BACKGROUND: Patients who have had a stroke with unknown time of onset have been previously excluded from thrombolysis. We aimed to establish whether intravenous alteplase is safe and effective in such patients when salvageable tissue has been identified with imaging biomarkers. METHODS: We did a systematic review and meta-analysis of individual patient data for trials published before Sept 21, 2020. Randomised trials of intravenous alteplase versus standard of care or placebo in adults with stroke with unknown time of onset with perfusion-diffusion MRI, perfusion CT, or MRI with diffusion weighted imaging-fluid attenuated inversion recovery (DWI-FLAIR) mismatch were eligible. The primary outcome was favourable functional outcome (score of 0-1 on the modified Rankin Scale [mRS]) at 90 days indicating no disability using an unconditional mixed-effect logistic-regression model fitted to estimate the treatment effect. Secondary outcomes were mRS shift towards a better functional outcome and independent outcome (mRS 0-2) at 90 days. Safety outcomes included death, severe disability or death (mRS score 4-6), and symptomatic intracranial haemorrhage. This study is registered with PROSPERO, CRD42020166903. FINDINGS: Of 249 identified abstracts, four trials met our eligibility criteria for inclusion: WAKE-UP, EXTEND, THAWS, and ECASS-4. The four trials provided individual patient data for 843 individuals, of whom 429 (51%) were assigned to alteplase and 414 (49%) to placebo or standard care. A favourable outcome occurred in 199 (47%) of 420 patients with alteplase and in 160 (39%) of 409 patients among controls (adjusted odds ratio [OR] 1·49 [95% CI 1·10-2·03]; p=0·011), with low heterogeneity across studies (I2=27%). Alteplase was associated with a significant shift towards better functional outcome (adjusted common OR 1·38 [95% CI 1·05-1·80]; p=0·019), and a higher odds of independent outcome (adjusted OR 1·50 [1·06-2·12]; p=0·022). In the alteplase group, 90 (21%) patients were severely disabled or died (mRS score 4-6), compared with 102 (25%) patients in the control group (adjusted OR 0·76 [0·52-1·11]; p=0·15). 27 (6%) patients died in the alteplase group and 14 (3%) patients died among controls (adjusted OR 2·06 [1·03-4·09]; p=0·040). The prevalence of symptomatic intracranial haemorrhage was higher in the alteplase group than among controls (11 [3%] vs two [<1%], adjusted OR 5·58 [1·22-25·50]; p=0·024). INTERPRETATION: In patients who have had a stroke with unknown time of onset with a DWI-FLAIR or perfusion mismatch, intravenous alteplase resulted in better functional outcome at 90 days than placebo or standard care. A net benefit was observed for all functional outcomes despite an increased risk of symptomatic intracranial haemorrhage. Although there were more deaths with alteplase than placebo, there were fewer cases of severe disability or death. FUNDING: None.

BACKGROUND AND PURPOSE: Accurate automated infarct segmentation is needed for acute ischemic stroke studies relying on infarct volumes as an imaging phenotype or biomarker that require large numbers of subjects. This study investigated whether an ensemble of convolutional neural networks trained on multiparametric DWI maps outperforms single networks trained on solo DWI parametric maps. MATERIALS AND METHODS: Convolutional neural networks were trained on combinations of DWI, ADC, and low b-value-weighted images from 116 subjects. The performances of the networks (measured by the Dice score, sensitivity, and precision) were compared with one another and with ensembles of 5 networks. To assess the generalizability of the approach, we applied the best-performing model to an independent Evaluation Cohort of 151 subjects. Agreement between manual and automated segmentations for identifying patients with large lesion volumes was calculated across multiple thresholds (21, 31, 51, and 70 cm). RESULTS: An ensemble of convolutional neural networks trained on DWI, ADC, and low b-value-weighted images produced the most accurate acute infarct segmentation over individual networks ( < .001). Automated volumes correlated with manually measured volumes (Spearman ρ = 0.91, < .001) for the independent cohort. For the task of identifying patients with large lesion volumes, agreement between manual outlines and automated outlines was high (Cohen κ, 0.86-0.90; < .001). CONCLUSIONS: Acute infarcts are more accurately segmented using ensembles of convolutional neural networks trained with multiparametric maps than by using a single model trained with a solo map. Automated lesion segmentation has high agreement with manual techniques for identifying patients with large lesion volumes.

OBJECTIVE: The role of heparin in acute ischemic stroke is controversial. We investigated the effect of heparin on ischemic lesion growth. METHODS: Data were analyzed on nonthrombolyzed ischemic stroke patients in whom diffusion-weighted imaging (DWI)/perfusion-weighted imaging (PWI) MRI was performed less than 12 hours of last known well and showed a PWI-DWI lesion mismatch, and who underwent follow-up neuroimaging at least 4 days after admission. Lesion growth was assessed by (1) absolute lesion growth and (2) percentage mismatch lost (PML). Univariate and multivariate regression analysis, and propensity score matching, were used to determine the effects of heparin on ischemic lesion growth. RESULTS: Of the 113 patients meeting study criteria, 59 received heparin within 24 hours. Heparin use was associated with ∼5-fold reductions in PML (3.5% versus 19.2%, P = .002) and absolute lesion growth (4.7 versus 20.5 mL, P = .009). In multivariate regression models, heparin independently predicted reduced PML (P = .04) and absolute lesion growth (P = .04) in the entire cohort, and in multiple subgroups (patients with and without proximal artery occlusion; DWI volume greater than 5 mL; cardio-embolic mechanism; DEFUSE-3 target mismatch). In propensity score matching analysis where patients were matched by admission NIHSS, DWI volume and proximal artery occlusion, heparin remained an independent predictor of PML (P = .048) and tended to predict absolute lesion growth (P = .06). Heparin treatment did not predict functional outcome at discharge or 90 days. CONCLUSION: Early heparin treatment in acute ischemic stroke patients with PWI-DWI mismatch attenuates ischemic lesion growth. Clinical trials with careful patient selection are warranted to investigate the potential ischemic protective effects of heparin.

Background and Purpose- We evaluated deep learning algorithms' segmentation of acute ischemic lesions on heterogeneous multi-center clinical diffusion-weighted magnetic resonance imaging (MRI) data sets and explored the potential role of this tool for phenotyping acute ischemic stroke. Methods- Ischemic stroke data sets from the MRI-GENIE (MRI-Genetics Interface Exploration) repository consisting of 12 international genetic research centers were retrospectively analyzed using an automated deep learning segmentation algorithm consisting of an ensemble of 3-dimensional convolutional neural networks. Three ensembles were trained using data from the following: (1) 267 patients from an independent single-center cohort, (2) 267 patients from MRI-GENIE, and (3) mixture of (1) and (2). The algorithms' performances were compared against manual outlines from a separate 383 patient subset from MRI-GENIE. Univariable and multivariable logistic regression with respect to demographics, stroke subtypes, and vascular risk factors were performed to identify phenotypes associated with large acute diffusion-weighted MRI volumes and greater stroke severity in 2770 MRI-GENIE patients. Stroke topography was investigated. Results- The ensemble consisting of a mixture of MRI-GENIE and single-center convolutional neural networks performed best. Subset analysis comparing automated and manual lesion volumes in 383 patients found excellent correlation (ρ=0.92; P<0.0001). Median (interquartile range) diffusion-weighted MRI lesion volumes from 2770 patients were 3.7 cm (0.9-16.6 cm). Patients with small artery occlusion stroke subtype had smaller lesion volumes ( P<0.0001) and different topography compared with other stroke subtypes. Conclusions- Automated accurate clinical diffusion-weighted MRI lesion segmentation using deep learning algorithms trained with multi-center and diverse data is feasible. Both lesion volume and topography can provide insight into stroke subtypes with sufficient sample size from big heterogeneous multi-center clinical imaging phenotype data sets.

BACKGROUND AND PURPOSE:

Background and aims: Assessment of ischemic lesions on computed tomography or MRI diffusion-weighted imaging (DWI) using the Alberta Stroke Program Early Computed Tomography Score (ASPECTS) is widely used to guide acute stroke treatment. However, it has never been defined how many voxels need to be affected to label a DWI-ASPECTS region ischemic. We aimed to assess the effect of various lesion load thresholds on DWI-ASPECTS and compare this automated analysis with visual rating. Materials and methods: We analyzed overlap of individual DWI lesions of 315 patients from the previously published predictive value of fluid-attenuated inversion recovery study with a probabilistic ASPECTS template derived from 221 CT images. We applied multiple lesion load thresholds per DWI-ASPECTS region (>0, >1, >10, and >20% in each DWI-ASPECTS region) to compute DWI-ASPECTS for each patient and compared the results to visual reading by an experienced stroke neurologist. Results: By visual rating, median ASPECTS was 9, 84 patients had a DWI-ASPECTS score ≤7. Mean DWI lesion volume was 22.1 (±35) ml. In contrast, by use of >0, >1-, >10-, and >20%-thresholds, median DWI-ASPECTS was 1, 5, 8, and 10; 97.1% (306), 72.7% (229), 41% (129), and 25.7% (81) had DWI-ASPECTS ≤7, respectively. Overall agreement between automated assessment and visual rating was low for every threshold used (>0%: κ = 0.020 1%: κ = 0.151; 10%: κ = 0.386; 20% κ = 0.381). Agreement for dichotomized DWI-ASPECTS ranged from fair to substantial (≤7: >10% κ = 0.48; >20% κ = 0.45; ≤5: >10% κ = 0.528; and >20% κ = 0.695). Conclusion: Overall agreement between automated and the standard used visual scoring is low regardless of the lesion load threshold used. However, dichotomized scoring achieved more comparable results. Varying lesion load thresholds had a critical impact on patient selection by ASPECTS. Of note, the relatively low lesion volume and lack of patients with large artery occlusion in our cohort may limit generalizability of these findings.

BACKGROUND AND PURPOSE: Tissue signatures from acute MR imaging of the brain may be able to categorize physiological status and thereby assist clinical decision making. We designed and analyzed statistical algorithms to evaluate the risk of infarction for each voxel of tissue using acute human functional MRI. METHODS: Diffusion-weighted MR images (DWI) and perfusion-weighted MR images (PWI) from acute stroke patients scanned within 12 hours of symptom onset were retrospectively studied and used to develop thresholding and generalized linear model (GLM) algorithms predicting tissue outcome as determined by follow-up MRI. The performances of the algorithms were evaluated for each patient by using receiver operating characteristic curves. RESULTS: At their optimal operating points, thresholding algorithms combining DWI and PWI provided 66% sensitivity and 83% specificity, and GLM algorithms combining DWI and PWI predicted with 66% sensitivity and 84% specificity voxels that proceeded to infarct. Thresholding algorithms that combined DWI and PWI provided significant improvement to algorithms that utilized DWI alone (P=0.02) but no significant improvement over algorithms utilizing PWI alone (P=0.21). GLM algorithms that combined DWI and PWI showed significant improvement over algorithms that used only DWI (P=0.02) or PWI (P=0.04). The performances of thresholding and GLM algorithms were comparable (P>0.2). CONCLUSIONS: Algorithms that combine acute DWI and PWI can assess the risk of infarction with higher specificity and sensitivity than algorithms that use DWI or PWI individually. Methods for quantitatively assessing the risk of infarction on a voxel-by-voxel basis show promise as techniques for investigating the natural spatial evolution of ischemic damage in humans.

PURPOSE: To (a) determine the optimal choice of a scalar metric of anisotropy and (b) determine by means of magnetic resonance imaging if changes in diffusion anisotropy occurred in acute human ischemic stroke. MATERIALS AND METHODS: The full diffusion tensor over the entire brain was measured. To optimize the choice of a scalar anisotropy metric, the performances of scalar indices in simulated models and in a healthy volunteer were analyzed. The anisotropy, trace apparent diffusion coefficient (ADC), and eigenvalues of the diffusion tensor in lesions and contralateral normal brain were compared in 50 patients with stroke. RESULTS: Changes in anisotropy in patients were quantified by using fractional anisotropy because it provided the best performance in terms of contrast-to-noise ratio as a function of signal-to-noise ratio in simulations. The anisotropy of ischemic white matter decreased (P = .01). Changes in anisotropy in ischemic gray matter were not significant (P = .63). The trace ADC decreased for ischemic gray matter and white matter (P < .001). The first and second eigenvalues decreased in both ischemic gray and ischemic white matter (P < .001). The third eigenvalue decreased in ischemic gray (P = .001) and white matter (P = .03). CONCLUSION: Gray matter is mildly anisotropic in normal and early ischemic states. However, early white matter ischemia is associated with not only changes in trace ADC values but also significant changes in the anisotropy, or shape, of the water self-diffusion tensor.

BACKGROUND: Diffusion-weighted MRI (DWI) represents a major advance in the early diagnosis of acute ischemic stroke. When abnormal in patients with stroke-like deficit, DWI usually establishes the presence and location of ischemic brain injury. However, this is not always the case. OBJECTIVE: To investigate patients with stroke-like deficits occurring without DWI abnormalities in brain regions clinically suspected to be responsible. METHODS: We identified 27 of 782 consecutive patients scanned when stroke-like neurologic deficits were still present and who had normal DWI in the brain region(s) clinically implicated. Based on all the clinical and radiologic data, we attempted to arrive at a pathophysiologic diagnosis in each. RESULTS: Best final diagnosis was a stroke mimic in 37% and a cerebral ischemic event in 63%. Stroke mimics (10 patients) included migraine, seizures, functional disorder, transient global amnesia, and brain tumor. The remaining patients were considered to have had cerebral ischemic events: lacunar syndrome (7 patients; 3 with infarcts demonstrated subsequently) and hemispheric cortical syndrome (10 patients; 5 with TIA, 2 with prolonged reversible deficits, 3 with infarction on follow-up imaging). In each of the latter three patients, the regions destined to infarct showed decreased perfusion on the initial hemodynamically weighted MRI (HWI). CONCLUSIONS: Normal DWI in patients with stroke-like deficits should stimulate a search for nonischemic cause of symptoms. However, more than one-half of such patients have an ischemic cause as the best clinical diagnosis. Small brainstem lacunar infarctions may escape detection. Concomitant HWI can identify some patients with brain ischemia that is symptomatic but not yet to the stage of causing DWI abnormality.

PURPOSE: To determine the probability that regions of decreased apparent diffusion coefficient (ADC) return to normal without persistent symptoms or T2 change and the settings in which these ADC reversals occur. MATERIALS AND METHODS: Three hundred magnetic resonance (MR) imaging studies were selected at random from a database of 7,147 examinations to determine the probability of a pathologically decreased ADC. In cases with decreased ADC, the clinical history was recorded and, if available, follow-up MR imaging findings were evaluated. Five cases of ADC reversal became known during the same period and were evaluated to determine the initial ADC decrease, clinical outcome, and findings at follow-up imaging. RESULTS: Findings in 116 of 300 MR imaging studies revealed regions of decreased ADC. In 49 of 116 studies, follow-up MR imaging examinations were performed at least 4 weeks after the onset of symptoms; ADC did not reverse. Five cases of ADC reversal were identified in the same period, giving an estimated 0.2%-0.4% probability of ADC reversal. Clinical settings were venous sinus thrombosis and seizure (n = 3), hemiplegic migraine (n = 1), and hyperacute arterial infarction (n = 1). Both white matter (n = 3) and gray matter (n = 3) regions were involved. CONCLUSION: Reversal of ADC lesions is rare, occurs in complicated clinical settings, and can involve white or gray matter.

The brains of six healthy volunteers were scanned with a full tensor diffusion MRI technique to study the effect of a high b value on diffusion-weighted images (DWIs). The b values ranged from 500 to 5000 s/mm(2). Isotropic DWIs, trace apparent diffusion coefficient (ADC) maps, and fractional anisotropy (FA) maps were created for each b value. As the b value increased, ADC decreased in both the gray and white matter. Furthermore, ADC of the white matter became lower than that of the gray matter, and, as a result, the white matter became brighter than the gray matter in the isotropic DWIs. Quantitative analysis showed that these changes were due to nonmonoexponential diffusion signal decay of the brain tissue, which was more prominent in white matter than in gray matter. There was no significant change in relation to the b value in the FA maps. High b value appears to have a dissociating effect on gray and white matter in DWIs.

BACKGROUND AND PURPOSE: The heterogeneity of microvascular flows is known to be an important determinant of the efficacy of oxygen delivery to tissue. Studies in animals have demonstrated decreased flow heterogeneity (FH) in states of decreased perfusion pressure. The purpose of the present study was to assess microvascular FH changes in acute stroke with use of a novel perfusion-weighted MRI technique and to evaluate the ability of combined diffusion-weighted MRI and FH measurements to predict final infarct size. METHODS: Cerebral blood flow, FH, and plasma mean transit time (MTT) were measured in 11 patients who presented with acute (<12 hours after symptom onset) stroke. Final infarct size was determined with follow-up MRI or CT scanning. RESULTS: In normal brain tissue, the distribution of relative flows was markedly skewed toward high capillary flow velocities. Within regions of decreased cerebral blood flow, plasma MTT was prolonged. Furthermore, subregions were identified with significant loss of the high-flow component of the flow distribution, thereby causing increased homogeneity of flow velocities. In parametric maps that quantify the acute deviation of FH from that of normal tissue, areas of extreme homogenization of capillary flows predicted final infarct size on follow-up scans of 10 of 11 patients. CONCLUSIONS: Flow heterogeneity and MTT can be rapidly assessed as part of a routine clinical MR examination and may provide a tool for planning of individual stroke treatment, as well as in targeting and evaluation of emerging therapeutic strategies.