Systematic Review of Artificial Intelligence for Abnormality Detection in High-volume Neuroimaging and Subgroup Meta-analysis for Intracranial Hemorrhage Detection
Clinical neuroradiology. 2023;:1-14
PURPOSE Most studies evaluating artificial intelligence (AI) models that detect abnormalities in neuroimaging are either tested on unrepresentative patient cohorts or are insufficiently well-validated, leading to poor generalisability to real-world tasks. The aim was to determine the diagnostic test accuracy and summarise the evidence supporting the use of AI models performing first-line, high-volume neuroimaging tasks. METHODS Medline, Embase, Cochrane library and Web of Science were searched until September 2021 for studies that temporally or externally validated AI capable of detecting abnormalities in first-line computed tomography (CT) or magnetic resonance (MR) neuroimaging. A bivariate random effects model was used for meta-analysis where appropriate. This study was registered on PROSPERO as CRD42021269563. RESULTS Out of 42,870 records screened, and 5734 potentially eligible full texts, only 16 studies were eligible for inclusion. Included studies were not compromised by unrepresentative datasets or inadequate validation methodology. Direct comparison with radiologists was available in 4/16 studies and 15/16 had a high risk of bias. Meta-analysis was only suitable for intracranial hemorrhage detection in CT imaging (10/16 studies), where AI systems had a pooled sensitivity and specificity 0.90 (95% confidence interval [CI] 0.85-0.94) and 0.90 (95% CI 0.83-0.95), respectively. Other AI studies using CT and MRI detected target conditions other than hemorrhage (2/16), or multiple target conditions (4/16). Only 3/16 studies implemented AI in clinical pathways, either for pre-read triage or as post-read discrepancy identifiers. CONCLUSION The paucity of eligible studies reflects that most abnormality detection AI studies were not adequately validated in representative clinical cohorts. The few studies describing how abnormality detection AI could impact patients and clinicians did not explore the full ramifications of clinical implementation.
A systematic review on viscoelastic testing in subarachnoid haemorrhage patients
World Neurosurgery. 2023
OBJECTIVES Bleeding and thromboembolic complications frequently occur following subarachnoid haemorrhage (SAH) and substantially contribute to poor outcome. Viscoelastic testing could be used for detection of coagulopathies following SAH. This review summarizes literature on the utility of viscoelastic testing to detect coagulopathy in SAH patients and explores whether viscoelastic parameters are associated with SAH-related complications and clinical outcome. MATERIALS AND METHODS PUBMED, EMBASE and Google Scholar were systematically searched on August 18(th), 2022. Two authors independently selected studies which performed viscoelastic testing in SAH patients and assessed the quality of studies using the Newcastle Ottawa Scale or a previously published framework for quality assessment. Data was meta-analysed if methodologically possible. RESULTS The search yielded 19 studies (1160 SAH patients). Pooling of data including all relevant studies was not possible for any of the outcome measurements due to methodological differences. Thirteen of 19 studies evaluated the association of coagulation profiles and SAH, of which 11 studies showed a hypercoagulable profile. Rebleeding was associated with platelet dysfunction, deep venous thrombosis was associated with faster clot initiation and both delayed cerebral ischemia and poor outcome were associated with increased clot strength. CONCLUSIONS This explorative review shows that SAH patients frequently have a hypercoagulable profile. TEG- and ROTEM-parameters are associated with rebleeding, delayed cerebral ischemia, deep venous thrombosis and poor clinical outcome after SAH, however more research on the subject is needed. Future studies should focus on determining the optimal time frame and cut-off values for TEG or ROTEM to predict these complications.
Perihematomal Edema and Clinical Outcome After Intracerebral Hemorrhage: A Systematic Review and Meta-Analysis
Neurocritical care. 2022
BACKGROUND Perihematomal edema (PHE) has been proposed as a radiological marker of secondary injury and therapeutic target in intracerebral hemorrhage (ICH). We conducted a systematic review and meta-analysis to assess the prognostic impact of PHE on functional outcome and mortality in patients with ICH. METHODS We searched major databases through December 2020 using predefined keywords. Any study using logistic regression to examine the association between PHE or its growth and functional outcome was included. We examined the overall pooled effect and conducted secondary analyses to explore the impact of individual PHE measures on various outcomes separately. Study quality was assessed by three independent raters using the Newcastle-Ottawa Scale. Odds ratios (per 1-unit increase in PHE) and their confidence intervals (CIs) were log transformed and entered into a DerSimonian-Laird random-effects meta-analysis to obtain pooled estimates of the effect. RESULTS Twenty studies (n = 6633 patients) were included in the analysis. The pooled effect size for overall outcome was 1.05 (95% CI 1.02-1.08; p < 0.00). For the following secondary analyses, the effect size was weak: mortality (1.01; 95% CI 0.90-1.14), functional outcome (1.04; 95% CI 1.02-1.07), both 90-day (1.06; 95% CI 1.02-1.11), and in-hospital assessments (1.04; 95% CI 1.00-1.08). The effect sizes for PHE volume and PHE growth were 1.04 (95% CI 1.01-1.07) and 1.14 (95% CI 1.04-1.25), respectively. Heterogeneity across studies was substantial except for PHE growth. CONCLUSIONS This meta-analysis demonstrates that PHE volume within the first 72 h after ictus has a weak effect on functional outcome and mortality after ICH, whereas PHE growth might have a slightly larger impact during this time frame. Definitive conclusions are limited by the large variability of PHE measures, heterogeneity, and different evaluation time points between studies.
Accuracy of artificial intelligence for the detection of intracranial hemorrhage and chronic cerebral microbleeds: a systematic review and pooled analysis
La Radiologia medica. 2022
BACKGROUND Artificial intelligence (AI)-driven software has been developed and become commercially available within the past few years for the detection of intracranial hemorrhage (ICH) and chronic cerebral microbleeds (CMBs). However, there is currently no systematic review that summarizes all of these tools or provides pooled estimates of their performance. METHODS In this PROSPERO-registered, PRISMA compliant systematic review, we sought to compile and review all MEDLINE and EMBASE published studies that have developed and/or tested AI algorithms for ICH detection on non-contrast CT scans (NCCTs) or MRI scans and CMBs detection on MRI scans. RESULTS In total, 40 studies described AI algorithms for ICH detection in NCCTs/MRIs and 19 for CMBs detection in MRIs. The overall sensitivity, specificity, and accuracy were 92.06%, 93.54%, and 93.46%, respectively, for ICH detection and 91.6%, 93.9%, and 92.7% for CMBs detection. Some of the challenges encountered in the development of these algorithms include the laborious work of creating large, labeled and balanced datasets, the volumetric nature of the imaging examinations, the fine tuning of the algorithms, and the reduction in false positives. CONCLUSIONS Numerous AI-driven software tools have been developed over the last decade. On average, they are characterized by high performance and expert-level accuracy for the diagnosis of ICH and CMBs. As a result, implementing these tools in clinical practice may improve workflow and act as a failsafe for the detection of such lesions. REGISTRATION-URL: https://www.crd.york.ac.uk/prospero/ Unique Identifier: CRD42021246848.
Causes and Risk Factors of Pediatric Spontaneous Intracranial Hemorrhage-A Systematic Review
Diagnostics (Basel, Switzerland). 2022;12(6)
Previous studies suggest that the most common cause of spontaneous intracerebral hemorrhage in children and adolescents is arteriovenous malformations (AVMs). However, an update containing recently published data on pediatric spontaneous intracranial hemorrhages is lacking. The aim of this study is to systematically analyze the published data on the etiologies and risk factors of pediatric spontaneous intracranial hemorrhage. This systematic review was performed in compliance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. A search in PubMed, Embase, Scopus, Web of Science and Cochrane Library was conducted aiming for articles published in year 2000 and later, containing data on etiology and risk factors of spontaneous intracranial hemorrhages in unselected cohorts of patients aged between 1 month and 18 years. As a result, forty studies were eligible for data extraction and final analysis. These included 7931 children and adolescents with 4009 reported etiologies and risk factors. A marked variety of reported etiologies and risk factors among studies was observed. Vascular etiologies were the most frequently reported cause of pediatric spontaneous intracranial hemorrhages (n = 1727, 43.08% of all identified etiologies or risk factors), with AVMs being the most common vascular cause (n = 1226, 70.99% of all vascular causes). Hematological and systemic causes, brain tumors, intracranial infections and cardiac causes were less commonly encountered risk factors and etiologies.
A New Nomogram for Predicting the Risk of Intracranial Hemorrhage in Acute Ischemic Stroke Patients After Intravenous Thrombolysis
Frontiers in neurology. 2022;13:774654
BACKGROUND We aimed to develop and validate a new nomogram for predicting the risk of intracranial hemorrhage (ICH) in patients with acute ischemic stroke (AIS) after intravenous thrombolysis (IVT). METHODS A retrospective study enrolled 553 patients with AIS treated with IVT. The patients were randomly divided into two cohorts: the training set (70%, n = 387) and the testing set (30%, n = 166). The factors in the predictive nomogram were filtered using multivariable logistic regression analysis. The performance of the nomogram was assessed based on the area under the receiver operating characteristic curve (AUC-ROC), calibration plots, and decision curve analysis (DCA). RESULTS After multivariable logistic regression analysis, certain factors, such as smoking, National Institutes of Health of Stroke Scale (NIHSS) score, blood urea nitrogen-to-creatinine ratio (BUN/Cr), and neutrophil-to-lymphocyte ratio (NLR), were found to be independent predictors of ICH and were used to construct a nomogram. The AUC-ROC values of the nomogram were 0.887 (95% CI: 0.842-0.933) and 0.776 (95% CI: 0.681-0.872) in the training and testing sets, respectively. The AUC-ROC of the nomogram was higher than that of the Multicenter Stroke Survey (MSS), Glucose, Race, Age, Sex, Systolic blood Pressure, and Severity of stroke (GRASPS), and stroke prognostication using age and NIH Stroke Scale-100 positive index (SPAN-100) scores for predicting ICH in both the training and testing sets (p < 0.05). The calibration plot demonstrated good agreement in both the training and testing sets. DCA indicated that the nomogram was clinically useful. CONCLUSIONS The new nomogram, which included smoking, NIHSS, BUN/Cr, and NLR as variables, had the potential for predicting the risk of ICH in patients with AIS after IVT.
Viscoelastic Testing in the Clinical Management of Subarachnoid Hemorrhage and Intracerebral Hemorrhage
Seminars in thrombosis and hemostasis. 2022
Subarachnoid hemorrhage (SAH) and intracerebral hemorrhage (ICH) are both debilitating and life-threatening incidents calling for immediate action and treatment. This review focuses on the applicability of viscoelastic testing (rotational thromboelastometry or thromboelastography [TEG]) in the management of SAH and ICH. A systematic literature search was performed in PubMed and EMBASE. Studies including patients with SAH or ICH, in which viscoelastic testing was performed, were identified. In total, 24 studies were included for analysis, and further subdivided into studies on SAH patients investigated prior to stenting or coiling (n = 12), ICH patients (n = 8) and studies testing patients undergoing stenting or coiling, or ischemic stroke patients undergoing thrombolysis or thrombectomy and developing ICH as a complication (n = 5). SAH patients had increased clot firmness, and this was associated with a higher degree of early brain injury and higher Hunt-Hess score. SAH patients with delayed cerebral ischemia had higher clot firmness than patients not developing delayed cerebral ischemia. ICH patients showed accelerated clot formation and increased clot firmness in comparison to healthy controls. Patients with hematoma expansion had longer clot initiation and lower platelet aggregation than patients with no hematoma expansion. During stent procedures for SAH, adjustment of antiplatelet therapy according to TEG platelet mapping did not change prevalence of major bleeding, thromboembolic events, or functional outcome. Viscoelastic testing prior to thrombolysis showed conflicting results in predicting ICH as complication. In conclusion, viscoelastic testing suggests hypercoagulation following SAH and ICH. Further investigation of the predictive value of increased clot firmness in SAH seems relevant. In ICH, the prediction of hematoma expansion and ICH as a complication to thrombolysis might be clinically relevant.
Can Artificial Intelligence Be Applied to Diagnose Intracerebral Hemorrhage under the Background of the Fourth Industrial Revolution? A Novel Systemic Review and Meta-Analysis
International journal of clinical practice. 2022;2022:9430097
AIM: We intended to provide the clinical evidence that artificial intelligence (AI) could be used to assist doctors in the diagnosis of intracerebral hemorrhage (ICH). METHODS Studies published in 2021 were identified after the literature search of PubMed, Embase, and Cochrane. Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) was used to perform the quality assessment of studies. Data extraction of diagnosis effect included accuracy (ACC), sensitivity (SEN), specificity (SPE), positive predictive value (PPV), negative predictive value (NPV), area under curve (AUC), and Dice scores (Dices). The pooled effect with its 95% confidence interval (95%CI) was calculated by the random effects model. I-Square (I (2)) was used to test heterogeneity. To check the stability of the overall results, sensitivity analysis was conducted by recalculating the pooled effect of the remaining studies after omitting the study with the highest quality or the random effects model was switched to the fixed effects model. Funnel plot was used to evaluate publication bias. To reduce heterogeneity, recalculating the pooled effect of the remaining studies after omitting the study with the lowest quality or perform subgroup analysis. RESULTS Twenty-five diagnostic tests of ICH via AI and doctors with overall high quality were included. Pooled ACC, SEN, SPE, PPV, NPV, AUC, and Dices were 0.88 (0.83∼0.93), 0.85 (0.81∼0.89), 0.90 (0.88∼0.92), 0.80 (0.75∼0.85), 0.93 (0.91∼0.95), 0.84 (0.80∼0.89), and 0.90 (0.85∼0.95), respectively. There was no publication bias. All of results were stable as revealed by sensitivity analysis and were accordant as outcomes via subgroups analysis. CONCLUSION Under the background of the fourth industrial revolution, AI might be an effective and efficient tool to assist doctors in the clinical diagnosis of ICH.
Postsurgical functional outcome prediction model using deep learning framework (Prediction One, Sony Network Communications Inc.) for hypertensive intracerebral hemorrhage
Surgical neurology international. 2021;12:203
BACKGROUND Reliable prediction models of intracerebral hemorrhage (ICH) outcomes are needed for decision-making of the treatment. Statistically making such prediction models needs a large number of samples and time-consuming statistical analysis. Deep learning (DL), one of the artificial intelligence, is attractive, but there were no reports on DL-based functional outcome prediction models for ICH outcomes after surgery. We herein made a functional outcome prediction model using DLframework, Prediction One (Sony Network Communications Inc., Tokyo, Japan), and compared it to original ICH score, ICH Grading Scale, and FUNC score. METHODS We used 140 consecutive hypertensive ICH patients' data in our hospital between 2012 and 2019. All patients were surgically treated. Modified Rankin Scale 0-3 at 6 months was defined as a favorable outcome. We randomly divided them into 100 patients training dataset and 40 patients validation dataset. Prediction One made the prediction model using the training dataset with 5-fold cross-validation. We calculated area under the curves (AUCs) regarding the outcome using the DL-based model, ICH score, ICH Grading Scale, and FUNC score. The AUCs were compared. RESULTS The model made by Prediction One using 64 variables had AUC of 0.997 in the training dataset and that of 0.884 in the validation dataset. These AUCs were superior to those derived from ICH score, ICH Grading Scale, and FUNC score. CONCLUSION We easily and quickly made prediction models using Prediction One, even with a small single-center dataset. The accuracy of the DL-based model was superior to those of previous statistically calculated models.
Convolutional neural network performance compared to radiologists in detecting intracranial hemorrhage from brain computed tomography: A systematic review and meta-analysis
European journal of radiology. 2021;146:110073
PURPOSE To compare the diagnostic accuracy of convolutional neural networks (CNN) with radiologists as the reference standard in the diagnosis of intracranial hemorrhages (ICH) with non contrast computed tomography of the cerebrum (NCTC). METHODS PubMed, Embase, Scopus, and Web of Science were searched for the period from 1 January 2012 to 20 July 2020; eligible studies included patients with and without ICH as the target condition undergoing NCTC, studies had deep learning algorithms based on CNNs and radiologists reports as the minimum reference standard. Pooled sensitivities, specificities and a summary receiver operating characteristics curve (SROC) were employed for meta-analysis. RESULTS 5,119 records were identified through database searching. Title-screening left 47 studies for full-text assessment and 6 studies for meta-analysis. Comparing the CNN performance to reference standards in the retrospective studies found a pooled sensitivity of 96.00% (95% CI: 93.00% to 97.00%), pooled specificity of 97.00% (95% CI: 90.00% to 99.00%) and SROC of 98.00% (95% CI: 97.00% to 99.00%), and combining retrospective and studies with external datasets found a pooled sensitivity of 95.00% (95% CI: 91.00% to 97.00%), pooled specificity of 96.00% (95% CI: 91.00% to 98.00%) and a pooled SROC of 98.00% (95% CI: 97.00% to 99.00%). CONCLUSION This review found the diagnostic performance of CNNs to be equivalent to that of radiologists for retrospective studies. Out-of-sample external validation studies pooled with retrospective studies found CNN performance to be slightly worse. There is a critical need for studies with a robust reference standard and external data-set validation.