Two closely-spaced Aneurysms of the Supraclinoid Internal Carotid Artery: How Does One Influence the Other?

Augmenting Prediction of Intracranial Aneurysms' Risk Status Using Velocity-Informatics: Initial Experience

Our primary goal here is to demonstrate that innovative analytics of aneurismal velocities, named velocity-informatics, enhances intracranial aneurysm (IA) rupture status prediction. 3D computer models were generated using imaging data from 112 subjects harboring anterior IAs (4-25 mm; 44 ruptured and 68 unruptured). Computational fluid dynamics simulations and geometrical analyses were performed. Then, computed 3D velocity vector fields within the IA dome were processed for velocity-informatics. Four machine learning methods (support vector machine, random forest, generalized linear model, and GLM with Lasso or elastic net regularization) were employed to assess the merits of the proposed velocity-informatics. All 4 ML methods consistently showed that, with velocity-informatics metrics, the area under the curve and prediction accuracy both improved by approximately 0.03. Overall, with velocity-informatics, the support vector machine's prediction was most promising: an AUC of 0.86 and total accuracy of 77%, with 60% and 88% of ruptured and unruptured IAs being correctly identified, respectively.

Deep-learning-based image segmentation for image-based computational hemodynamic analysis of abdominal aortic aneurysms: a comparison study

Computational hemodynamics is increasingly being used to quantify hemodynamic characteristics in and around abdominal aortic aneurysms (AAA) in a patient-specific fashion. However, the time-consuming manual annotation hinders the clinical translation of computational hemodynamic analysis. Thus, we investigate the feasibility of using deep-learning-based image segmentation methods to reduce the time required for manual segmentation. Two of the latest deep-learning-based image segmentation methods, ARU-Net and CACU-Net, were used to test the feasibility of automated computer model creation for computational hemodynamic analysis. Morphological features and hemodynamic metrics of 30 computed tomography angiography (CTA) scans were compared between pre-dictions and manual models. The DICE score for both networks was 0.916, and the correlation value was above 0.95, indicating their ability to generate models comparable to human segmentation. The Bland-Altman analysis shows a good agreement between deep learning and manual segmentation results. Compared with manual (computational hemodynamics) model recreation, the time for automated computer model generation was significantly reduced (from ∼2 h to ∼10 min). Automated image segmentation can significantly reduce time expenses on the recreation of patient-specific AAA models. Moreover, our study showed that both CACU-Net and ARU-Net could accomplish AAA segmentation, and CACU-Net outperformed ARU-Net in terms of accuracy and time-saving.

USING CONVOLUTIONAL NEURAL NETWORK-BASED SEGMENTATION FOR IMAGE-BASED COMPUTATIONAL FLUID DYNAMICS SIMULATIONS OF BRAIN ANEURYSMS: INITIAL EXPERIENCE IN AUTOMATED MODEL CREATION

"Image-based" computational fluid dynamics (CFD) simulations provide insights into each patient's hemodynamic environment. However, current standard procedures for creating CFD models start with manual segmentation and are time-consuming, hindering the clinical translation of image-based CFD simulations. This feasibility study adopts deep-learning-based image segmentation (hereafter referred to as Artificial Intelligence (AI) segmentation) to replace manual segmentation to accelerate CFD model creation. Two published convolutional neural network-based AI methods (MIScnn and DeepMedic) were selected to perform CFD model extraction from three-dimensional (3D) rotational angiography data containing intracranial aneurysms. In this study, aneurysm morphological and hemodynamic results using models generated by AI segmentation methods were compared with those obtained by two human users for the same data. Interclass coefficients (ICC), Bland-Altman plots, and Pearson's correlation coefficients (PCC) were combined to assess how well AI-generated CFD models were performed. We found that almost perfect agreement was obtained between the human and AI results for all eleven morphological and five out of eight hemodynamic parameters, while a moderate agreement was obtained from the remaining three hemodynamic parameters. Given this level of agreement, using AI segmentation to create CFD models is feasible, given more developments.

Can we explain machine learning-based prediction for rupture status assessments of intracranial aneurysms?

Although applying machine learning (ML) algorithms to rupture status assessment of intracranial aneurysms (IA) has yielded promising results, the opaqueness of some ML methods has limited their clinical translation. We presented the first explainability comparison of six commonly used ML algorithms: multivariate logistic regression (LR), support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost), multi-layer perceptron neural network (MLPNN), and Bayesian additive regression trees (BART). A total of 112 IAs with known rupture status were selected for this study. The ML-based classification used two anatomical features, nine hemodynamic parameters, and thirteen morphologic variables. We utilized permutation feature importance, local interpretable model-agnostic explanations (LIME), and SHapley Additive exPlanations (SHAP) algorithms to explain and analyze 6 Ml algorithms. All models performed comparably: LR area under the curve (AUC) was 0.71; SVM AUC was 0.76; RF AUC was 0.73; XGBoost AUC was 0.78; MLPNN AUC was 0.73; BART AUC was 0.73. Our interpretability analysis demonstrated consistent results across all the methods; i.e., the utility of the top 12 features was broadly consistent. Furthermore, contributions of 9 important features (aneurysm area, aneurysm location, aneurysm type, wall shear stress maximum during systole, ostium area, the size ratio between aneurysm width, (parent) vessel diameter, one standard deviation among time-averaged low shear area, and one standard deviation of temporally averaged low shear area less than 0.4 Pa) were nearly the same. This research suggested that ML classifiers can provide explainable predictions consistent with general domain knowledge concerning IA rupture. With the improved understanding of ML algorithms, clinicians' trust in ML algorithms will be enhanced, accelerating their clinical translation.

Computational Assessment of Hemodynamics Vortices Within the Cerebral Vasculature Using Informational Entropy

Propper assessment of hemodynamic swirling flow patterns, vortices, may help understand the influence of disturbed flow on arterial wall pathophysiology and remodeling. Studies have shown that vortices trigger pathologic cellular changes within the vasculature such as increased inflammation and cellular apoptosis, leading to weakening of the vessel wall indicative of aneurysm development and rupture. Yet many studies qualitatively assess the presence of vortices within the vasculature or assess only their centermost region (critical point analysis) which overlooks the broader characteristics of flow, leading to a narrow view of vortices. This chapter provides a protocol for utilizing commercially available computational fluid dynamic software (ANSYS-FLUENT) to simulate realistic hemodynamic flow patterns, fluid velocity, and wall shear stress in the complex geometry of the cerebral vasculature, as well as an innovative method for assessing flow vortices. This innovative analytic methodology can identify areas of flow vortices and quantify how the broader bulk-flow (opposed to critical point) characteristics change in space and time over the cardiac cycle. Analysis of such flow structures can be used to identify specific characteristics such as vortex stability and the portion of an aneurysmal sac that is dominated by swirling flow, which may be indicative of vascular pathologies.

Quantitative analysis of flow vortices: differentiation of unruptured and ruptured medium-sized middle cerebral artery aneurysms

BACKGROUND

Surgical intervention for unruptured intracranial aneurysms (IAs) carries inherent health risks. The analysis of "patient-specific" IA geometric and computational fluid dynamics (CFD) simulated wall shear stress (WSS) data has been investigated to differentiate IAs at high and low risk of rupture to help clinical decision making. Yet, outcomes vary among studies, suggesting that novel analysis could improve rupture characterization. The authors describe a CFD analytic method to assess spatiotemporal characteristics of swirling flow vortices within IAs to improve characterization.

METHODS

CFD simulations were performed for 47 subjects harboring one medium-sized (4-10 mm) middle cerebral artery (MCA) aneurysm with available 3D digital subtraction angiography data. Alongside conventional indices, quantified IA flow vortex spatiotemporal characteristics were applied during statistical characterization. Statistical supervised machine learning using a support vector machine (SVM) method was run with cross-validation (100 iterations) to assess flow vortex-based metrics' strength toward rupture characterization.

RESULTS

Relying solely on vortex indices for statistical characterization underperformed compared with established geometric characteristics (total accuracy of 0.77 vs 0.80) yet showed improvements over wall shear stress models (0.74). However, the application of vortex spatiotemporal characteristics into the combined geometric and wall shear stress parameters augmented model strength for assessing the rupture status of middle cerebral artery aneurysms (0.85).

CONCLUSIONS

This preliminary study suggests that the spatiotemporal characteristics of flow vortices within MCA aneurysms are of value to improve the differentiation of ruptured aneurysms from unruptured ones.

Multivariate analysis of hemodynamic parameters on intracranial aneurysm initiation of the internal carotid artery

Although fluctuating hemodynamic wall stressors are known to impact intracranial aneurysms (IA) initiation, specificity of those stressors has not been evaluated. In this study, using human IA data, we investigated: (1) specificity of stressors in regions with and without IA eventual IA formation; and (2) how combinations of multiple stressors could improve IA formation prediction. 3D computational vasculatures were constructed based on angiographic images of 18 subjects having multiple closely-spaced IAs in the internal carotid artery. Two models were created: Model A with all IAs computationally removed, Model B which kept keep one IA. Computational fluid dynamics (CFD) simulated flow within models. Based on simulated flow fields, wall shear stress and its gradient (WSS, WSSG), oscillatory shear index (OSI), gradient oscillatory number (GON), aneurysm formation index (AFI), and mean number of swirling flow vortices (MV) were analysed. Multivariate logistic regression determined the accuracy of different combinations of those above-mentioned stressors. Overall, we found that combining hemodynamic stressors improves IA formation prediction over individual indices. Both Model A and Model B's parsimonious model was MV+WSS+GON: AUROC 0.88 and 0.83, respectively. Future studies are planned to understand biological meanings induced by fluctuating stressors.