Automated segmentation of subarachnoid hemorrhages with convolutional neural networks

Automated segmentation of ventricular volumes and subarachnoid hemorrhage from computed tomography images: Evaluation of a rule-based pipeline approach

Changes in ventricular size, related to brain edema and hydrocephalus, as well as the extent of hemorrhage are associated with adverse outcomes in patients with subarachnoid hemorrhage (SAH). Frequently, these are measured manually using consecutive non-contrast computed tomography scans. Here, we developed a rule-based approach which incorporates both intensity and spatial normalization and utilizes user-defined thresholds and anatomical templates to segment both lateral ventricle (LV) and SAH blood volumes automatically from CT images. The algorithmic segmentations were evaluated against two expert neuroradiologists on representative slices from 20 admission scans from aneurysmal SAH patients. Previous methods have been developed to automate this time-consuming task, but they lack user feedback and are hard to implement due to large-scale data and complex design processes. Our results using automatic ventricular segmentation aligned well with expert reviewers with a median Dice coefficient of 0.81, AUC of 0.91, sensitivity of 81%, and precision of 84%. Automatic segmentation of SAH blood was most reliable near the base of the brain with a median Dice coefficient of 0.51, an AUC of 0.75, precision of 68%, and sensitivity of 50%. Ultimately, we developed a rule-based method that is easily adaptable through user feedback, generates spatially normalized segmentations that are comparable regardless of brain morphology or acquisition conditions, and automatically segments LV with good overall reliability and basal SAH blood with good precision. Our approach could benefit longitudinal studies in patients with SAH by streamlining assessment of edema and hydrocephalus progression, as well as blood resorption.

Improved hypertensive stroke classification based on multi-scale feature fusion of head axial CT angiogram and multimodal learning

PURPOSE

Strokes are severe cardiovascular and circulatory diseases with two main types: ischemic and hemorrhagic. Clinically, brain images such as computed tomography (CT) and computed tomography angiography (CTA) are widely used to recognize stroke types. However, few studies have combined imaging and clinical data to classify stroke or consider a factor as an Independent etiology.

METHODS

In this work, we propose a classification model that automatically distinguishes stroke types with hypertension as an independent etiology based on brain imaging and clinical data. We first present a preprocessing workflow for head axial CT angiograms, including noise reduction and feature enhancement of the images, followed by an extraction of regions of interest. Next, we develop a multi-scale feature fusion model that combines the location information of position features and the semantic information of deep features. Furthermore, we integrate brain imaging with clinical information through a multimodal learning model to achieve more reliable results.

RESULTS

Experimental results show our proposed models outperform state-of-the-art models on real imaging and clinical data, which reveals the potential of multimodal learning in brain disease diagnosis.

CONCLUSION

The proposed methodologies can be extended to create AI-driven diagnostic assistance technology for categorizing strokes.

Concurrent Ischemic Lesion Age Estimation and Segmentation of CT Brain Using a Transformer-Based Network

The cornerstone of stroke care is expedient management that varies depending on the time since stroke onset. Consequently, clinical decision making is centered on accurate knowledge of timing and often requires a radiologist to interpret Computed Tomography (CT) of the brain to confirm the occurrence and age of an event. These tasks are particularly challenging due to the subtle expression of acute ischemic lesions and the dynamic nature of their appearance. Automation efforts have not yet applied deep learning to estimate lesion age and treated these two tasks independently, so, have overlooked their inherent complementary relationship. To leverage this, we propose a novel end-to-end multi-task transformer-based network optimized for concurrent segmentation and age estimation of cerebral ischemic lesions. By utilizing gated positional self-attention and CT-specific data augmentation, the proposed method can capture long-range spatial dependencies while maintaining its ability to be trained from scratch under low-data regimes commonly found in medical imaging. Furthermore, to better combine multiple predictions, we incorporate uncertainty by utilizing quantile loss to facilitate estimating a probability density function of lesion age. The effectiveness of our model is then extensively evaluated on a clinical dataset consisting of 776 CT images from two medical centers. Experimental results demonstrate that our method obtains promising performance, with an area under the curve (AUC) of 0.933 for classifying lesion ages ≤ 4.5 hours compared to 0.858 using a conventional approach, and outperforms task-specific state-of-the-art algorithms.

Clinical potential of automated convolutional neural network-based hematoma volumetry after aneurysmal subarachnoid hemorrhage

OBJECTIVES

Cerebrospinal fluid hemoglobin has been positioned as a potential biomarker and drug target for aneurysmal subarachnoid hemorrhage-related secondary brain injury (SAH-SBI). The maximum amount of hemoglobin, which may be released into the cerebrospinal fluid, is defined by the initial subarachnoid hematoma volume (ISHV). In patients without external ventricular or lumbar drain, there remains an unmet clinical need to predict the risk for SAH-SBI. The aim of this study was to explore automated segmentation of ISHV as a potential surrogate for cerebrospinal fluid hemoglobin to predict SAH-SBI.

METHODS

This study is based on a retrospective analysis of imaging and clinical data from 220 consecutive patients with aneurysmal subarachnoid hemorrhage collected over a five-year period. 127 annotated initial non-contrast CT scans were used to train and test a convolutional neural network to automatically segment the ISHV in the remaining cohort. Performance was reported in terms of Dice score and intraclass correlation. We characterized the associations between ISHV and baseline cohort characteristics, SAH-SBI, ventriculoperitoneal shunt dependence, functional outcome, and survival. Established clinical (World Federation of Neurosurgical Societies, Hunt & Hess) and radiological (modified Fisher, Barrow Neurological Institute) scores served as references.

RESULTS

A strong volume agreement (0.73 Dice, range 0.43 - 0.93) and intraclass correlation (0.89, 95% CI, 0.81-0.94) were shown. While ISHV was not associated with the use of antithrombotics or cardiovascular risk factors, there was strong evidence for an association with a lower Glasgow Coma Scale at hospital admission. Aneurysm size and location were not associated with ISHV, but the presence of intracerebral or intraventricular hemorrhage were independently associated with higher ISHV. Despite strong evidence for a positive association between ISHV and SAH-SBI, the discriminatory ability of ISHV for SAH-SBI was insufficient. The discriminatory ability of ISHV was, however, higher regarding ventriculoperitoneal shunt dependence and functional outcome at three-months follow-up. Multivariate survival analysis provided strong evidence for an independent negative association between survival probability and both ISHV and intraventricular hemorrhage.

CONCLUSIONS

The proposed algorithm demonstrates strong performance in volumetric segmentation of the ISHV on the admission CT. While the discriminatory ability of ISHV for SAH-SBI was similar to established clinical and radiological scores, it showed a high discriminatory ability for ventriculoperitoneal shunt dependence and functional outcome at three-months follow-up.

Predicting vasospasm risk using first presentation aneurysmal subarachnoid hemorrhage volume: A semi-automated CT image segmentation analysis using ITK-SNAP

PURPOSE

Cerebral vasospasm following aneurysmal subarachnoid hemorrhage (aSAH) is a significant complication associated with poor neurological outcomes. We present a novel, semi-automated pipeline, implemented in the open-source medical imaging analysis software ITK-SNAP, to segment subarachnoid blood volume from initial CT head (CTH) scans and use this to predict future radiological vasospasm.

METHODS

42 patients were admitted between February 2020 and December 2021 to our tertiary neurosciences center, and whose initial referral CTH scan was used for this retrospective cohort study. Blood load was segmented using a semi-automated random forest classifier and active contour evolution implemented in ITK-SNAP. Clinical data were extracted from electronic healthcare records in order to fit models aimed at predicting radiological vasospasm risk.

RESULTS

Semi-automated segmentations demonstrated excellent agreement with manual, expert-derived volumes (mean Dice coefficient = 0.92). Total normalized blood volume, extracted from CTH images at first presentation, was significantly associated with greater odds of later radiological vasospasm, increasing by approximately 7% for each additional cm3 of blood (OR = 1.069, 95% CI: 1.021-1.120; p < .005). Greater blood volume was also significantly associated with vasospasm of a higher Lindegaard ratio, of longer duration, and a greater number of discrete episodes. Total blood volume predicted radiological vasospasm with a greater accuracy as compared to the modified Fisher scale (AUC = 0.86 vs 0.70), and was of independent predictive value.

CONCLUSION

Semi-automated methods provide a plausible pipeline for the segmentation of blood from CT head images in aSAH, and total blood volume is a robust, extendable predictor of radiological vasospasm, outperforming the modified Fisher scale. Greater subarachnoid blood volume significantly increases the odds of subsequent vasospasm, its time course and its severity.

Prehospital transdermal glyceryl trinitrate in patients with presumed acute stroke (MR ASAP): an ambulance-based, multicentre, randomised, open-label, blinded endpoint, phase 3 trial

BACKGROUND

Pooled analyses of previous randomised studies have suggested that very early treatment with glyceryl trinitrate (also known as nitroglycerin) improves functional outcome in patients with acute ischaemic stroke or intracerebral haemorrhage, but this finding was not confirmed in a more recent trial (RIGHT-2). We aimed to assess whether patients with presumed acute stroke benefit from glyceryl tr initrate started within 3 h after symptom onset.

METHODS

MR ASAP was a phase 3, randomised, open-label, blinded endpoint trial done at six ambulance services serving 18 hospitals in the Netherlands. Eligible participants (aged ≥18 years) had a probable diagnosis of acute stroke (as assessed by a paramedic), a face-arm-speech-time test score of 2 or 3, systolic blood pressure of at least 140 mm Hg, and could start treatment within 3 h of symptom onset. Participants were randomly assigned (1:1) by ambulance personnel, using a secure web-based electronic application with random block sizes stratified by ambulance service, to receive either transdermal glyceryl trinitrate 5 mg/day for 24 h plus standard care (glyceryl trinitrate group) or to standard care alone (control group) in the prehospital setting. Informed consent was deferred until after arrival at the hospital. The primary outcome was functional outcome assessed with the modified Rankin Scale (mRS) at 90 days. Safety outcomes included death within 7 days, death within 90 days, and serious adverse events. Analyses were based on modified intention to treat, and treatment effects were expressed as odds ratios (ORs) or common ORs, with adjustment for baseline prognostic factors. We separately analysed the total population and the target population (ie, patients with intracerebral haemorrhage, ischaemic stroke, or transient ischaemic attack). The target sample size was 1400 patients. The trial is registered as ISRCTN99503308.

FINDINGS

On June 24, 2021, the MR ASAP trial was prematurely terminated on the advice of the data and safety monitoring board, with recruitment stopped because of safety concerns in patients with intracerebral haemorrhage. Between April 4, 2018, and Feb 12, 2021, 380 patients were randomly allocated to a study group. 325 provided informed consent or died before consent could be obtained, of whom 170 were assigned to the glyceryl trinitrate group and 155 to the control group. These patients were included in the total population. 201 patients (62%) had ischaemic stroke, 34 (10%) transient ischaemic attack, 56 (17%) intracerebral haemorrhage, and 34 (10%) a stroke-mimicking condition. In the total population (n=325), the median mRS score at 90 days was 2 (IQR 1-4) in both the glyceryl trinitrate and control groups (adjusted common OR 0·97 [95% CI 0·65-1·47]). In the target population (n=291), the 90-day mRS score was 2 (2-4) in the glyceryl trinitrate group and 3 (1-4) in the control group (0·92 [0·59-1·43]). In the total population, there were no differences between the two study groups with respect to death within 90 days (adjusted OR 1·07 [0·53-2·14]) or serious adverse events (unadjusted OR 1·23 [0·76-1·99]). In patients with intracerebral haemorrhage, 12 (34%) of 35 patients allocated to glyceryl trinitrate versus two (10%) of 21 allocated to the control group died within 7 days (adjusted OR 5·91 [0·78-44·81]); death within 90 days occurred in 16 (46%) of 35 in the glyceryl trinitrate group and 11 (55%) of 20 in the control group (adjusted OR 0·87 [0·18-4·17]).

INTERPRETATION

We found no sign of benefit of transdermal glyceryl trinitrate started within 3 h of symptom onset in the prehospital setting in patients with presumed acute stroke. The signal of potential early harm of glyceryl trinitrate in patients with intracerebral haemorrhage suggests that glyceryl trinitrate should be avoided in this setting.

FUNDING

The Collaboration for New Treatments of Acute Stroke consortium, the Brain Foundation Netherlands, the Ministry of Economic Affairs, Stryker, Medtronic, Cerenovus, and the Dutch Heart Foundation.

Unsupervised Deep Learning for Stroke Lesion Segmentation on Follow-up CT Based on Generative Adversarial Networks

BACKGROUND AND PURPOSE

Supervised deep learning is the state-of-the-art method for stroke lesion segmentation on NCCT. Supervised methods require manual lesion annotations for model development, while unsupervised deep learning methods such as generative adversarial networks do not. The aim of this study was to develop and evaluate a generative adversarial network to segment infarct and hemorrhagic stroke lesions on follow-up NCCT scans.

MATERIALS AND METHODS

Training data consisted of 820 patients with baseline and follow-up NCCT from 3 Dutch acute ischemic stroke trials. A generative adversarial network was optimized to transform a follow-up scan with a lesion to a generated baseline scan without a lesion by generating a difference map that was subtracted from the follow-up scan. The generated difference map was used to automatically extract lesion segmentations. Segmentation of primary hemorrhagic lesions, hemorrhagic transformation of ischemic stroke, and 24-hour and 1-week follow-up infarct lesions were evaluated relative to expert annotations with the Dice similarity coefficient, Bland-Altman analysis, and intraclass correlation coefficient.

RESULTS

The median Dice similarity coefficient was 0.31 (interquartile range, 0.08-0.59) and 0.59 (interquartile range, 0.29-0.74) for the 24-hour and 1-week infarct lesions, respectively. A much lower Dice similarity coefficient was measured for hemorrhagic transformation (median, 0.02; interquartile range, 0-0.14) and primary hemorrhage lesions (median, 0.08; interquartile range, 0.01-0.35). Predicted lesion volume and the intraclass correlation coefficient were good for the 24-hour (bias, 3 mL; limits of agreement, -64-59 mL; intraclass correlation coefficient, 0.83; 95% CI, 0.78-0.88) and excellent for the 1-week (bias, -4 m; limits of agreement,-66-58 mL; intraclass correlation coefficient, 0.90; 95% CI, 0.83-0.93) follow-up infarct lesions.

CONCLUSIONS

An unsupervised generative adversarial network can be used to obtain automated infarct lesion segmentations with a moderate Dice similarity coefficient and good volumetric correspondence.

Prediction and Risk Assessment Models for Subarachnoid Hemorrhage: A Systematic Review on Case Studies

Subarachnoid hemorrhage (SAH) is one of the major health issues known to society and has a higher mortality rate. The clinical factors with computed tomography (CT), magnetic resonance image (MRI), and electroencephalography (EEG) data were used to evaluate the performance of the developed method. In this paper, various methods such as statistical analysis, logistic regression, machine learning, and deep learning methods were used in the prediction and detection of SAH which are reviewed. The advantages and limitations of SAH prediction and risk assessment methods are also being reviewed. Most of the existing methods were evaluated on the collected dataset for the SAH prediction. In some researches, deep learning methods were applied, which resulted in higher performance in the prediction process. EEG data were applied in the existing methods for the prediction process, and these methods demonstrated higher performance. However, the existing methods have the limitations of overfitting problems, imbalance data problems, and lower efficiency in feature analysis. The artificial neural network (ANN) and support vector machine (SVM) methods have been applied for the prediction process, and considerably higher performance is achieved by using this method.

A Review on Computer Aided Diagnosis of Acute Brain Stroke

Amongst the most common causes of death globally, stroke is one of top three affecting over 100 million people worldwide annually. There are two classes of stroke, namely ischemic stroke (due to impairment of blood supply, accounting for ~70% of all strokes) and hemorrhagic stroke (due to bleeding), both of which can result, if untreated, in permanently damaged brain tissue. The discovery that the affected brain tissue (i.e., 'ischemic penumbra') can be salvaged from permanent damage and the bourgeoning growth in computer aided diagnosis has led to major advances in stroke management. Abiding to the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines, we have surveyed a total of 177 research papers published between 2010 and 2021 to highlight the current status and challenges faced by computer aided diagnosis (CAD), machine learning (ML) and deep learning (DL) based techniques for CT and MRI as prime modalities for stroke detection and lesion region segmentation. This work concludes by showcasing the current requirement of this domain, the preferred modality, and prospective research areas.

A simplified framework for the detection of intracranial hemorrhage in CT brain images using deep learning

BACKGROUND

The need for accurate and timely detection of Intracranial hemorrhage (ICH) is utmost important to avoid untoward incidents that may even lead to death.Hence, this presented work leverages the ability of a pretrained deep convolutional neural network (CNN) for the detection of ICH in computed tomography (CT) brain images.

METHODS

Different frameworks have been analyzed for their effectiveness for the classification of CT brain images into hemorrhage or non-hemorrhage conditions. All these frameworks were investigated on CQ500 dataset. Furthermore, an exclusive preprocessing pipeline was designed for both normal and ICH CT images. Firstly, a framework involving the pretrained deep CNN, AlexNet, has been exploited for both feature extraction and classification using the transfer learning method, secondly, a modified AlexNet-Support vector machine (SVM) classifier is explored and finally, a feature selection method, Principal Component Analysis (PCA) has been introduced in the AlexNet-SVM classifier model and its efficacy is explored.These models were trained and tested on two different sets of CT images, one containing the original images without preprocessing and another set consisting of preprocessed images.

RESULTS

The modified AlexNet-SVM classifier has shown an improved performance in comparison to the other investigated frameworks and has achieved a classification accuracy of 99.86%, sensitivity and specificity of 0.9986 for the detection of ICH in brain CT images.

CONCLUSION

This research has given an overview of a simple and efficient framework for the classification of hemorrhage and non-hemorrhage images. Also, the proposed simplified deep learning framework manifests its ability as a screening tool to assist the radiological trainees for the accurate detection of ICH.

Automatic segmentation of cerebral infarcts in follow-up computed tomography images with convolutional neural networks

Background and purpose

Infarct volume is a valuable outcome measure in treatment trials of acute ischemic stroke and is strongly associated with functional outcome. Its manual volumetric assessment is, however, too demanding to be implemented in clinical practice.

Objective

To assess the value of convolutional neural networks (CNNs) in the automatic segmentation of infarct volume in follow-up CT images in a large population of patients with acute ischemic stroke.

Materials and methods

We included CT images of 1026 patients from a large pooling of patients with acute ischemic stroke. A reference standard for the infarct segmentation was generated by manual delineation. We introduce three CNN models for the segmentation of subtle, intermediate, and severe hypodense lesions. The fully automated infarct segmentation was defined as the combination of the results of these three CNNs. The results of the three-CNNs approach were compared with the results from a single CNN approach and with the reference standard segmentations.

Results

The median infarct volume was 48 mL (IQR 15–125 mL). Comparison between the volumes of the three-CNNs approach and manually delineated infarct volumes showed excellent agreement, with an intraclass correlation coefficient (ICC) of 0.88. Even better agreement was found for severe and intermediate hypodense infarcts, with ICCs of 0.98 and 0.93, respectively. Although the number of patients used for training in the single CNN approach was much larger, the accuracy of the three-CNNs approach strongly outperformed the single CNN approach, which had an ICC of 0.34.

Conclusion

Convolutional neural networks are valuable and accurate in the quantitative assessment of infarct volumes, for both subtle and severe hypodense infarcts in follow-up CT images. Our proposed three-CNNs approach strongly outperforms a more straightforward single CNN approach.