Data Extraction from Free-Text Reports on Mechanical Thrombectomy in Acute Ischemic Stroke Using ChatGPT: A Retrospective Analysis.

Background Procedural details of mechanical thrombectomy in patients with ischemic stroke are important predictors of clinical outcome and are collected for prospective studies or national stroke registries. To date, these data are collected manually by human readers, a labor-intensive task that is prone to errors. Purpose To evaluate the use of the large language models (LLMs) GPT-4 and GPT-3.5 to extract data from neuroradiology reports on mechanical thrombectomy in patients with ischemic stroke. Materials and Methods This retrospective study included consecutive reports from patients with ischemic stroke who underwent mechanical thrombectomy between November 2022 and September 2023 at institution 1 and between September 2016 and December 2019 at institution 2. A set of 20 reports was used to optimize the prompt, and the ability of the LLMs to extract procedural data from the reports was compared using the McNemar test. Data manually extracted by an interventional neuroradiologist served as the reference standard. Results A total of 100 internal reports from 100 patients (mean age, 74.7 years ± 13.2 [SD]; 53 female) and 30 external reports from 30 patients (mean age, 72.7 years ± 13.5; 18 male) were included. All reports were successfully processed by GPT-4 and GPT-3.5. Of 2800 data entries, 2631 (94.0% [95% CI: 93.0, 94.8]; range per category, 61%-100%) data points were correctly extracted by GPT-4 without the need for further postprocessing. With 1788 of 2800 correct data entries, GPT-3.5 produced fewer correct data entries than did GPT-4 (63.9% [95% CI: 62.0, 65.6]; range per category, 14%-99%; P < .001). For the external reports, GPT-4 extracted 760 of 840 (90.5% [95% CI: 88.3, 92.4]) correct data entries, while GPT-3.5 extracted 539 of 840 (64.2% [95% CI: 60.8, 67.4]; P < .001). Conclusion Compared with GPT-3.5, GPT-4 more frequently extracted correct procedural data from free-text reports on mechanical thrombectomy performed in patients with ischemic stroke. © RSNA, 2024 Supplemental material is available for this article.

A novel NODAL variant in a young embolic stroke patient with visceral heterotaxy.

BACKGROUND: Ischemic stroke in young adults can be caused by a variety of etiologies including the monogenic disorders. Visceral heterotaxy is a condition caused by abnormal left-right determinations during embryonic development. We aimed to determine the cause of a young ischemic stroke patient with visceral heterotaxy. CASE PRESENTATION: We performed neurological, radiological, and genetic evaluations in a 17-year-old male patient presenting ischemic stroke and visceral heterotaxy to determine the underlying cause of this rare disease combination. Brain magnetic resonance imaging (MRI) showed evidence of embolic stroke, abdominal computed tomography (CT) showed visceral heterotaxy, and echocardiogram showed cardiac anomaly with right-to-left-shunt (RLS). Whole genome sequencing (WGS) revealed a heterozygous missense variant (NM_018055.5: c.1016 T > C, p.(Met339Val)) in the NODAL gene, which is essential to the determination of the left-right body axis. CONCLUSIONS: Our study highlights the importance of evaluating genetic etiology in young ischemic stroke and the need for stroke risk management in visceral heterotaxy patients with RLS. To the best of our knowledge, we report the first genetically-confirmed case of visceral heterotaxy with young embolic stroke reported to date.

Application of radiomics in ischemic stroke.

In recent years, radiomics has emerged as a novel research methodology that plays a crucial role in the diagnosis and treatment of ischemic stroke. By integrating multimodal medical imaging techniques such as computed tomography and magnetic resonance imaging, radiomics offers in-depth insights into aspects such as the extent of brain tissue damage and hemodynamics. These data help physicians to accurately assess patient condition, select optimal treatment strategies, and predict recovery trajectories and long-term prognoses, thereby enhancing treatment efficacy and reducing the risk of complications. With the anticipated further advancements in radiomic technology, this methodology has great potential for expanded applications in the early detection, treatment, and prognosis of ischemic stroke. The present narrative review explores the burgeoning field of radiomics and its transformative impact on ischemic stroke.

Synchronous subarachnoid haemorrhage and ischaemic stroke as a result of complete internal carotid artery occlusion.

A previously healthy man in his 60s presents with a one-day history of insidious onset headache and 'walking into doors'. He reported transient right arm pain and tingling but no weakness. A CT brain showed a right middle cerebral artery (MCA) infarct as well as a synchronous right frontal lobe convexal subarachnoid haemorrhage (cSAH). An arch to vertex CT angiogram demonstrated right MCA occlusion and complete right internal carotid artery (ICA) occlusion from its origin. Reconstitution of flow was seen within the distal right ICA at the level of the distal foramen lacerum. A repeat CT of the brain after one week showed a stable appearance of the bleeding and infarct. He was commenced on antiplatelet therapy for the treatment of the ischaemic stroke as well as secondary prevention.This is a rare case of synchronous right MCA infarct as well as a right frontal cSAH secondary to ipsilateral carotid artery occlusion.

Contrast extravasation mimicking intracerebral and intraventricular hemorrhage after intravenous thrombolytic treatment of ischemic stroke: a case report.

BACKGROUND: Although contrast extravasation on follow-up head computed tomography (CT) is frequently visualized after endovascular treatment, this phenomenon is rare after intravenous thrombolytic treatment in patients with acute ischemic stroke (AIS). Here, we report a case of contrast extravasation mimicking intracerebral hemorrhage (ICH) with intraventricular extension after intravenous thrombolytic treatment and computed tomography angiography (CTA). CASE PRESENTATION: A 52-year-old man presented with right-sided hemiparesis and hypoesthesia. Initial non-contrast head CT was negative for intracranial hemorrhage and acute ischemic changes. He received intravenous treatment with tenecteplase 3.8 h after the onset of stroke. CTA of the head and neck was performed at 4.3 h after stroke onset. It showed no stenosis or occlusion of the carotid and major intracranial arteries. At about 1.5 h after CTA, the right-sided hemiparesis deteriorated, accompanied by drowsiness, aphasia, and urinary incontinence. Immediate head CT showed hyperdense lesions with mild space-occupying effect in the left basal ganglia and both lateral ventricles. The hyperdense lesions were reduced in size on follow-up CT after 5 h. Two days later, CT showed that the hyperdense lesions in the lateral ventricles almost completely disappeared and only a small amount remained in the infarcted area. CONCLUSIONS: Contrast extravasation into the brain tissue and lateral ventricles, mimicking ICH with intraventricular extension, could occur after intravenous thrombolytic treatment and CTA in a patient with AIS, which might lead to misdiagnosis and wrong treatment of the patient. The rapid resolution of intracranial hyperdense lesions is key to differentiate contrast extravasation from ICH on serial non-enhanced CT.

CT Imaging Computed Tomography/Computed Tomography Angiography/Perfusion in Acute Ischemic Stroke and Vasospasm.

Computed tomography (CT), CT angiography (CTA), and CT perfusion (CTP) play crucial roles in the comprehensive evaluation and management of acute ischemic stroke, aneurysmal subarachnoid hemorrhage (SAH), and vasospasm. CTP provides functional data about cerebral blood flow, allowing radiologists, neurointerventionalists, and stroke neurologists to more accurately delineate the volume of core infarct and ischemic penumbra allowing for patient-specific treatment decisions to be made. CTA and CTP are used in tandem to evaluate for vasospasm associated with aneurysmal SAH and can help provide an insight into the physiologic impact of angiographic vasospasm, better triaging patients for medical and interventional treatment.

MR Imaging-based Biomarker Development in Hemorrhagic Stroke Patients Including Brain Iron Quantification, Diffusion Tensor Imaging, and Phenomenon of Ultra-early Erythrolysis.

This review article discusses the role of MR imaging-based biomarkers in understanding and managing hemorrhagic strokes, focusing on intracerebral hemorrhage (ICH) and aneurysmal subarachnoid hemorrhage. ICH is a severe type of stroke with high mortality and morbidity rates, primarily caused by the rupture of small blood vessels in the brain, resulting in hematoma formation. MR imaging-based biomarkers, including brain iron quantification, ultra-early erythrolysis detection, and diffusion tensor imaging, offer valuable insights for hemorrhagic stroke management. These biomarkers could improve early diagnosis, risk stratification, treatment monitoring, and patient outcomes in the future, revolutionizing our approach to hemorrhagic strokes.

MR Imaging Techniques for Acute Ischemic Stroke and Delayed Cerebral Ischemia Following Subarachnoid Hemorrhage.

Acute ischemic stroke (AIS) is a leading cause of death and disability worldwide, and its prevalence is expected to increase with global population aging and the burgeoning obesity epidemic. Clinical care for AIS has evolved during the past 3 decades, and it comprises of 3 major tenants: (1) timely recanalization of occluded vessels with intravenous thrombolysis or endovascular thrombectomy, (2) prompt initiation of antithrombotic agents to prevent stroke recurrences, and (3) poststroke supportive care and rehabilitation. In this article, we summarize commonly used MR sequences for AIS and DCI and highlight their clinical applications.

Dual-Energy Computed Tomography in the Evaluation and Management of Subarachnoid Hemorrhage, Intracranial Hemorrhage, and Acute Ischemic Stroke.

Dual-energy computed tomography (DECT) has emerged as a valuable imaging modality in the diagnosis and management of various cerebrovascular pathologies, including subarachnoid hemorrhage, intracranial hemorrhage, and acute ischemic stroke. This article reviews the principles of DECT and its applications in the evaluation and management of these conditions. The authors discuss the advantages of DECT over conventional computed tomography, as well as its limitations, and provide an overview of current research and future directions in the field.

Multimaterial decomposition in dual-energy CT for characterization of clots from acute ischemic stroke patients.

BACKGROUND: Nowadays, there is no method to quantitatively characterize the material composition of acute ischemic stroke thrombi prior to intervention, but dual-energy CT (DE-CT) offers imaging-based multimaterial decomposition. We retrospectively investigated the material composition of thrombi ex vivo using DE-CT with histological analysis as a reference. METHODS: Clots of 70 patients with acute ischemic stroke were extracted by mechanical thrombectomy and scanned ex vivo in formalin-filled tubes with DE-CT. Multimaterial decomposition in the three components, i.e., red blood cells (RBC), white blood cells (WBC), and fibrin/platelets (F/P), was performed and compared to histology (hematoxylin/eosin staining) as reference. Attenuation and effective Z values were assessed, and histological composition was compared to stroke etiology according to the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) criteria. RESULTS: Histological and imaging analysis showed the following correlation coefficients for RBC (r = 0.527, p < 0.001), WBC (r = 0.305, p = 0.020), and F/P (r = 0.525, p < 0.001). RBC-rich thrombi presented higher clot attenuation in Hounsfield units than F/P-rich thrombi (51 HU versus 42 HU, p < 0.01). In histological analysis, cardioembolic clots showed less RBC (40% versus 56%, p = 0.053) and more F/P (53% versus 36%, p = 0.024), similar to cryptogenic clots containing less RBC (34% versus 56%, p = 0.006) and more F/P (58% versus 36%, p = 0.003) than non-cardioembolic strokes. No difference was assessed for the mean WBC portions in all TOAST groups. CONCLUSIONS: DE-CT has the potential to quantitatively characterize the material composition of ischemic stroke thrombi. RELEVANCE STATEMENT: Using DE-CT, the composition of ischemic stroke thrombi can be determined. Knowledge of histological composition prior to intervention offers the opportunity to define personalized treatment strategies for each patient to accomplish faster recanalization and better clinical outcomes. KEY POINTS: • Acute ischemic stroke clots present different recanalization success according to histological composition. • Currently, no method can determine clot composition prior to intervention. • DE-CT allows quantitative material decomposition of thrombi ex vivo in red blood cells, white blood cells, and fibrin/platelets. • Histological clot composition differs between stroke etiology. • Insights into the histological composition in situ offer personalized treatment strategies.

Diagnostic Performance of Carotid Ring Sign on CT-Angiography in Internal Carotid True Occlusion.

BACKGROUND: To differentiate between pseudo occlusion (PO) and true occlusion (TO) of internal carotid artery (ICA) is important in thrombectomy treatment planning for patients with acute ischemic stroke. Although delayed contrast filling has been differentiated carotid PO from TO, its application has been limited by the implementations of multiphasic computed tomography angiography. In this study, we hypothesized that carotid ring sign, which is readily acquired from single-phasic CTA, can sufficiently differentiate carotid TO from PO. METHODS: One thousand four hundred and twenty patients with anterior circulation stroke receiving endovascular therapy were consecutively recruited through a hospital- and web-based registry. Two hundred patients with nonvisualization of the proximal ICA were included in the analysis after a retrospective screening. Diagnosis of PO or TO of the cervical segment of ICA was made based on digital subtraction angiography. Diagnostic performances of carotid ring sign on arterial-phasic CTA and delayed contrast filling on multiphasic computed tomography angiography were evaluated and compared. RESULTS: One-hundred twelve patients had ICA PO and 88 had TO. Carotid ring sign was more common in patients with TO (70.5% versus 6.3%; P<0.001), whereas delayed contrast filling was more common in PO (94.9% versus 7.7%; P<0.001). The sensitivity and specificity of carotid ring sign in diagnosing carotid TO were 0.70 and 0.94, respectively, whereas sensitivity and specificity of delayed contrast filling was 0.95 and 0.92 in judging carotid PO. CONCLUSIONS: Carotid ring sign is a potent imaging marker in diagnosing ICA TO. Carotid ring sign could be complementary to delayed contrast filling sign in differentiating TO from PO, in particular in centers with only single-phasic CTA.

Magnetic resonance imaging-based deep learning imaging biomarker for predicting functional outcomes after acute ischemic stroke.

PURPOSE: Clinical risk scores are essential for predicting outcomes in stroke patients. The advancements in deep learning (DL) techniques provide opportunities to develop prediction applications using magnetic resonance (MR) images. We aimed to develop an MR-based DL imaging biomarker for predicting outcomes in acute ischemic stroke (AIS) and evaluate its additional benefit to current risk scores. METHOD: This study included 3338 AIS patients. We trained a DL model using deep neural network architectures on MR images and radiomics to predict poor functional outcomes at three months post-stroke. The DL model generated a DL score, which served as the DL imaging biomarker. We compared the predictive performance of this biomarker to five risk scores on a holdout test set. Additionally, we assessed whether incorporating the imaging biomarker into the risk scores improved the predictive performance. RESULTS: The DL imaging biomarker achieved an area under the receiver operating characteristic curve (AUC) of 0.788. The AUCs of the five studied risk scores were 0.789, 0.793, 0.804, 0.810, and 0.826, respectively. The imaging biomarker's predictive performance was comparable to four of the risk scores but inferior to one (p = 0.038). Adding the imaging biomarker to the risk scores improved the AUCs (p-values) to 0.831 (0.003), 0.825 (0.001), 0.834 (0.003), 0.836 (0.003), and 0.839 (0.177), respectively. The net reclassification improvement and integrated discrimination improvement indices also showed significant improvements (all p < 0.001). CONCLUSIONS: Using DL techniques to create an MR-based imaging biomarker is feasible and enhances the predictive ability of current risk scores.

Magnetic Resonance Imaging Assists With Determining Etiology After Transient Ischemic Attack or Minor Stroke.

BACKGROUND: Magnetic resonance imaging infarct topography may assist with determining stroke etiology. The influence of diffusion-weighted imaging (DWI)-positive lesions on etiology determination in patients with transient ischemic attack or minor stroke is not well studied. METHODS AND RESULTS: We prospectively enrolled patients between 2010 and 2017 in 2 studies; participants with a final diagnosis of probable or definite transient ischemic attack or stroke were pooled for analysis. The primary outcome was the adjudicated ischemic etiology. We compared proportion of each etiology (cardioembolic, large-vessel, small-vessel disease, other) in patients who had DWI positivity compared with DWI negativity. We used logistic regression to determine the adjusted odds ratio (OR) for each etiology compared with undetermined by DWI positivity. The final analysis included 1498 patients: 832 (55.5%) were DWI-positive. DWI-positive patients were more likely to be diagnosed with small-vessel disease (19.1% versus 5.3%) and less likely with undetermined etiology (36.9% versus 53.0%; P<0.001). After adjustment, the presence of any DWI lesion was associated with increased odds of assigning any etiology (OR, 1.8 [95% CI, 1.3-2.5]). A single DWI lesion was associated with increased odds of small-vessel disease diagnosis (OR, 9.5 [95% CI, 6.4-14.0]), and multiple DWI lesions with reduced odds of small-vessel disease (OR, 0.2 [95% CI, 0.1-0.4]) but increased odds of all other etiologies compared with undetermined etiology. CONCLUSIONS: Any DWI-positive lesion after suspected transient ischemic attack or minor stroke was associated with increased odds of assigning a etiology. Presence and topography of DWI lesions on magnetic resonance imaging may assist with etiology determination and may impact stroke prevention therapies.

Thrombolysis for Wake-Up Stroke Versus Non-Wake-Up Unwitnessed Stroke: EOS Individual Patient Data Meta-Analysis.

BACKGROUND: Stroke with unknown time of onset can be categorized into 2 groups; wake-up stroke (WUS) and unwitnessed stroke with an onset time unavailable for reasons other than wake-up (non-wake-up unwitnessed stroke, non-WUS). We aimed to assess potential differences in the efficacy and safety of intravenous thrombolysis (IVT) between these subgroups. METHODS: Patients with an unknown-onset stroke were evaluated using individual patient-level data of 2 randomized controlled trials (WAKE-UP [Efficacy and Safety of MRI-Based Thrombolysis in Wake-Up Stroke], THAWS [Thrombolysis for Acute Wake-Up and Unclear-Onset Strokes With Alteplase at 0.6 mg/kg]) comparing IVT with placebo or standard treatment from the EOS (Evaluation of Unknown-Onset Stroke Thrombolysis trial) data set. A favorable outcome was prespecified as a modified Rankin Scale score of 0 to 1 at 90 days. Safety outcomes included symptomatic intracranial hemorrhage at 22 to 36 hours and 90-day mortality. The IVT effect was compared between the treatment groups in the WUS and non-WUS with multivariable logistic regression analysis. RESULTS: Six hundred thirty-four patients from 2 trials were analyzed; 542 had WUS (191 women, 272 receiving alteplase), and 92 had non-WUS (42 women, 43 receiving alteplase). Overall, no significant interaction was noted between the mode of onset and treatment effect (P value for interaction=0.796). In patients with WUS, the frequencies of favorable outcomes were 54.8% and 45.5% in the IVT and control groups, respectively (adjusted odds ratio, 1.47 [95% CI, 1.01-2.16]). Death occurred in 4.0% and 1.9%, respectively (P=0.162), and symptomatic intracranial hemorrhage in 1.8% and 0.3%, respectively (P=0.194). In patients with non-WUS, no significant difference was observed in favorable outcomes relative to the control (37.2% versus 29.2%; adjusted odds ratio, 1.76 [0.58-5.37]). One death and one symptomatic intracranial hemorrhage were reported in the IVT group, but none in the control. CONCLUSIONS: There was no difference in the effect of IVT between patients with WUS and non-WUS. IVT showed a significant benefit in patients with WUS, while there was insufficient statistical power to detect a substantial benefit in the non-WUS subgroup. REGISTRATION: URL: https://www.clinicaltrials.gov; Unique identifier: CRD42020166903.

Annotation-free prediction of treatment-specific tissue outcome from 4D CT perfusion imaging in acute ischemic stroke.

Acute ischemic stroke is a critical health condition that requires timely intervention. Following admission, clinicians typically use perfusion imaging to facilitate treatment decision-making. While deep learning models leveraging perfusion data have demonstrated the ability to predict post-treatment tissue infarction for individual patients, predictions are often represented as binary or probabilistic masks that are not straightforward to interpret or easy to obtain. Moreover, these models typically rely on large amounts of subjectively segmented data and non-standard perfusion analysis techniques. To address these challenges, we propose a novel deep learning approach that directly predicts follow-up computed tomography images from full spatio-temporal 4D perfusion scans through a temporal compression. The results show that this method leads to realistic follow-up image predictions containing the infarcted tissue outcomes. The proposed compression method achieves comparable prediction results to using perfusion maps as inputs but without the need for perfusion analysis or arterial input function selection. Additionally, separate models trained on 45 patients treated with thrombolysis and 102 treated with thrombectomy showed that each model correctly captured the different patient-specific treatment effects as shown by image difference maps. The findings of this work clearly highlight the potential of our method to provide interpretable stroke treatment decision support without requiring manual annotations.

Emotional processing impairments in patients with insula lesions following stroke.

Functional imaging has helped to understand the role of the human insula as a major processing network for integrating input with the current state of the body. However, these studies remain at a correlative level. Studies that have examined insula damage show lesion-specific performance deficits. Case reports have provided anecdotal evidence for deficits following insula damage, but group lesion studies offer a number of advances in providing evidence for functional representation of the insula. We conducted a systematic literature search to review group studies of patients with insula damage after stroke and identified 23 studies that tested emotional processing performance in these patients. Eight of these studies assessed emotional processing of visual (most commonly IAPS), auditory (e.g., prosody), somatosensory (emotional touch) and autonomic function (heart rate variability). Fifteen other studies looked at social processing, including emotional face recognition, gaming tasks and tests of empathy. Overall, there was a bias towards testing only patients with right-hemispheric lesions, making it difficult to consider hemisphere specificity. Although many studies included an overlay of lesion maps to characterise their patients, most did not differentiate lesion statistics between insula subunits and/or applied voxel-based associations between lesion location and impairment. This is probably due to small group sizes, which limit statistical comparisons. We conclude that multicentre analyses of lesion studies with comparable patients and performance tests are needed to definitively test the specific function of parts of the insula in emotional processing and social interaction.

Does Thrombosis Play a Causal Role in Lacunar Stroke and Cerebral Small Vessel Disease?

BACKGROUND: The importance of thromboembolism in the pathogenesis of lacunar stroke (LS), resulting from cerebral small vessel disease (cSVD), is debated, and although antiplatelets are widely used in secondary prevention after LS, there is limited trial evidence from well-subtyped patients to support this approach. We sought to evaluate whether altered anticoagulation plays a causal role in LS and cSVD using 2-sample Mendelian randomization. METHODS: From a recent genome-wide association study (n=81 190), we used 119 genetic variants associated with venous thrombosis at genome-wide significance (P<5*10-8) and with a linkage disequilibrium r2<0.001 as instrumental variables. We also used genetic associations with stroke from the GIGASTROKE consortium (62 100 ischemic stroke cases: 10 804 cardioembolic stroke, 6399 large-artery stroke, and 6811 LS). In view of the lower specificity for LS with the CT-based phenotyping mainly used in GIGASTROKE, we also used data from patients with magnetic resonance imaging-confirmed LS (n=3199). We also investigated associations with more chronic magnetic resonance imaging features of cSVD, namely, white matter hyperintensities (n=37 355) and diffusion tensor imaging metrics (n=36 533). RESULTS: Mendelian randomization analyses showed that genetic predisposition to venous thrombosis was associated with an increased odds of any ischemic stroke (odds ratio [OR], 1.19 [95% CI, 1.13-1.26]), cardioembolic stroke (OR, 1.32 [95% CI, 1.21-1.45]), and large-artery stroke (OR, 1.41 [95% CI, 1.26-1.57]) but not with LS (OR, 1.07 [95% CI, 0.99-1.17]) in GIGASTROKE. Similar results were found for magnetic resonance imaging-confirmed LS (OR, 0.94 [95% CI, 0.81-1.09]). Genetically predicted risk of venous thrombosis was not associated with imaging markers of cSVD. CONCLUSIONS: These findings suggest that altered thrombosis plays a role in the risk of cardioembolic and large-artery stroke but is not a causal risk factor for LS or imaging markers of cSVD. This raises the possibility that antithrombotic medication may be less effective in cSVD and underscores the necessity for further trials in well-subtyped cohorts with LS to evaluate the efficacy of different antithrombotic regimens in LS.