Study on the impact of digital transformation on the innovation potential based on evidence from Chinese listed companies.

Digital transformation has emerged as a powerful force in reshaping the business landscape and enabling organizations to enhance their capabilities. One critical aspect of this change is how it impacts an enterprise's innovation ability. To explore this question, we select data regarding China's A-share listed enterprises from 2007 to 2021 as the research sample. We employ crawler technology to gather keywords related to "digital transformation" from annual reports, portraying detailed journeys of enterprises' digital transformation. Through descriptive statistics and multiple covariance tests, a linear relationship is established between digital transformation and innovation ability. Benchmark regression is conducted and a robustness test is utilized to determine the robustness of the benchmark regression. The mechanism, heterogeneity, and moderating effects of this study are also tested. The results reveal that digital transformation makes a significant positive contribution to the innovation capability of enterprises. Meanwhile, among different types of enterprises, the impact of digital transformation on enterprise innovation capability shows heterogeneity. In terms of the impact mechanism, digital transformation can enhance the innovation output of enterprises by reducing the agency cost and improving the risk-taking level of enterprises, so as to further improve the innovation capability of enterprises. The research results of this paper provide essential theoretical support for the digital transformation of enterprises and the government's formulation of enterprises' digitalization strategies. More profoundly, it provides significant reference for how to further promote the digital transformation of Chinese enterprises.

A Clinical and Imaging Fused Deep Learning Model Matches Expert Clinician Prediction of 90-Day Stroke Outcomes.

BACKGROUND AND PURPOSE: Predicting long-term clinical outcome in acute ischemic stroke is beneficial for prognosis, clinical trial design, resource management, and patient expectations. This study used a deep learning-based predictive model (DLPD) to predict 90-day mRS outcomes and compared its predictions with those made by physicians. MATERIALS AND METHODS: A previously developed DLPD that incorporated DWI and clinical data from the acute period was used to predict 90-day mRS outcomes in 80 consecutive patients with acute ischemic stroke from a single-center registry. We assessed the predictions of the model alongside those of 5 physicians (2 stroke neurologists and 3 neuroradiologists provided with the same imaging and clinical information). The primary analysis was the agreement between the ordinal mRS predictions of the model or physician and the ground truth using the Gwet Agreement Coefficient. We also evaluated the ability to identify unfavorable outcomes (mRS >2) using the area under the curve, sensitivity, and specificity. Noninferiority analyses were undertaken using limits of 0.1 for the Gwet Agreement Coefficient and 0.05 for the area under the curve analysis. The accuracy of prediction was also assessed using the mean absolute error for prediction, percentage of predictions ±1 categories away from the ground truth (±1 accuracy [ACC]), and percentage of exact predictions (ACC). RESULTS: To predict the specific mRS score, the DLPD yielded a Gwet Agreement Coefficient score of 0.79 (95% CI, 0.71-0.86), surpassing the physicians' score of 0.76 (95% CI, 0.67-0.84), and was noninferior to the readers (P < .001). For identifying unfavorable outcome, the model achieved an area under the curve of 0.81 (95% CI, 0.72-0.89), again noninferior to the readers' area under the curve of 0.79 (95% CI, 0.69-0.87) (P < .005). The mean absolute error, ±1ACC, and ACC were 0.89, 81%, and 36% for the DLPD. CONCLUSIONS: A deep learning method using acute clinical and imaging data for long-term functional outcome prediction in patients with acute ischemic stroke, the DLPD, was noninferior to that of clinical readers.

Random expert sampling for deep learning segmentation of acute ischemic stroke on non-contrast CT.

BACKGROUND: Outlining acutely infarcted tissue on non-contrast CT is a challenging task for which human inter-reader agreement is limited. We explored two different methods for training a supervised deep learning algorithm: one that used a segmentation defined by majority vote among experts and another that trained randomly on separate individual expert segmentations. METHODS: The data set consisted of 260 non-contrast CT studies in 233 patients with acute ischemic stroke recruited from the multicenter DEFUSE 3 (Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3) trial. Additional external validation was performed using 33 patients with matched stroke onset times from the University Hospital Lausanne. A benchmark U-Net was trained on the reference annotations of three experienced neuroradiologists to segment ischemic brain tissue using majority vote and random expert sampling training schemes. The median of volume, overlap, and distance segmentation metrics were determined for agreement in lesion segmentations between (1) three experts, (2) the majority model and each expert, and (3) the random model and each expert. The two sided Wilcoxon signed rank test was used to compare performances (1) to 2) and (1) to (3). We further compared volumes with the 24 hour follow-up diffusion weighted imaging (DWI, final infarct core) and correlations with clinical outcome (modified Rankin Scale (mRS) at 90 days) with the Spearman method. RESULTS: The random model outperformed the inter-expert agreement ((1) to (2)) and the majority model ((1) to (3)) (dice 0.51±0.04 vs 0.36±0.05 (P<0.0001) vs 0.45±0.05 (P<0.0001)). The random model predicted volume correlated with clinical outcome (0.19, P<0.05), whereas the median expert volume and majority model volume did not. There was no significant difference when comparing the volume correlations between random model, median expert volume, and majority model to 24 hour follow-up DWI volume (P>0.05, n=51). CONCLUSION: The random model for ischemic injury delineation on non-contrast CT surpassed the inter-expert agreement ((1) to (2)) and the performance of the majority model ((1) to (3)). We showed that the random model volumetric measures of the model were consistent with 24 hour follow-up DWI.

White matter hyperintensities in cholinergic pathways correlates of cognitive impairment in moyamoya disease.

OBJECTIVE: To investigate the effect of cholinergic pathways damage caused by white matter hyperintensities (WMHs) on cognitive function in moyamoya disease (MMD). METHODS: We included 62 patients with MMD from a prospectively enrolled cohort. We evaluated the burden of cholinergic pathways damage caused by WMHs using the Cholinergic Pathways Hyperintensities Scale (CHIPS). Cognitive function was evaluated with the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA). Cognitive impairment was determined according to the cut-off of MMSE and education. Multivariate linear and logistic regression models were used to analyze whether CHIPS was independently associated with cognition. Receiver operating characteristic curve analysis was performed to identify the ability of CHIPS in discriminating cognitive impairment and normal cognition. RESULTS: CHIPS was associated with both MMSE and MoCA (ß = - 0.601 and ß = - 0.672, both p < 0.001). After correcting age, sex, education, volumes of limbic areas, and other factors, CHIPS remained to be independently associated with both MMSE and MoCA (ß = - 0.388 and ß = - 0.334, both p < 0.001). In the logistic regression, only CHIPS was associated with cognitive impairment (odds ratio = 1.431, 95% confidence interval = 1.103 to 1.856, p = 0.007). The optimal cut-off of CHIPS score was 10, yielding a sensitivity of 87.5% and a specificity of 78.3% in identifying MMD patients with cognitive impairment. CONCLUSIONS: The damage of cholinergic pathways caused by WMHs plays an independent effect on cognition and CHIPS could be a useful method in identifying MMD patients likely to be cognitive impairment. CLINICAL RELEVANCE STATEMENT: This study shows that Cholinergic Pathways Hyperintensities Scale (CHIPS) could be a simple and reliable method in identifying cognitive impairment for patients with moyamoya disease. CHIPS could be helpful in clinical practice, such as guiding treatment decisions and predicting outcome. KEY POINTS: • Cholinergic Pathways Hyperintensities Scale was significantly associated with cognitive screening tests in patients with moyamoya disease. • Cholinergic Pathways Hyperintensities Scale plays an independent effect on cognitive impairment in patients with moyamoya disease. • Cholinergic Pathways Hyperintensities Scale shows higher accuracy than education, volumes of limbic areas, and sex in identifying cognitive impairment in moyamoya disease.

Highly Sensitive, Portable Detection System for Multiplex Chemiluminescence Analysis.

Chemiluminescence (CL) has emerged as a critical tool for the sensing and quantification of various bioanalytes in virtually all clinical fields. However, the rapid nature of many CL reactions raises challenges for typical low-cost optical sensors such as cameras to achieve accurate and sensitive detection. Meanwhile, classic sensors such as photomultiplier tubes are highly sensitive but lack spatial multiplexing capabilities and are generally not suited for point-of-care applications outside a standard laboratory setting. To address this issue, in this paper, a miniaturized and versatile silicon-photomultiplier-based fiber-integrated CL device (SFCD) was designed for sensitive multiplex CL detection. The SFCD comprises a silicon photomultiplier array coupled to an array of high numerical aperture plastic optical fibers to achieve 16-plex detection. The optical fibers ensure efficient light collection while allowing the fixed detector to be mated with diverse sample geometries (e.g., circular or grid), simply by adjusting the fiber configuration. In a head-to-head comparison with a lens-based camera system featuring a cooled detector, the SFCD achieved a 14-fold improved limit of detection in both direct and enzyme-mediated CL reactions. The SFCD also features improved compactness and lower cost, as well as faster temporal resolution compared with camera-based systems while preserving spatial multiplexing and good environmental robustness. Thus, the SFCD has excellent potential for point-of-care biosensing applications.

Functional Outcome Prediction in Acute Ischemic Stroke Using a Fused Imaging and Clinical Deep Learning Model.

BACKGROUND: Predicting long-term clinical outcome based on the early acute ischemic stroke information is valuable for prognostication, resource management, clinical trials, and patient expectations. Current methods require subjective decisions about which imaging features to assess and may require time-consuming postprocessing. This study's goal was to predict ordinal 90-day modified Rankin Scale (mRS) score in acute ischemic stroke patients by fusing a Deep Learning model of diffusion-weighted imaging images and clinical information from the acute period. METHODS: A total of 640 acute ischemic stroke patients who underwent magnetic resonance imaging within 1 to 7 days poststroke and had 90-day mRS follow-up data were randomly divided into 70% (n=448) for model training, 15% (n=96) for validation, and 15% (n=96) for internal testing. Additionally, external testing on a cohort from Lausanne University Hospital (n=280) was performed to further evaluate model generalization. Accuracy for ordinal mRS, accuracy within ±1 mRS category, mean absolute prediction error, and determination of unfavorable outcome (mRS score >2) were evaluated for clinical only, imaging only, and 2 fused clinical-imaging models. RESULTS: The fused models demonstrated superior performance in predicting ordinal mRS score and unfavorable outcome in both internal and external test cohorts when compared with the clinical and imaging models. For the internal test cohort, the top fused model had the highest area under the curve of 0.92 for unfavorable outcome prediction and the lowest mean absolute error (0.96 [95% CI, 0.77-1.16]), with the highest proportion of mRS score predictions within ±1 category (79% [95% CI, 71%-88%]). On the external Lausanne University Hospital cohort, the best fused model had an area under the curve of 0.90 for unfavorable outcome prediction and outperformed other models with an mean absolute error of 0.90 (95% CI, 0.79-1.01), and the highest percentage of mRS score predictions within ±1 category (83% [95% CI, 78%-87%]). CONCLUSIONS: A Deep Learning-based imaging model fused with clinical variables can be used to predict 90-day stroke outcome with reduced subjectivity and user burden.

Automated detection of arterial landmarks and vascular occlusions in patients with acute stroke receiving digital subtraction angiography using deep learning.

BACKGROUND: Digital subtraction angiography (DSA) is the gold-standard method of assessing arterial blood flow and blockages prior to endovascular thrombectomy. OBJECTIVE: To detect anatomical features and arterial occlusions with DSA using artificial intelligence techniques. METHODS: We included 82 patients with acute ischemic stroke who underwent DSA imaging and whose carotid terminus was visible in at least one run. Two neurointerventionalists labeled the carotid location (when visible) and vascular occlusions on 382 total individual DSA runs. For detecting the carotid terminus, positive and negative image patches (either containing or not containing the internal carotid artery terminus) were extracted in a 1:1 ratio. Two convolutional neural network architectures (ResNet-50 pretrained on ImageNet and ResNet-50 trained from scratch) were evaluated. Area under the curve (AUC) of the receiver operating characteristic and pixel distance from the ground truth were calculated. The same training and analysis methods were used for detecting arterial occlusions. RESULTS: The ResNet-50 trained from scratch most accurately detected the carotid terminus (AUC 0.998 (95% CI 0.997 to 0.999), p<0.00001) and arterial occlusions (AUC 0.973 (95% CI 0.971 to 0.975), p<0.0001). Average pixel distances from ground truth for carotid terminus and occlusion localization were 63±45 and 98±84, corresponding to approximately 1.26±0.90 cm and 1.96±1.68 cm for a standard angiographic field-of-view. CONCLUSION: These results may serve as an unbiased standard for clinical stroke trials, as optimal standardization would be useful for core laboratories in endovascular thrombectomy studies, and also expedite decision-making during DSA-based procedures.

Predicting Hypoperfusion Lesion and Target Mismatch in Stroke from Diffusion-weighted MRI Using Deep Learning.

Background Perfusion imaging is important to identify a target mismatch in stroke but requires contrast agents and postprocessing software. Purpose To use a deep learning model to predict the hypoperfusion lesion in stroke and identify patients with a target mismatch profile from diffusion-weighted imaging (DWI) and clinical information alone, using perfusion MRI as the reference standard. Materials and Methods Imaging data sets of patients with acute ischemic stroke with baseline perfusion MRI and DWI were retrospectively reviewed from multicenter data available from 2008 to 2019 (Imaging Collaterals in Acute Stroke, Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution 2, and University of California, Los Angeles stroke registry). For perfusion MRI, rapid processing of perfusion and diffusion software automatically segmented the hypoperfusion lesion (time to maximum, ≥6 seconds) and ischemic core (apparent diffusion coefficient [ADC], ≤620 × 10-6 mm2/sec). A three-dimensional U-Net deep learning model was trained using baseline DWI, ADC, National Institutes of Health Stroke Scale score, and stroke symptom sidedness as inputs, with the union of hypoperfusion and ischemic core segmentation serving as the ground truth. Model performance was evaluated using the Dice score coefficient (DSC). Target mismatch classification based on the model was compared with that of the clinical-DWI mismatch approach defined by the DAWN trial by using the McNemar test. Results Overall, 413 patients (mean age, 67 years ± 15 [SD]; 207 men) were included for model development and primary analysis using fivefold cross-validation (247, 83, and 83 patients in the training, validation, and test sets, respectively, for each fold). The model predicted the hypoperfusion lesion with a median DSC of 0.61 (IQR, 0.45-0.71). The model identified patients with target mismatch with a sensitivity of 90% (254 of 283; 95% CI: 86, 93) and specificity of 77% (100 of 130; 95% CI: 69, 83) compared with the clinical-DWI mismatch sensitivity of 50% (140 of 281; 95% CI: 44, 56) and specificity of 89% (116 of 130; 95% CI: 83, 94) (P < .001 for all). Conclusion A three-dimensional U-Net deep learning model predicted the hypoperfusion lesion from diffusion-weighted imaging (DWI) and clinical information and identified patients with a target mismatch profile with higher sensitivity than the clinical-DWI mismatch approach. ClinicalTrials.gov registration nos. NCT02225730, NCT01349946, NCT02586415 © RSNA, 2022 Supplemental material is available for this article. See also the editorial by Kallmes and Rabinstein in this issue.

Predicting final ischemic stroke lesions from initial diffusion-weighted images using a deep neural network.

BACKGROUND: For prognosis of stroke, measurement of the diffusion-perfusion mismatch is a common practice for estimating tissue at risk of infarction in the absence of timely reperfusion. However, perfusion-weighted imaging (PWI) adds time and expense to the acute stroke imaging workup. We explored whether a deep convolutional neural network (DCNN) model trained with diffusion-weighted imaging obtained at admission could predict final infarct volume and location in acute stroke patients. METHODS: In 445 patients, we trained and validated an attention-gated (AG) DCNN to predict final infarcts as delineated on follow-up studies obtained 3 to 7 days after stroke. The input channels consisted of MR diffusion-weighted imaging (DWI), apparent diffusion coefficients (ADC) maps, and thresholded ADC maps with values less than 620 × 10-6 mm2/s, while the output was a voxel-by-voxel probability map of tissue infarction. We evaluated performance of the model using the area under the receiver-operator characteristic curve (AUC), the Dice similarity coefficient (DSC), absolute lesion volume error, and the concordance correlation coefficient (ρc) of the predicted and true infarct volumes. RESULTS: The model obtained a median AUC of 0.91 (IQR: 0.84-0.96). After thresholding at an infarction probability of 0.5, the median sensitivity and specificity were 0.60 (IQR: 0.16-0.84) and 0.97 (IQR: 0.93-0.99), respectively, while the median DSC and absolute volume error were 0.50 (IQR: 0.17-0.66) and 27 ml (IQR: 7-60 ml), respectively. The model's predicted lesion volumes showed high correlation with ground truth volumes (ρc = 0.73, p < 0.01). CONCLUSION: An AG-DCNN using diffusion information alone upon admission was able to predict infarct volumes at 3-7 days after stroke onset with comparable accuracy to models that consider both DWI and PWI. This may enable treatment decisions to be made with shorter stroke imaging protocols.

The efficacy and safety of acupuncture in the treatment of taste disorders after rehabilitation from COVID-19: A protocol for systematic review and meta-analysis.

BACKGROUND: Since the outbreak of the new crown pneumonia epidemic, although it has had a serious impact on people's lives and health in itself, the sequelae that accompany coronavirus disease 2019 (COVID-19) have also had a serious impact on people's mental health and quality of life. Taste disorder (TD) is one of the sequelae of COVID-19. Patients may experience reduced or even absent taste sensation, which seriously affects their real life. The efficacy of acupuncture in the treatment of taste disorders has been well reported, but there is a lack of evidence-based medical evidence. Therefore, this study set out to systematically evaluate the efficacy and safety of acupuncture in the treatment of post-COVID-19 taste disorder. METHODS: According to the retrieval strategies, randomized controlled trials on the acupuncture for COVID-19 TD were obtained from Cochrane Central Register of Controlled Trials, Embase, PubMed, Web of Science, the Chinese National Knowledge Infrastructure, the Chinese Biomedical Literature Database, the Chinese Scientific Journal Database and the Wanfang Database, regardless of publication date, or language. Studies were screened based on inclusion and exclusion criteria, and the Cochrane risk bias assessment tool was used to evaluate the quality of the studies. The meta-analysis was performed using Review Manager (RevMan 5.4) and StataSE 15.0 software. Ultimately, the evidentiary grade for the results will be evaluated. This systematic evaluation protocol is registered in PROSPERO under the registration ID CRD42022364653. RESULTS: The results of this meta-analysis will be submitted to a peer-reviewed journal for publication. CONCLUSION: This meta-analysis will evaluate the effect of acupuncture and moxibustion on TD caused by sequelae of COVID-19, providing evidence as to the treatment in these patients.

Microfluidic Paper-Based Preconcentration and Retrieval for Rapid Ribonucleic Acid Biomarker Detection and Visualization.

Microfluidic paper-based analytical devices (µPADs) have attracted significant attention in the field of point-of-care (POC) diagnostics. However, the heterogeneous structure of the paper often impairs the limit of detection (LOD) for low-abundance targets when those targets are directly analyzed. One viable solution to bypass this limitation is to elevate the target concentration above the LOD on-site to reach a valid readout. Here, we developed a 3D µPADs preconcentrator (3D-µP2) to increase sample concentration by electrokinetic trapping and demonstrated its application in increasing the LOD of a downstream colorimetric assay. The three-dimensional (3D) structure of this device was composed of a loading pad, a vertical fluid path formed by stacked absorbent pads, and an ion-selective membrane of PEDOT:PSS. This novel design facilitates fast preconcentration, high capacity in sample processing, and easy target retrieval. The concentration of an exemplary target, a single-stranded DNA sequence, was increased up to 170-fold within 80 s. The LOD of the colorimetric assay to verify the DNA target was increased 3 orders of magnitude with a preconcentrated sample compared to the control. The device and its analysis equipment used in this study were all cheap and portable. Thus, the 3D-µP2 can be a powerful POC tool for sample pretreatment in resource-limited areas.

Reproducibility of cerebrovascular reactivity measurements: A systematic review of neuroimaging techniques.

Cerebrovascular reactivity (CVR), the capacity of the brain to increase cerebral blood flow (CBF) to meet changes in physiological demand, is an important biomarker to evaluate brain health. Typically, this brain "stress test" is performed by using a medical imaging modality to measure the CBF change between two states: at baseline and after vasodilation. However, since there are many imaging modalities and many ways to augment CBF, a wide range of CVR values have been reported. An understanding of CVR reproducibility is critical to determine the most reliable methods to measure CVR as a clinical biomarker. This review focuses on CVR reproducibility studies using neuroimaging techniques in 32 articles comprising 427 total subjects. The literature search was performed in PubMed, Embase, and Scopus. The review was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). We identified 5 factors of the experimental subjects (such as sex, blood characteristics, and smoking) and 9 factors of the measuring technique (such as the imaging modality, the type of the vasodilator, and the quantification method) that have strong effects on CVR reproducibility. Based on this review, we recommend several best practices to improve the reproducibility of CVR quantification in neuroimaging studies.

Assessment of Therapeutic Response to Statin Therapy in Patients With Intracranial or Extracranial Carotid Atherosclerosis by Vessel Wall MRI: A Systematic Review and Updated Meta-Analysis.

Background and Aims: Statin therapy is an essential component of cardiovascular preventive care. In recent years, various vessel wall MRI (VW-MRI) techniques have been used to monitor atherosclerosis progression or regression in patients with extracranial or intracranial large-artery atherosclerosis. We aimed to perform a systematic review and meta-analysis on the effects of statin therapy on plaque evolution as assessed by VW-MRI. Materials and Methods: Prospective studies investigating carotid and intracranial atherosclerotic plaques in patients on statin therapy monitored by serial VW-MRI were systematically identified in the literature. The plaque burden and lipid-rich necrotic core (LRNC) volume of carotid plaque and the imaging features of intracranial plaques were extracted and summarized. For studies investigating carotid artery wall volume and LRNC volume, combined estimates were derived by meta-analysis. Results: The study identified 21 studies of carotid plaque and two studies of intracranial plaque. While 16 studies investigating carotid plaques that included 780 patients by High-resolution VW-MRI were included in the meta-analysis. There was no significant change in carotid wall volume from baseline to 12 months. A significant change in LRNC volume was observed at > 12 months compared with baseline (Effect = -10.69, 95% CI = -19.11, -2.28, P < 0.01), while no significant change in LRNC volume at 3-6 months or 7-12 months after statin therapy initiation in 6 studies. Increases in fibrous tissue and calcium and reduction in neovascularization density of the plaque were seen in 2/3 studies (including 48/59 patients), 1/3 studies (including 17/54 patients), and 2/2 studies (including 71 patients) after statin therapy, respectively. Two studies with 257 patients in intracranial atherosclerosis showed that statins could effectively decrease wall volume and plaque enhancement volume. Conclusions: Collective data indicated that statins could potentially stabilize carotid plaques by significantly reducing LRNC with 1 year of therapy as shown on serial carotid VW-MRI. There was no significant decrease in wall volume, which nonetheless indicated that plaque composition changes might be more sensitive to response monitoring than wall volume. It is likely that more sensitive, clinically relevant, and preferably quantitative indicators of therapeutic effects on intracranial vessel plaque morphology will be developed in the future.

Improving Ischemic Stroke Care With MRI and Deep Learning Artificial Intelligence.

ABSTRACT: Advanced magnetic resonance imaging has been used as selection criteria for both acute ischemic stroke treatment and secondary prevention. The use of artificial intelligence, and in particular, deep learning, to synthesize large amounts of data and to understand better how clinical and imaging data can be leveraged to improve stroke care promises a new era of stroke care. In this article, we review common deep learning model structures for stroke imaging, evaluation metrics for model performance, and studies that investigated deep learning application in acute ischemic stroke care and secondary prevention.

The bright vessel sign on arterial spin labeling MRI for heralding and localizing large vessel occlusions.

BACKGROUND AND PURPOSE: The significance of a bright vessel sign (BVS) at the site of a large vessel occlusion (LVO) on MR arterial spin labeling (ASL) sequence is not widely reported. We compared the utility of the ASL BVS to the gradient echo (GRE) susceptibility vessel sign (SVS) in heralding and localizing LVOs in a large cohort; most underwent digital subtraction angiography (DSA) and endovascular therapy for acute stroke. METHODS: A total of 171 patients with large hemispheric stroke symptoms had baseline and follow-up MRIs with ASL, GRE, and MR angiogram (MRA). Scans were evaluated for (1) presence versus absence and (2) location of ASL BVS and GRE SVS. For patients who underwent DSA, data comparing presence and location of ASL BVS and GRE SVS to occlusions found on angiography, as well as resolution of the signs after successful recanalization, were also evaluated. RESULTS: Compared to MRA, the sensitivity of the ASL BVS for an LVO was .83, significantly better than .67 for GRE SVS (p = .001). Localization of vessel occlusion was correct 60.4% of the time by ASL compared to 64.4% by GRE (p = .502). For the 107 patients who underwent DSA, the sensitivity of ASL BVS was .80 compared to .64 for GRE SVS (p = .009). Localization of LVO found on DSA was correct 63.5% of the time by ASL BVS compared to 72.9% by GRE SVS (p = .251). CONCLUSION: ASL BVS is significantly more sensitive than GRE SVS for identification of LVO on both MRA and DSA.

Early NT-ProBNP (N-Terminal Probrain Natriuretic Peptide) Elevation Predicts Malignant Edema and Death After Reperfusion Therapy in Acute Ischemic Stroke Patients.

BACKGROUND AND PURPOSE: We aimed to investigate the relationship between early NT-proBNP (N-terminal probrain natriuretic peptide) and all-cause death in patients receiving reperfusion therapy, including intravenous thrombolysis and endovascular thrombectomy (EVT). METHODS: This study included 1039 acute ischemic stroke patients with early NT-proBNP data at 2 hours after the beginning of alteplase infusion for those with intravenous thrombolysis only or immediately at the end of EVT for those with EVT. We performed natural log transformation for NT-proBNP (Ln(NT-proBNP)). Malignant brain edema was ascertained by using the SITS-MOST (Safe Implementation of Thrombolysis in Stroke-Monitoring Study) criteria. RESULTS: Median serum NT-proBNP level was 349 pg/mL (interquartile range, 89-1250 pg/mL). One hundred twenty-one (11.6%) patients died. Malignant edema was observed in 78 (7.5%) patients. Ln(NT-proBNP) was independently associated with 3-month mortality in patients with intravenous thrombolysis only (odds ratio, 1.465 [95% CI, 1.169-1.836]; P=0.001) and in those receiving EVT (odds ratio, 1.563 [95% CI, 1.139-2.145]; P=0.006). The elevation of Ln(NT-proBNP) was also independently associated with malignant edema in patients with intravenous thrombolysis only (odds ratio, 1.334 [95% CI, 1.020-1.745]; P=0.036), and in those with EVT (odds ratio, 1.455 [95% CI, 1.057-2.003]; P=0.022). CONCLUSIONS: An early increase in NT-proBNP levels was related to malignant edema and stroke mortality after reperfusion therapy.

Absent filling of the superficial middle cerebral vein is associated with reperfusion but not parenchymal hematoma in stroke patients undergoing thrombectomy: an observational study.

BACKGROUND: Parenchymal hematoma (PH) is the most feared complication of reperfusion therapy after stroke. The opacification of the superficial middle cerebral vein (SMCV) on computed tomography perfusion (CTP) has been associated with poor functional outcomes after stroke, while its association with PH has not been verified for acute stroke patients undergoing thrombectomy. METHODS: Consecutive patients with acute anterior large artery occlusion (LAO) who received thrombectomy were retrospectively enrolled between May 2018 and May 2019. Absent filing of the SMCV (SMCV-) on CTP-derived CT angiography was defined as no contrast filling of the SMCV across the whole venous phase in the ischemic hemisphere, while SMCV+ was defined as the presence of contrast filling of the SMCV at any time point of the venous phase. RESULTS: A total of 52 patients were enrolled in the study, and 15 patients (28.8%) developed a PH within 48 hours after thrombectomy. SMCV- was not associated with PH in both the univariate and multivariate logistic regression analyses (all P>0.05), but was an independent risk factor for reperfusion [modified thrombolysis in cerebral infarction score of 2b-3; odds ratio (OR) =0.172, 95% confidence interval (CI): 0.031-0.960, P=0.045]. Reperfusion was associated with a reduced risk of PH (OR =0.110, 95% CI: 0.013-0.913, P=0.041). However, in a subgroup analysis of patients who had reperfusion, the SMCV- group had a higher rate of PH than the SMCV+ group (40.0% vs. 13.8%, P=0.049). CONCLUSIONS: In patients who received thrombectomy, SMCV- did not predict PH, but was a risk factor for reperfusion. Although reperfusion was a protective factor for PH, the SMCV- group was still at a higher risk of PH compared with the SMCV+ group when reperfusion was successfully achieved.

Absent Contrast Filling of Ipsilateral Superficial Middle Cerebral Vein Predicts Midline Shift in Acute Middle Cerebral Artery Occlusion.

Background and purpose: Midline shift is a life-threatening complication of acute large artery occlusion (LAO). The value of superficial middle cerebral vein (SMCV) for predicting midline shift is currently unclear for patients with acute LAO. Methods: Consecutive acute LAO (middle cerebral artery M1 ± intracranial internal carotid artery) patients between March 2018 and May 2019 were included. Absent filling of ipsilateral cortical vein (marked as SMCV-) was defined as no contrast filling into the vein across the whole venous phase of four-dimensional computed tomography (CT) angiography derived from CT perfusion in the ischemic hemisphere. Results: In the total of 81 patients, 31 (38.4%) were identified as SMCV-. SMCV- independently predicted midline shift, with sensitivity of 87.5% and specificity of 82.5%. Receiver operating characteristic analysis showed that including SMCV- as a predictor in addition to baseline ischemic core volume significantly increased the area under the curve in predicting midline shift (SMCV- with baseline ischemic core volume vs. baseline ischemic core volume: AUC = 0.903 vs. 0.841, Z = 2.451, P = 0.014). Conclusion: In acute LAO patients, the presence of SMCV- was a sensitive and specific imaging marker for midline shift. SMCV- had supplementary value to baseline ischemic core volume in predicting midline shift.

Use of Deep Learning to Predict Final Ischemic Stroke Lesions From Initial Magnetic Resonance Imaging.

Importance: Predicting infarct size and location is important for decision-making and prognosis in patients with acute stroke. Objectives: To determine whether a deep learning model can predict final infarct lesions using magnetic resonance images (MRIs) acquired at initial presentation (baseline) and to compare the model with current clinical prediction methods. Design, Setting, and Participants: In this multicenter prognostic study, a specific type of neural network for image segmentation (U-net) was trained, validated, and tested using patients from the Imaging Collaterals in Acute Stroke (iCAS) study from April 14, 2014, to April 15, 2018, and the Diffusion Weighted Imaging Evaluation for Understanding Stroke Evolution Study-2 (DEFUSE-2) study from July 14, 2008, to September 17, 2011 (reported in October 2012). Patients underwent baseline perfusion-weighted and diffusion-weighted imaging and MRI at 3 to 7 days after baseline. Patients were grouped into unknown, minimal, partial, and major reperfusion status based on 24-hour imaging results. Baseline images acquired at presentation were inputs, and the final true infarct lesion at 3 to 7 days was considered the ground truth for the model. The model calculated the probability of infarction for every voxel, which can be thresholded to produce a prediction. Data were analyzed from July 1, 2018, to March 7, 2019. Main Outcomes and Measures: Area under the curve, Dice score coefficient (DSC) (a metric from 0-1 indicating the extent of overlap between the prediction and the ground truth; a DSC of ≥0.5 represents significant overlap), and volume error. Current clinical methods were compared with model performance in subgroups of patients with minimal or major reperfusion. Results: Among the 182 patients included in the model (97 women [53.3%]; mean [SD] age, 65 [16] years), the deep learning model achieved a median area under the curve of 0.92 (interquartile range [IQR], 0.87-0.96), DSC of 0.53 (IQR, 0.31-0.68), and volume error of 9 (IQR, -14 to 29) mL. In subgroups with minimal (DSC, 0.58 [IQR, 0.31-0.67] vs 0.55 [IQR, 0.40-0.65]; P = .37) or major (DSC, 0.48 [IQR, 0.29-0.65] vs 0.45 [IQR, 0.15-0.54]; P = .002) reperfusion for which comparison with existing clinical methods was possible, the deep learning model had comparable or better performance. Conclusions and Relevance: The deep learning model appears to have successfully predicted infarct lesions from baseline imaging without reperfusion information and achieved comparable performance to existing clinical methods. Predicting the subacute infarct lesion may help clinicians prepare for decompression treatment and aid in patient selection for neuroprotective clinical trials.