuroimaging sessions. Ordering neurooncologists' intended and actual management plans on the surveys completed before and after advanced neuroimaging, respectively, differed in 44% (31/70; 95% CI: 33-56%) of episodes, which differed from the published frequency of 8.5% (5/59) (p < .001). These management plan changes included selection of a different plan for six of eight episodes with an intended plan to enroll patients in a clinical trial, 12 of 19 episodes with an intended plan to change chemotherapeutic agents, four of eight episodes with an intended plan of surgical intervention, and one of two episodes with an intended plan of reirradiation. The ordering neurooncologists found advanced neuroimaging to be helpful in 93% (65/70; 95% CI: 87-99%) of episodes. CONCLUSION. Neurooncologists' management plans changed in a substantial fraction of adult patients with HGG who underwent advanced neuroimaging to further evaluate conventional MRI findings equivocal for tumor progression versus treatment effect. CLINICAL IMPACT. The findings support incorporation of advanced neuroimaging into HGG posttreatment monitoring protocols.

Machine learning and radiomics for ventricular tachyarrhythmia prediction in hypertrophic cardiomyopathy: insights from an MRI-based analysis.

BACKGROUND: Myocardial fibrosis is often detected in patients with hypertrophic cardiomyopathy (HCM), which causes left ventricular (LV) dysfunction and tachyarrhythmias. PURPOSE: To evaluate the potential value of a machine learning (ML) approach that uses radiomic features from late gadolinium enhancement (LGE) and cine images for the prediction of ventricular tachyarrhythmia (VT) in patients with HCM. MATERIAL AND METHODS: Hyperenhancing areas of LV myocardium on LGE images were manually segmented, and the segmentation was propagated to corresponding areas on cine images. Radiomic features were extracted using the PyRadiomics library. The least absolute shrinkage and selection operator (LASSO) method was employed for radiomic feature selection. Our model development employed the TabPFN algorithm, an adapted Prior-Data Fitted Network design. Model performance was evaluated graphically and numerically over five-repeat fivefold cross-validation. SHapley Additive exPlanations (SHAP) were employed to determine the relative importance of selected radiomic features. RESULTS: Our cohort consisted of 60 patients with HCM (73.3% male; median age = 51.5 years), among whom 17 had documented VT during the follow-up. A total of 1612 radiomic features were extracted for each patient. The LASSO algorithm led to a final selection of 18 radiomic features. The model achieved a mean area under the receiver operating characteristic curve of 0.877, demonstrating good discrimination, and a mean Brier score of 0.119, demonstrating good calibration. CONCLUSION: Radiomics-based ML models are promising for predicting VT in patients with HCM during the follow-up period. Developing predictive models as clinically useful decision-making tools may significantly improve risk assessment and prognosis.

Trends in stroke-related journals: Examination of publication patterns using topic modeling.

OBJECTIVES: This study aims to demonstrate the capacity of natural language processing and topic modeling to manage and interpret the vast quantities of scholarly publications in the landscape of stroke research. These tools can expedite the literature review process, reveal hidden themes, and track rising research areas. MATERIALS AND METHODS: Our study involved reviewing and analyzing articles published in five prestigious stroke journals, namely Stroke, International Journal of Stroke, European Stroke Journal, Translational Stroke Research, and Journal of Stroke and Cerebrovascular Diseases. The team extracted document titles, abstracts, publication years, and citation counts from the Scopus database. BERTopic was chosen as the topic modeling technique. Using linear regression models, current stroke research trends were identified. Python 3.1 was used to analyze and visualize data. RESULTS: Out of the 35,779 documents collected, 26,732 were classified into 30 categories and used for analysis. "Animal Models," "Rehabilitation," and "Reperfusion Therapy" were identified as the three most prevalent topics. Linear regression models identified "Emboli," "Medullary and Cerebellar Infarcts," and "Glucose Metabolism" as trending topics, whereas "Cerebral Venous Thrombosis," "Statins," and "Intracerebral Hemorrhage" demonstrated a weaker trend. CONCLUSIONS: The methodology can assist researchers, funders, and publishers by documenting the evolution and specialization of topics. The findings illustrate the significance of animal models, the expansion of rehabilitation research, and the centrality of reperfusion therapy. Limitations include a five-journal cap and a reliance on high-quality metadata.

Correlating volumetric and linear measurements of brain metastases on MRI scans using intelligent automation software: a preliminary study.

PURPOSE: The Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) working group proposed a guide for treatment responses for BMs by utilizing the longest diameter; however, despite recognizing that many patients with BMs have sub-centimeter lesions, the group referred to these lesions as unmeasurable due to issues with repeatability and interpretation. In light of RANO-BM recommendations, we aimed to correlate linear and volumetric measurements in sub-centimeter BMs on contrast-enhanced MRI using intelligent automation software. METHODS: In this retrospective study, patients with BMs scanned with MRI between January 1, 2018, and December 31, 2021, were screened. Inclusion criteria were: (1) at least one sub-centimeter BM with an integer millimeter-longest diameter was noted in the MRI report; (2) patients were a minimum of 18 years of age; (3) patients with available pre-treatment three-dimensional T1-weighted spoiled gradient-echo MRI scan. The screening was terminated when there were 20 lesions in each group. Lesion volumes were measured with the help of intelligent automation software Jazz (AI Medical, Zollikon, Switzerland) by two readers. The Kruskal-Wallis test was used to compare volumetric differences. RESULTS: Our study included 180 patients. The agreement for volumetric measurements was excellent between the two readers. The volumes of the following groups were not significantly different: 1-2 mm, 1-3 mm, 1-4 mm, 2-3 mm, 2-4 mm, 3-4 mm, 3-5 mm, 4-5 mm, 5-6 mm, 5-7 mm, 6-7 mm, 6-8 mm, 6-9 mm, 7-8 mm, 7-9 mm, 8-9 mm. CONCLUSION: Our findings indicate that the largest diameter of a lesion may not accurately represent its volume. Additional research is required to determine which method is superior for measuring radiologic response to therapy and which parameter correlates best with clinical improvement or deterioration.

Radiomics-based machine learning models in STEMI: a promising tool for the prediction of major adverse cardiac events.

OBJECTIVE: To evaluate the potential value of the machine learning (ML) models using radiomic features of late gadolinium enhancement (LGE) and cine images on magnetic resonance imaging (MRI) along with relevant clinical information and conventional MRI parameters for the prediction of major adverse cardiac events (MACE) in ST-segment elevation myocardial infarction (STEMI) patients. METHODS: This retrospective study included 60 patients with the first STEMI. MACE consisted of new-onset congestive heart failure, ventricular arrhythmia, and cardiac death. Radiomic features were extracted from cine and LGE images. Inter-class correlation coefficients (ICCs) were calculated to assess inter-observer reproducibility. LASSO (least absolute shrinkage and selection operator) method was used for radiomic feature selection. Seven separate models using a different combination of the available information were investigated. Classifications with repeat random sampling were done using adaptive boosting, k-nearest neighbor, naive Bayes, neural network, random forest, stochastic gradient descent, and support vector machine algorithms. RESULTS: Of the 1748 extracted radiomic features, 1393 showed good inter-observer agreement. With LASSO, 25 features were selected. Among the ML algorithms, the neural network showed the highest predictive performance on average (area under the curve (AUC) 0.822 ± 0.181). Of the best-calculated model, the one using clinical parameters, CMRI parameters, and selected radiomic features (model 7), the diagnostic performance was as follows: 0.965 AUC, 0.894 classification accuracy, 0.906 sensitivity, 0.883 specificity, 0.875 positive predictive value (PPV), and 0.912 negative predictive value (NPV). CONCLUSION: The radiomics-based ML models incorporating clinical and conventional MRI parameters are promising for predicting MACE occurrence in STEMI patients in the follow-up period. KEY POINTS: • Acute coronary occlusion results in variable changes at the cellular level ranging from myocyte swelling to myonecrosis depending on the duration of the ischemia and the metabolic state of the heart, which causes subtle heterogeneous signal changes that are imperceptible to the human eye with cardiac MRI. • Radiomics-based machine learning analysis of cardiac MR images is promising for risk prediction. • Combining MRI-derived parameters and clinical variables increases the accuracy of predictive models.

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.

PURPOSE: 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). METHODS: 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. RESULTS: 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 6.0 mL with a wide 95% confidence interval (CI) of - 68.32 to 80.22, and for midline shift, the average difference was 1.4 mm with a 95% CI of - 3.4 to 6.2. CONCLUSION: 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.

Radiomics-based machine learning models for prediction of medulloblastoma subgroups: a systematic review and meta-analysis of the diagnostic test performance.

BACKGROUND: Medulloblastomas are a major cause of cancer-related mortality in the pediatric population. Four molecular groups have been identified, and these molecular groups drive risk stratification, prognostic modeling, and the development of novel treatment modalities. It has been demonstrated that radiomics-based machine learning (ML) models are effective at predicting the diagnosis, molecular class, and grades of CNS tumors. PURPOSE: To assess radiomics-based ML models' diagnostic performance in predicting medulloblastoma subgroups and the methodological quality of the studies. MATERIAL AND METHODS: A comprehensive literature search was performed on PubMed; the last search was conducted on 1 May 2022. Studies that predicted all four medulloblastoma subgroups in patients with histopathologically confirmed medulloblastoma and reporting area under the curve (AUC) values were included in the study. The quality assessments were conducted according to the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) and Checklist for Artificial Intelligence in Medical Imaging (CLAIM). A meta-analysis of radiomics-based ML studies' diagnostic performance for the preoperative evaluation of medulloblastoma subgrouping was performed. RESULTS: Five studies were included in this meta-analysis. Regarding patient selection, two studies indicated an unclear risk of bias according to the QUADAS-2. The five studies had an average CLAIM score and compliance score of 23.2 and 0.57, respectively. The meta-analysis showed pooled AUCs of 0.88, 0.82, 0.83, and 0.88 for WNT, SHH, group 3, and group 4 for classification, respectively. CONCLUSION: Radiomics-based ML studies have good classification performance in predicting medulloblastoma subgroups, with AUCs >0.80 in every subgroup. To be applied to clinical practice, they need methodological quality improvement and stability.

Final-year medical student mental wellness during preparation for the examination for specialty in Turkey: a cross-sectional survey study.

BACKGROUND: In Turkey, most final-year medical students prepare for the Examination for Specialty in Medicine in a high-stress environment. To the best of our knowledge, this is the first study on final-year medical student general psychological distress during preparation for the Examination for Specialty in Turkey. We aim to evaluate psychological distress and understand the variables associated with depression, anxiety, and stress levels among final-year medical students preparing for the Examination for Specialty. METHODS: A self-reporting, anonymous, cross-sectional survey with 21 items consisting of demographic variables, custom variables directed for this study, and the DASS-21 was utilized. Survey results were expounded based on univariate analysis and multivariate linear regression analysis. RESULTS: Our study revealed four variables associated with impaired mental wellness among final-year medical students during preparation for the examination for Specialty: attendance to preparatory courses, duration of preparation, consideration of quitting studying, and psychiatric drug usage/ongoing psychotherapy. DISCUSSION: Considering that physician mental wellness is one of the most crucial determinants of healthcare quality, impaired mental wellness among future physicians is an obstacle to a well-functioning healthcare system. Our study targets researchers and authorities, who should focus on medical student mental wellness, and medical students themselves.

A Pilot Project to Promote Research Competency in Medical Students Through Journal Clubs: Mixed Methods Study.

Background: Undergraduate medical students often lack hands-on research experience and fundamental scientific research skills, limiting their exposure to the practical aspects of scientific investigation. The Cerrahpasa Neuroscience Society introduced a program to address this deficiency and facilitate student-led research. Objective: The primary goal of this initiative was to enhance medical students' research output by enabling them to generate and publish peer-reviewed papers within the framework of this pilot project. The project aimed to provide an accessible, global model for research training through structured journal clubs, mentorship from experienced peers, and resource access. Methods: In January 2022, a total of 30 volunteer students from various Turkish medical schools participated in this course-based undergraduate research experience program. Students self-organized into 2 groups according to their preferred study type: original research or systematic review. Two final-year students with prior research experience led the project, developing training modules using selected materials. The project was implemented entirely online, with participants completing training modules before using their newly acquired theoretical knowledge to perform assigned tasks. Results: Based on student feedback, the project timeline was adjusted to allow for greater flexibility in meeting deadlines. Despite these adjustments, participants successfully completed their tasks, applying the theoretical knowledge they had gained to their respective assignments. As of April 2024, the initiative has culminated in 3 published papers and 3 more under peer review. The project has also seen an increase in student interest in further involvement and self-paced learning. Conclusions: This initiative leverages globally accessible resources for research training, effectively fostering research competency among participants. It has successfully demonstrated the potential for undergraduates to contribute to medical research output and paved the way for a self-sustaining, student-led research program. Despite some logistical challenges, the project provided valuable insights for future implementations, showcasing the potential for students to engage in meaningful, publishable research.

Enhancing hospital course and outcome prediction in patients with traumatic brain injury: A machine learning study.

PURPOSE: We aimed to use machine learning (ML) algorithms with clinical, lab, and imaging data as input to predict various outcomes in traumatic brain injury (TBI) patients. METHODS: In this retrospective study, blood samples were analyzed for glial fibrillary acidic protein (GFAP) and ubiquitin C-terminal hydrolase L1 (UCH-L1). The non-contrast head CTs were reviewed by two neuroradiologists for TBI common data elements (CDE). Three outcomes were designed to predict: discharged or admitted for further management (prediction 1), deceased or not deceased (prediction 2), and admission only, prolonged stay, or neurosurgery performed (prediction 3). Five ML models were trained. SHapley Additive exPlanations (SHAP) analyses were used to assess the relative significance of variables. RESULTS: Four hundred forty patients were used to predict predictions 1 and 2, while 271 patients were used in prediction 3. Due to Prediction 3's hospitalization requirement, deceased and discharged patients could not be utilized. The Random Forest model achieved an average accuracy of 1.00 for prediction 1 and an accuracy of 0.99 for prediction 2. The Random Forest model achieved a mean accuracy of 0.93 for prediction 3. Key features were extracranial injury, hemorrhage, UCH-L1 for prediction 1; The Glasgow Coma Scale, age, GFAP for prediction 2; and GFAP, subdural hemorrhage volume, and pneumocephalus for prediction 3, per SHAP analysis. CONCLUSION: Combining clinical and laboratory parameters with non-contrast CT CDEs allowed our ML models to accurately predict the designed outcomes of TBI patients. GFAP and UCH-L1 were among the significant predictor variables, demonstrating the importance of these biomarkers.

Factors associated with substance use among preclinical medical students in Turkey: a cross-sectional study.

Introduction: Medical students experience high levels of stress due to their rigorous training, which can negatively affect their mental health. This study aimed to investigate substance use habits of medical students at Istanbul University-Cerrahpasa and the association on their mental health and demographic factors. Methods: This cross-sectional survey study was conducted in March-April 2022 among preclinical medical students (years 1-3 of a 6-year program). A confidential, anonymous online survey consisting of four sections on sociodemographic and educational characteristics, nicotine use and dependence [Fagerström Test for Nicotine Dependence (FTND)], alcohol use [Alcohol Use Disorders Identification Test (AUDIT)], mental health status [12-item General Health Questionnaire (GHQ-12)], was distributed to 1131 students via WhatsApp and Telegram text messages. Mann-Whitney U and Kruskal Wallis tests compared variables' distribution in the questionnaire categories. Spearman's correlation assessed associations between scales. Significance was p < 0.05. Results: The study included 190 medical students. A total of 26.3% of the participants were smokers, with 8.4% showing moderate to high levels of nicotine dependence. An estimated 45.8% and 8.4%reported low-risk consumption and risky usage of alcohol, respectively. There were statistically significant associations between substance use and demographic factors such as sex, GPA, and religious belief. The study found a statistically significant correlation between FTND scores and GHQ-12 scores, and, between FTND scores and AUDIT scores. Conclusion: The findings of this study will inform the development of interventions to improve the mental health and academic performance of medical students at Istanbul University-Cerrahpasa. Furthermore, it will raise awareness about the importance of addressing substance use among medical students in Turkey.

Introduction: Les étudiants en médecine sont assujettis à des niveaux élevés de stress en raison de leur formation rigoureuse, ce qui peut avoir un impact négatif sur leur santé mentale. Cette étude avait pour but d'étudier les habitudes de consommation de substances des étudiants en médecine de l'Université d'Istanbul-Cerrahpasa et l'association avec leur santé mentale et les facteurs démographiques. Méthodes: Cette étude transversale a été menée en mars-avril 2022 parmi les étudiants en médecine au pré-clinique (années 1 à 3 d'un programme de 6 ans). Un questionnaire en ligne confidentiel et anonyme comprenant quatre sections sur les caractéristiques sociodémographiques et éducatives, l'usage et la dépendance à la nicotine [Test de Fagerström pour la dépendance à la nicotine (FTND)], la consommation d'alcool [Test d'identification des troubles liés à la consommation d'alcool (AUDIT)], l'état de santé mentale [Questionnaire général sur la santé en 12 points (GHQ-12)], a été distribué à 1131 étudiants au moyen de messages texte WhatsApp et Telegram. Les tests de Mann-Whitney U et de Kruskal Wallis ont comparé la distribution des variables dans les catégories du questionnaire. La corrélation de Spearman a évalué les associations entre les échelles. Le niveau de signification statistique était p<0,05. Résultats: L'étude a porté sur 190 étudiants en médecine. Au total, 26,3 % des participants étaient des fumeurs, dont 8,4 % présentaient des niveaux modérés à élevés de dépendance à la nicotine. On estime que 45,8 % et 8,4 % des participants ont déclaré une consommation d'alcool à faible risque et une consommation d'alcool à risque, respectivement. Des associations statistiquement significatives ont été observées entre la consommation de substances et des facteurs démographiques tels que le sexe, la moyenne générale et les croyances religieuses. L'étude a mis en évidence une corrélation statistiquement significative entre les scores FTND et les scores GHQ-12, ainsi qu'entre les scores FTND et les scores AUDIT. Conclusion: Les résultats de cette étude permettront d'élaborer des interventions visant à améliorer la santé mentale et les résultats universitaires des étudiants en médecine de l'université d'Istanbul-Cerrahpasa. En outre, elle sensibilisera à l'importance de la prise en charge de l'utilisation de substances chez les étudiants en médecine en Turquie.

Utilizing imaging parameters for functional outcome prediction in acute ischemic stroke: A machine learning study.

BACKGROUND AND PURPOSE: We aimed to predict the functional outcome of acute ischemic stroke patients with anterior circulation large vessel occlusions (LVOs), irrespective of how they were treated or the severity of the stroke at admission, by only using imaging parameters in machine learning models. METHODS: Consecutive adult patients with anterior circulation LVOs who were scanned with CT angiography (CTA) and CT perfusion were queried in this single-center, retrospective study. The favorable outcome was defined as a modified Rankin score (mRS) of 0-2 at 90 days. Predictor variables included only imaging parameters. CatBoost, XGBoost, and Random Forest were employed. Algorithms were evaluated using the area under the receiver operating characteristic curve (AUROC), the area under the precision-recall curve (AUPRC), accuracy, Brier score, recall, and precision. SHapley Additive exPlanations were implemented. RESULTS: A total of 180 patients (102 female) were included, with a median age of 69.5. Ninety-two patients had an mRS between 0 and 2. The best algorithm in terms of AUROC was XGBoost (0.91). Furthermore, the XGBoost model exhibited a precision of 0.72, a recall of 0.81, an AUPRC of 0.83, an accuracy of 0.78, and a Brier score of 0.17. Multiphase CTA collateral score was the most significant feature in predicting the outcome. CONCLUSIONS: Using only imaging parameters, our model had an AUROC of 0.91 which was superior to most previous studies, indicating that imaging parameters may be as accurate as conventional predictors. The multiphase CTA collateral score was the most predictive variable, highlighting the importance of collaterals.

Artificial intelligence/machine learning for neuroimaging to predict hemorrhagic transformation: Systematic review/meta-analysis.

BACKGROUND AND PURPOSE: Early and reliable prediction of hemorrhagic transformation (HT) in patients with acute ischemic stroke (AIS) is crucial for treatment decisions and early intervention. The purpose of this study was to conduct a systematic review and meta-analysis on the performance of artificial intelligence (AI) and machine learning (ML) models that utilize neuroimaging to predict HT. METHODS: A systematic search of PubMed, EMBASE, and Web of Science was conducted until February 19, 2024. Inclusion criteria were as follows: patients with AIS who received reperfusion therapy; AI/ML algorithm using imaging to predict HT; or presence of sufficient data on the predictive performance. Exclusion criteria were as follows: articles with less than 20 patients; articles lacking algorithms that operate solely on images; or articles not detailing the algorithm used. The quality of eligible studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 and Checklist for Artificial Intelligence in Medical Imaging. Pooled sensitivity, specificity, and diagnostic odds ratio (DOR) were calculated using a random-effects model, and a summary receiver operating characteristic curve was constructed using the Reitsma method. RESULTS: We identified six eligible studies, which included 1640 patients. Aside from an unclear risk of bias regarding flow and timing identified in two of the studies, all studies showed low risk of bias and applicability concerns in all categories. Pooled sensitivity, specificity, and DOR were .849, .878, and 45.598, respectively. CONCLUSION: AI/ML models can reliably predict the occurrence of HT in AIS patients. More prospective studies are needed for subgroup analyses and higher clinical certainty and usefulness.

Diagnostic Performance of ASL-MRI and FDG-PET in Frontotemporal Dementia: A Systematic Review and Meta-Analysis.

BACKGROUND: While the diagnosis of frontotemporal dementia (FTD) is based mostly on clinical features, [18F]-FDG PET has been investigated as a potential imaging golden standard in ambiguous cases, with arterial spin labeling (ASL) MRI gaining recent interest. PURPOSE: The purpose of this study is to conduct a systematic review and meta-analysis on the diagnostic performance of ASL MRI in FTD patients and compare it to that of [18F]-FDG PET. DATA SOURCES: A systematic search of PubMed, Scopus and EMBASE was conducted until March 13, 2024. STUDY SELECTION: Inclusion criteria were: original articles, patients with FTD and/or its variants, use of ASL MR perfusion imaging with or without [18F]-FDG PET, presence of sufficient diagnostic performance data. Exclusion criteria were: meeting abstracts, comments, summaries, protocols, letters and guidelines, longitudinal studies, overlapping cohorts. DATA ANALYSIS: The quality of eligible studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2. Pooled sensitivity, specificity, and diagnostic odds ratio (DOR) for [18F]-FDG PET and ASL MRI were calculated, and a summary receiver operating characteristic curve was plotted. DATA SYNTHESIS: Seven eligible studies were identified, which included a total of 102 FTD patients. Aside from some of the studies showing at worst an unclear risk of bias in patient selection, index test, flow and timing, all studies showed low risk of bias and applicability concerns in all categories. Data from 4 studies was included in our meta-analysis for ASL MRI and 3 studies for [18F]-FDG PET. Pooled sensitivity, specificity and DOR were 0.70 (95% CI: 0.59-0.79), 0.81 (95% CI: 0.71-0.88) and 8.00 (95% CI: 3.74-17.13) for ASL MRI, and 0.88 (95% CI: 0.71-0.96), 0.89 (95% CI: 0.43-0.99) and 47.18 (95% CI: 10.77-206.75) for [18F]-FDG PET. LIMITATIONS: The number of studies was relatively small, with a small sample size. The studies used different scanning protocols as well as a mix of diagnostic metrics, all of which might have introduced heterogeneity in the data. CONCLUSIONS: While ASL MRI performed worse than [18F]-FDG PET in the diagnosis of FTD, it exhibited a decent diagnostic performance to justify its further investigation as a quicker and more convenient alternative. ABBREVIATIONS: 3DPCASL, 3D pseudocontinuous ASL; AD, Alzheimer's disease; ASL, arterial spin labeling; AUC, area under the curve; CI, confidence interval; DOR, diagnostic odds ratio; FN, false negative; FP, false positive; FTD, frontotemporal dementia; LE, limbic encephalitis; NLR, negative likelihood ratio; PASL, pulsed ASL; PLD, post-label delay; PLR, positive likelihood ratio; PRISMA, PSP, progressive supranuclear palsy; Preferred Reporting Items for Systematic Reviews and Meta-Analysis; SROC, summary receiver operative characteristic; TN, true negative; TP, true positive; QUADAS-2, Quality Assessment of Diagnostic Accuracy Studies-2.

Data-Driven Prognostication in Distal Medium Vessel Occlusions Using Explainable Machine Learning.

BACKGROUND AND PURPOSE: Distal medium vessel occlusions (DMVOs) are estimated to cause acute ischemic stroke (AIS) in 25-40% of cases. Prognostic models can inform patient counseling and research by enabling outcome predictions. However, models designed specifically for DMVOs are lacking. MATERIALS AND METHODS: This retrospective study developed a machine learning model to predict 90-day unfavorable outcome [defined as a modified Rankin Scale (mRS) score of 3-6] in 164 primary DMVO patients. A model developed with the TabPFN algorithm utilized selected clinical, laboratory, imaging, and treatment data with the Least Absolute Shrinkage and Selection Operator feature selection. Performance was evaluated via 5-repeat 5-fold cross-validation. Model discrimination and calibration were evaluated. SHapley Additive Explanations (SHAP) identified influential features. A web application deployed the model for individualized predictions. RESULTS: The model achieved an area under the receiver operating characteristic curve of 0.815 (95% CI: 0.79-0.841) for predicting unfavorable outcome, demonstrating good discrimination, and a Brier score of 0.19 (95% CI: 0.177-0.202), demonstrating good calibration. SHAP analysis ranked admission National Institutes of Health Stroke Scale (NIHSS) score, premorbid mRS, type of thrombectomy, modified thrombolysis in cerebral infarction score, and history of malignancy as top predictors. The web application enables individualized prognostication. CONCLUSIONS: Our machine learning model demonstrated good discrimination and calibration for predicting 90-day unfavorable outcomes in primary DMVO strokes. This study demonstrates the potential for personalized prognostic counseling and research to support precision medicine in stroke care and recovery. ABBREVIATIONS: DMVO = Distal medium vessel occlusion; AIS = acute ischemic stroke; mRS = modified Rankin Scale; SHAP = SHapley Additive Explanations; NIHSS = National Institutes of Health Stroke Scale; ST = stroke thrombectomy; TRIPOD = Transparent Reporting of Multivariable Prediction Models for Individual Prognosis or Diagnosis; CT = computed tomography; CTP = CT perfusion; MRI = magnetic resonance imaging; CTA = CT angiography; DVT = deep vein thrombosis; PE = pulmonary emboli; TIA = transient ischemic attack; BMI = body mass index; ALP = alkaline phosphatase; ALT = alanine transaminase; AST = aspartate aminotransferase; NCCT-ASPECTS = Alberta Stroke Program Early CT Score; IVT = intravenous thrombolysis; mTICI = modified thrombolysis in cerebral infarction; ER = emergency room; kNN = k-nearest neighbor; LASSO = Least Absolute Shrinkage and Selection Operator; PDPs = partial dependence plots; ROC = receiver operating characteristic; PRC = precision-recall curve; AUROC = area under the ROC curve; AUPRC = area under the PRC; CI = confidence interval.

Artificial Intelligence-Generated Editorials in Radiology: Can Expert Editors Detect Them?

BACKGROUND AND PURPOSE: We aimed to evaluate GPT-4's ability to write radiology editorials and to compare these with human-written counterparts, thereby determining their real-world applicability for scientific writing. MATERIALS AND METHODS: Sixteen editorials from eight journals were included. To generate the AI-written editorials, the summary of 16 human-written editorials was fed into GPT-4. Six experienced editors reviewed the articles. First, an unpaired approach was used. The raters were asked to evaluate the content of each article using a 1-5 Likert scale across specified metrics. Then, they determined whether the editorials were written by humans or AI. The articles were then evaluated in pairs to determine which article was generated by AI and which should be published. Finally, the articles were analyzed with an AI detector and for plagiarism. RESULTS: The human-written articles had a median AI probability score of 2.0%, whereas the AI-written articles had 58%. The median similarity score among AI-written articles was 3%. 58% of unpaired articles were correctly classified regarding authorship. Rating accuracy was increased to 70% in the paired setting. AI-written articles received slightly higher scores in most metrics. When stratified by perception, human-written perceived articles were rated higher in most categories. In the paired setting, raters strongly preferred publishing the article they perceived as human-written (82%). CONCLUSIONS: GPT-4 can write high-quality articles that iThenticate does not flag as plagiarized, which may go undetected by editors, and that detection tools can detect to a limited extent. Editors showed a positive bias toward human-written articles. ABBREVIATIONS: AI = Artificial intelligence; LLM = large language model; SD = standard deviation.

CT perfusion based rCBF <38% volume is independently and negatively associated with digital subtraction angiography collateral score in anterior circulation large vessel occlusions.

BACKGROUND: Collateral status (CS) is an important biomarker of functional outcomes in patients with acute ischemic stroke secondary to large vessel occlusion (AIS-LVO). Pretreatment CT perfusion (CTP) parameters serve as reliable surrogates of collateral status (CS). In this study, we aim to assess the relationship between the relative cerebral blood flow less than 38% (rCBF <38%), with the reference standard American Society of Interventional and Therapeutic Neuroradiology (ASITN) collateral score (CS) on DSA. METHODS: In this prospectively collected, retrospectively reviewed analysis, inclusion criteria were as follows: (a) CT angiography (CTA) confirmed anterior circulation large vessel occlusion from 9/1/2017 to 10/01/2023; (b) diagnostic CT perfusion; and (c) underwent mechanical thrombectomy with documented ASITN CS. The ratios of the CTP-derived CBF values were calculated by dividing the values of the ischemic lesion by the corresponding values of the contralateral normal region (which were defined as rCBF). Spearman's rank correlation and logistic regression analysis were performed to determine the relationship of rCBF <38% lesion volume with DSA ASITN CS. p ≤ .05 was considered significant. RESULTS: In total, 223 patients [mean age: 67.77 ± 15.76 years, 56.1% (n = 125) female] met our inclusion criteria. Significant negative correlation was noted between rCBF <38% volume and DSA CS (ρ = -0.37, p < .001). On multivariate logistic regression analysis, rCBF <38% volume was found to be independently associated with worse ASITN CS (unadjusted OR: 3.03, 95% CI: 1.60-5.69, p < .001, and adjusted OR: 2.73, 95% CI: 1.34-5.50, p < .01). CONCLUSION: Greater volume of tissue with rCBF <38% is independently associated with better DSA CS. rCBF <38% is a useful adjunct tool in collateralization-based prognostication. Future studies are needed to expand our understanding of the role of rCBF <38% within the decision-making in patients with AIS-LVO.

CT Perfusion Derived rCBV < 42% Lesion Volume Is Independently Associated with Followup FLAIR Infarct Volume in Anterior Circulation Large Vessel Occlusion.

Pretreatment CT Perfusion (CTP) parameter rCBV < 42% lesion volume has recently been shown to predict 90-day mRS. In this study, we aim to assess the relationship between rCBV < 42% and a radiographic follow-up infarct volume delineated on FLAIR images. In this retrospective evaluation of our prospectively collected database, we included acute stroke patients triaged by multimodal CT imaging, including CT angiography and perfusion imaging, with confirmed anterior circulation large vessel occlusion between 9 January 2017 and 10 January 2023. Follow-up FLAIR imaging was used to determine the final infarct volume. Student t, Mann-Whitney-U, and Chi-Square tests were used to assess differences. Spearman's rank correlation and linear regression analysis were used to assess associations between rCBV < 42% and follow-up infarct volume on FLAIR. In total, 158 patients (median age: 68 years, 52.5% female) met our inclusion criteria. rCBV < 42% (ρ = 0.56, p < 0.001) significantly correlated with follow-up-FLAIR infarct volume. On multivariable linear regression analysis, rCBV < 42% lesion volume (beta = 0.60, p < 0.001), ASPECTS (beta = -0.214, p < 0.01), mTICI (beta = -0.277, p < 0.001), and diabetes (beta = 0.16, p < 0.05) were independently associated with follow-up infarct volume. The rCBV < 42% lesion volume is independently associated with FLAIR follow-up infarct volume.

Pretreatment parameters associated with hemorrhagic transformation among successfully recanalized medium vessel occlusions.

Although pretreatment radiographic biomarkers are well established for hemorrhagic transformation (HT) following successful mechanical thrombectomy (MT) in large vessel occlusion (LVO) strokes, they are yet to be explored for medium vessel occlusion (MeVO) acute ischemic strokes. We aim to investigate pretreatment imaging biomarkers representative of collateral status, namely the hypoperfusion intensity ratio (HIR) and cerebral blood volume (CBV) index, and their association with HT in successfully recanalized MeVOs. A prospectively collected registry of acute ischemic stroke patients with MeVOs successfully recanalized with MT between 2019 and 2023 was retrospectively reviewed. A multivariate logistic regression for HT of any subtype was derived by combining significant univariate predictors into a forward stepwise regression with minimization of Akaike information criterion. Of 60 MeVO patients successfully recanalized with MT, HT occurred in 28.3% of patients. Independent factors for HT included: diabetes mellitus history (p = 0.0005), CBV index (p = 0.0071), and proximal versus distal occlusion location (p = 0.0062). A multivariate model with these factors had strong diagnostic performance for predicting HT (area under curve [AUC] 0.93, p < 0.001). Lower CBV indexes, distal occlusion location, and diabetes history are significantly associated with HT in MeVOs successfully recanalized with MT. Of note, HIR was not found to be significantly associated with HT.

The Relative Cerebral Blood Volume (rCBV) < 42% Is Independently Associated with Collateral Status in Anterior Circulation Large Vessel Occlusion.

Background: The pretreatment CT perfusion (CTP) marker the relative cerebral blood volume (rCBV) < 42% lesion volume has recently been shown to predict 90-day functional outcomes; however, studies assessing correlations of the rCBV < 42% lesion volume with other outcomes remain sparse. Here, we aim to assess the relationship between the rCBV < 42% lesion volume and the reference standard digital subtraction angiography (DSA)-derived American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology (ASITN) collateral score, hereby referred as the DSA CS. Methods: In this retrospective evaluation of our prospectively collected database, we included acute stroke patients triaged by multimodal CT imaging, including CT angiography and perfusion imaging, with confirmed anterior circulation large vessel occlusion between 1 September 2017 and 1 October 2023. Group differences were assessed using the Student's t test, Mann-Whitney U test and Chi-Square test. Spearman's rank correlation and logistic regression analyses were used to assess associations between rCBV < 42% and DSA CS. Results: In total, 222 patients (median age: 69 years, 56.3% female) met our inclusion criteria. In the multivariable logistic regression analysis, taking into account age, sex, race, hypertension, hyperlipidemia, diabetes, atrial fibrillation, prior stroke or transient ischemic attack, the admission National Institute of Health stroke scale, the premorbid modified Rankin score, the Alberta stroke program early CT score (ASPECTS), and segment occlusion, the rCBV < 42% lesion volume (adjusted OR: 0.98, p < 0.05) was independently associated with the DSA CS. Conclusion: The rCBV < 42% lesion volume is independently associated with the DSA CS.