Prediction of hematoma expansion in spontaneous intracerebral hemorrhage using a multimodal neural network.

Hematoma expansion occasionally occurs in patients with intracerebral hemorrhage (ICH), associating with poor outcome. Multimodal neural networks incorporating convolutional neural network (CNN) analysis of images and neural network analysis of tabular data are known to show promising results in prediction and classification tasks. We aimed to develop a reliable multimodal neural network model that comprehensively analyzes CT images and clinical variables to predict hematoma expansion. We retrospectively enrolled ICH patients at four hospitals between 2017 and 2021, assigning patients from three hospitals to the training and validation dataset and patients from one hospital to the test dataset. Admission CT images and clinical variables were collected. CT findings were evaluated by experts. Three types of models were developed and trained: (1) a CNN model analyzing CT images, (2) a multimodal CNN model analyzing CT images and clinical variables, and (3) a non-CNN model analyzing CT findings and clinical variables with machine learning. The models were evaluated on the test dataset, focusing first on sensitivity and second on area under the receiver operating curve (AUC). Two hundred seventy-three patients (median age, 71 years [59-79]; 159 men) in the training and validation dataset and 106 patients (median age, 70 years [62-82]; 63 men) in the test dataset were included. Sensitivity and AUC of a CNN model were 1.000 (95% confidence interval [CI] 0.768-1.000) and 0.755 (95% CI 0.704-0.807); those of a multimodal CNN model were 1.000 (95% CI 0.768-1.000) and 0.799 (95% CI 0.749-0.849); and those of a non-CNN model were 0.857 (95% CI 0.572-0.982) and 0.733 (95% CI 0.625-0.840). We developed a multimodal neural network model incorporating CNN analysis of CT images and neural network analysis of clinical variables to predict hematoma expansion in ICH. The model was externally validated and showed the best performance of all the models.

Machine learning prediction of hematoma expansion in acute intracerebral hemorrhage.

To examine whether machine learning (ML) approach can be used to predict hematoma expansion in acute intracerebral hemorrhage (ICH) with accuracy and widespread applicability, we applied ML algorithms to multicenter clinical data and CT findings on admission. Patients with acute ICH from three hospitals (n = 351) and those from another hospital (n = 71) were retrospectively assigned to the development and validation cohorts, respectively. To develop ML predictive models, the k-nearest neighbors (k-NN) algorithm, logistic regression, support vector machines (SVMs), random forests, and XGBoost were applied to the patient data in the development cohort. The models were evaluated for their performance on the patient data in the validation cohort, which was compared with previous scoring methods, the BAT, BRAIN, and 9-point scores. The k-NN algorithm achieved the highest area under the receiver operating characteristic curve (AUC) of 0.790 among all ML models, and the sensitivity, specificity, and accuracy were 0.846, 0.733, and 0.775, respectively. The BRAIN score achieved the highest AUC of 0.676 among all previous scoring methods, which was lower than the k-NN algorithm (p = 0.016). We developed and validated ML predictive models of hematoma expansion in acute ICH. The models demonstrated good predictive ability, showing better performance than the previous scoring methods.

A Case of Suspected Metallic Embolism after Carotid Artery Stenting.

A case in which metallic embolism was suspected after carotid artery stenting (CAS) is described. A 79-year-old woman was referred to our hospital because of a severe stenosis of the left cervical internal carotid artery (ICA). Carotid ultrasound revealed that the plaque was fibrous and was accompanied with partial calcification. The carotid stenosis was treated by CAS. The magnetic resonance imaging (MRI) taken in the following day of the CAS demonstrated that a new abnormal spot at the left frontal lobe. The spot appeared as a signal void on T1, T2, diffusion, susceptibility-weighted image (SWI), and fluid attenuated inversion recovery (FLAIR) image, and was surrounded by a high-signal halo on T2 and diffusion-weighted images (DWIs). The spot also demonstrated "blooming" appearance on SWIs. Despite the lesion she was asymptomatic all through the postoperative course, and she left our hospital on postoperative day 6. Follow-up MRI obtained 27 months after the CAS demonstrated that the lesion remained at the left frontal lobe without any signal changes. The patient remained asymptomatic at the last follow-up. Considering the location of the new abnormal spot (in the vascular territory of the catheterized vessel), these imaging characteristics and asymptomatic clinical course, the spot likely suggested metallic embolism. This is the first case in which the metallic embolism was suspected after CAS.

Prediction of the extent of thrombus formation in the parent artery after endovascular occlusion of a distal anterior cerebral artery aneurysm using computational fluid dynamics.

Endovascular coiling of a cerebral aneurysm and coil occlusion of the parent artery have been occasionally performed to treat cerebral aneurysms; however, it is difficult to predict the accurate extent of thrombus formation in the parent artery proximal to the coiled aneurysm and the coil-occluded parent artery preoperatively, and unexpected occlusion of the arterial branches can occur by thrombus extension into or in the parent artery. The authors describe a case of a distal anterior cerebral artery (ACA) aneurysm treated by endovascular parent artery occlusion (PAO) with preoperative computational fluid dynamics (CFD) prediction of the extent of thrombus formation. A 73-year-old woman presented with subarachnoid hemorrhage and an aneurysm that was located on the right pericallosal artery distal to the paracentral artery bifurcation. Endovascular coiling of the aneurysm and the pericallosal artery was planned. In advance of the treatment, CFD was performed to predict the extent of thrombus formation with specific wall shear stress and shear rate thresholds. The hemodynamic results indicated that coiling of the aneurysm resulted in thrombus formation in the pericallosal artery up to just distal to the paracentral artery ostium; therefore, the treatment was implemented according to the CFD prediction. Postoperative digital subtraction angiography revealed that the extent of thrombus formation was consistent with the preoperative CFD prediction. This technique may prevent unexpected occlusion of arterial branches.

Quantification of hemodynamic irregularity using oscillatory velocity index in the associations with the rupture status of cerebral aneurysms.

BACKGROUND: Complex and unstable flow patterns are reported to be associated with the rupture status of cerebral aneurysms, while their evaluation depends on qualitative analysis of streamlines of bloodflow. Oscillatory velocity index (OVI) is a hemodynamic parameter to quantify flow patterns. The aim of this study is to elucidate the associations between OVI and the rupture status of cerebral aneurysms. METHODS: One hundred and twenty-nine ruptured and unruptured cerebral aneurysms were analyzed with computational fluid dynamics under pulsatile flow conditions. With the use of median value of OVI, all aneurysms were divided into high and low OVI groups. Statistical analysis was performed to compare rupture status, and morphological and hemodynamic parameters between the two groups. RESULTS: The median value of OVI was 0.006. High OVI was more likely observed in ruptured aneurysms (P=0.028) and associated with irregular shape, complex flow patterns, and unstable flow patterns (P<0.001, respectively). In morphological parameters, maximum size, aspect, projection, size, and volume-to-ostium area ratios were significantly higher in the high OVI group (P<0.001, respectively). In hemodynamic parameters, wall shear stress and wall shear stress gradient were significantly lower, and oscillatory shear index and gradient oscillatory number were significantly higher in the high OVI group (P<0.001, respectively). CONCLUSION: High OVI was associated with rupture status, and morphological and hemodynamic characteristics of ruptured aneurysms. These results indicate that OVI may serve as a valuable hemodynamic parameter for diagnosing rupture status and risks of aneurysms.

Hemodynamic characteristics of hyperplastic remodeling lesions in cerebral aneurysms.

BACKGROUND & PURPOSE: Hyperplastic remodeling (HR) lesions are sometimes found on cerebral aneurysm walls. Atherosclerosis is the results of HR, which may cause an adverse effect on surgical treatment for cerebral aneurysms. Previous studies have demonstrated that atherosclerotic changes had a correlation with certain hemodynamic characteristics. Therefore, we investigated local hemodynamic characteristics of HR lesions of cerebral aneurysms using computational fluid dynamics (CFD). METHODS: Twenty-four cerebral aneurysms were investigated using CFD and intraoperative video recordings. HR lesions and red walls were confirmed on the intraoperative images, and the qualification points were determined on the center of the HR lesions and the red walls. The qualification points were set on the virtual operative images for evaluation of wall shear stress (WSS), normalized WSS (NWSS), oscillatory shear index (OSI), relative residence time (RRT), and aneurysm formation indicator (AFI). These hemodynamic parameters at the qualification points were compared between HR lesions and red walls. RESULTS: HR lesions had lower NWSS, lower AFI, higher OSI and prolonged RRT compared with red walls. From analysis of the receiver-operating characteristic curve for hemodynamic parameters, OSI was the most optimal hemodynamic parameter to predict HR lesions (area under the curve, 0.745; 95% confidence interval, 0.603-0.887; cutoff value, 0.00917; sensitivity, 0.643; specificity, 0.893; P<0.01). With multivariate logistic regression analyses using stepwise method, NWSS was significantly associated with the HR lesions. CONCLUSIONS: Although low NWSS was independently associated with HR lesions, OSI is the most valuable hemodynamic parameter to distinguish HR lesions from red walls.

Glutathione peroxidase 2 in Saccharomyces cerevisiae is distributed in mitochondria and involved in sporulation.

Gpx2, one of three glutathione peroxidase homologs (Gpx1, Gpx2, and Gpx3) in Saccharomyces cerevisiae, is an atypical 2-Cys peroxiredoxin that prefers to use thioredoxin as a reducing agent in vitro. Despite Gpx2 being an antioxidant, no obvious phenotype of gpx2Δ mutant cells in terms of oxidative stress has yet been found. To gain a clue as to Gpx2's physiological function in vivo, here we identify its intracellular distribution. Gpx2 was found to exist in the cytoplasm and mitochondria. In mitochondria, Gpx2 was associated with the outer membrane of the cytoplasmic-side, as well as the inner membrane of the matrix-side. The redox state of the mitochondrial Gpx2 was regulated by Trx1 and Trx2 (cytoplasmic thioredoxin), and by Trx3 (mitochondrial matrix thioredoxin). In addition, we found that the disruption of GPX2 reduced the sporulation efficiency of diploid cells.