Machine Learning Models for Brain Arteriovenous Malformations Presenting with Hemorrhage Based on Clinical and Angioarchitectural Characteristics.

RATIONALE AND OBJECTIVES: This study aims to develop the best diagnostic model for brain arteriovenous malformations (bAVMs) rupture by using machine learning (ML) algorithms. MATERIALS AND METHODS: We retrospectively included 353 adult patients with ruptured and unruptured bAVMs. The clinical and radiological data on patients were collected. The significant variables were selected using univariable logistic regression. We constructed and compared the predictive models using five supervised ML algorithms, multivariable logistic regression, and R2eDAVM scoring system. A complementary systematic review and meta-analysis of studies was aggregated to explore the potential predictors for bAVMs rupture. RESULTS: We found that a small nidus size of <3 cm, deep and infratentorial location, longer filling time, and deep and single venous drainage were associated with a higher risk of bAVMs rupture. The multilayer perceptron model showed the best performance with an area under the curve value of 0.736 (95% CI 0.67-0.801) and 0.713 (95% CI 0.647-0.779) in the training and test dataset, respectively. In our pooled analysis, we also found that the male sex, a single feeding artery, hypertension, non-White race, low Spetzler-Martin grade, and coexisting aneurysms are risk factors for bAVMs rupture. CONCLUSION: This study further demonstrated the clinical and angioarchitectural characteristics in predicting bAVMs hemorrhage.

Genetically predicted tobacco consumption and risk of intracranial aneurysm: a Mendelian randomization study.

Several observational studies have suggested that tobacco consumption is a risk factor for intracranial aneurysms (IAs). We here genetically predict the causal association between specific smoking features and biomarkers for smokers and IA risk. The Mendelian randomization (MR) analysis considered summary statistics from the largest current genome-wide association studies of smoking and IA. The inverse-variance weighted (IVW) method, weighted median method, MR-RAPS, and multiple variants Mendelian randomization (MVMR) were performed to estimate the effect of different smoking features and drinking in IA. We observed significant causal effects of smoking on the risk of both aneurysmal subarachnoid hemorrhage (aSAH) and unruptured IA (uIA). The ORs of IAs based on the IVW method were 1.890 (95% CI 1.486-2.405) of ever smoking regularly. MVMR analysis afforded odds ratios of 1.685 (95% CI 1.136-2.501). In the further subgroup analysis, a similar causal relationship was observed in aSAH. Moreover, our analyses suggested that higher blood cotinine level and cadmium increases aSAH risk, and ORs were 1.235 (95%CI 1.009-1.186) and 1.235 (95%CI 1.046-1.458), respectively. Our study suggests that ever smoking regularly is associated with the IA risk, which includes both uIA and aSAH. Besides, higher blood cadmium and cotinine level may increases IA and aSAH risk. Thus, tobacco control should be promoted as primordial prevention for IAs, and screening for patients with a smoking history is emphasized.

Angioarchitectural features of brain arteriovenous malformation presented with seizures.

Seizures are the second most common manifestations of brain arteriovenous malformations (bAVMs). This study was conducted to investigate the clinical and angioarchitectural features of bAVMs with seizures and provide guidelines for the clinical management of these patients. We collected clinical and radiological data on patients with bAVMs diagnosed by digital subtraction angiography between January 2013 and December 2020 and dichotomized the patients into the seizures and non-seizures groups. We identified differences in demographic and angiographic features. Logistic regression and random forest (RF) models were developed and compared. The diagnostic capacity was assessed using receiver operating characteristic (ROC) curves. A nomogram was constructed, and the clinical impact was determined by decision curve analysis. A total of 414 patients with bAVMs were included in the analysis, of which 78 (18.8%) had bAVM-related seizures. In the multivariable logistic regression model, the location and side of bAVMs were independently associated with seizures. In RF models, the maximal diameter of veins and the cross-sectional area of feeding arteries and draining veins were the most important features. ROC curves showed that the RF model was not better than MLR in predicting seizures. Decision curve analysis revealed that the use of a constructed nomogram to stratify the seizure patients was beneficial at all threshold probabilities in our study. The side and location of bAVMs are specific angioarchitectural features independently associated with the occurrences of seizures with bAVMs.

Molecular feature of arterial remodeling in the brain arteriovenous malformation revealed by arteriovenous shunt rat model and RNA sequencing.

PURPOSE: Morphological research suggested the feeding artery of brain arteriovenous malformation (bAVM) had vascular remodeling under the high blood flow; however, the underlying molecular mechanisms were unclear. METHODS: We constructed 32 simplified AVM rat models in four groups: the control group (n = 6), 1-week high-blood-flow group (n = 9), 3-week high-blood-flow group (n = 7) and 6-week high-blood-flow group (n = 10). The circumference, blood velocity, blood flow, pressure, and wall shear of the feeding artery were measured or calculated. The arterial wall change was observed by Masson staining. RNA sequencing (RNA-seq) of feeding arteries was performed, followed by bioinformatics analysis to detect the potential molecular mechanism for bAVM artery remodeling under the high blood flow. RESULTS: We observed hemodynamic injury and vascular remodeling on the feeding artery under the high blood flow. RNA-seq showed immune/inflammation infiltration and vascular smooth muscle cell (VSMC) phenotype transformation during remodeling. Weighted gene co-expression network analysis (WGCNA) and time series analysis further identified 27 key genes and pathways involved in remodeling. Upstream miRNA and molecular drugs were predicted targeting these key genes. CONCLUSIONS: We depicted molecular change of bAVM arterial remodeling via RNA-seq in high-blood-flow rat models. Twenty-seven key genes may regulate immune/inflammation infiltration and VSMC phenotype transform in bAVM arterial remodeling.

The Transcriptional Landscapes and Key Genes in Brain Arteriovenous Malformation Progression in a Venous Hypertension Rat Model Revealed by RNA Sequencing.

BACKGROUND: Brain arteriovenous malformations (bAVM) are abnormal vascular lesions characterized by direct connections between arteries and veins without an intervening capillary bed. The primary goal for brain AVM treatment is to prevent rupture and hemorrhage; however, the underlying molecular mechanisms are still unknown. METHODS: We constructed venous hypertension (VH) rat model with end-to-end anastomosis of the proximal left common carotid artery and the left distal external jugular vein. Thirty-eight adult rats were randomly assigned to four groups: the 0-week (n=5), the 1-week VH group (n=12), the 3-week VH group (n=9), and the 6-week VH group (n=12). We measured the hemodynamics and diameter of the arterialized veins. An RNA sequencing of arterialized veins was conducted, followed by comprehensive bioinformatics analysis to identify key genes and biological pathways involved in VH progression. The candidate genes from RNA-Seq were validated by RT-qPCR and immunostaining in human tissues. RESULTS: We observed high-flow and low resistance characteristics in VH models. A total of 317 upregulated and 258 downregulated common genes were consistently differentially expressed during VH progression. Thirteen co-expression modules were obtained by WGCNA analysis, and 4 key modules were identified. Thirteen genes: Adamts8, Adamtsl3, Spon2, Adamtsl2, Chad, Itga7, Comp, Itga8, Bmp6, Fst, Smad6, Smad7, Grem1, and Nog with differential expressions were identified using the density of maximum neighborhood component (DMNC) algorithm in Cytohubba. The expression of five potential genes (Adamts8, Adamtsl3, Spon2, Adamtsl2, Itga8) were increased in RT-qPCR, while in human bAVM tissue, the protein levels of Adamtsl2 and Itga8 were significant elevated and Spon2 and Adamtsl3 were decreased. CONCLUSION: The identified gene networks of Adamtsl3, Spon2, Adamtsl2, and Itga8 provided key genes for further intervention.

Medial Gap: A Structural Factor at the Arterial Bifurcation Aggravating Hemodynamic Insult.

Previous studies have reported that intracranial aneurysms frequently occur adjacent to the medial gap. However, the role of the medial gap in aneurysm formation is controversial. We designed this study to explore the potential role of the medial gap in aneurysm formation. Widened artery bifurcations with or without medial gaps were microsurgically created and pathologically stained in the carotid arteries of 30 rats. Numerical artery bifurcation models were constructed, and bidirectional fluid-solid interaction analyses were performed. Animal experiments showed that the apexes of widened bifurcations with a medial gap were prone to being insulted by blood flow compared to those without a medial gap. The bidirectional fluid-solid interaction analyses indicated that artery bifurcations with the medial gap exhibited higher wall shear stress (WSS) and von Mises stress (VMS) at the apex of the bifurcation. The disparity of stress between the gap and no-gap model was larger for widened bifurcations, peaking at 180° with a maximum of 1.9 folds. The maximum VMS and relatively high WSS were located at the junction between the medial gap and the adjacent arterial wall. Our results suggest that the medial gap at the widened arterial bifurcation may promote aneurysm formation.

Comprehensive analysis of immunocyte infiltration and the key genes associated with intraplaque hemorrhage in carotid atherosclerotic plaques.

PURPOSE: To identify key biomarkers associated with intraplaque hemorrhage (IPH). METHODS: We conducted a comprehensive analysis combined with DEGs, xCell, WGCNA, GSEA, and GSVA methods to identify immune infiltration cells and key genes involved in IPH by using GSE163154 from the gene expression omnibus (GEO). E-MTAB-1470 and E-MTAB-2055 from the ArrayExpress database were utilized as the verification datasets. Finally, the candidate hub genes were further validated by RT-qPCR in clinical samples. RESULTS: A total of 280 genes were upregulated and 234 genes were downregulated in GSE163154. Among the upregulated pathways, the lysosome and chemokine signaling pathway were enriched, while the vascular smooth muscle (VSMC) contraction and focal adhesion were downregulated. In addition, ten coexpression modules were obtained by using the WGCNA method and two IPH and immunity-related modules were identified. In total, 454 genes overlapped by DEGs and WGCNA results were imported into Cytoscape to construct a protein-protein network. Eight genes (FCER1G, ITGB2, VAV1, CSF1R, ITGAM, TYROBP, PTK2, and PTPN11) were identified as the IPH-related gene set with area under curves (AUC) of 0.961, 0.905, and 0.857 in GSE163154, E-MTAB-2055, and E-MTAB-1470, respectively. The expression of four genes (ITGB2, VAV1, ITGAM, TYROBP) from our analysis were consistent with RT-qPCR results. CONCLUSION: Eight genes were found to be involved in IPH, and four genes (ITGB2, VAV1, ITGAM, TYROBP) may be an important biomarkers for IPH.

Factors affecting the performance of brain arteriovenous malformation rupture prediction models.

BACKGROUND: In many cases, both the rupture rate of cerebral arteriovenous malformation (bAVM) in patients and the risk of endovascular or surgical treatment (when radiosurgery is not appropriate) are not low, it is important to assess the risk of rupture more cautiously before treatment. Based on the current high-risk predictors and clinical data, different sample sizes, sampling times and algorithms were used to build prediction models for the risk of hemorrhage in bAVM, and the accuracy and stability of the models were investigated. Our purpose was to remind researchers that there may be some pitfalls in developing similar prediction models. METHODS: The clinical data of 353 patients with bAVMs were collected. During the creation of prediction models for bAVM rupture, we changed the ratio of the training dataset to the test dataset, increased the number of sampling times, and built models for predicting bAVM rupture by the logistic regression (LR) algorithm and random forest (RF) algorithm. The area under the curve (AUC) was used to evaluate the predictive performances of those models. RESULTS: The performances of the prediction models built by both algorithms were not ideal (AUCs: 0.7 or less). The AUCs from the models built by the LR algorithm with different sample sizes were better than those built by the RF algorithm (0.70 vs 0.68, p < 0.001). The standard deviations (SDs) of the AUCs from both prediction models with different sample sizes displayed wide ranges (max range > 0.1). CONCLUSIONS: Based on the current risk predictors, it may be difficult to build a stable and accurate prediction model for the hemorrhagic risk of bAVMs. Compared with sample size and algorithms, meaningful predictors are more important in establishing an accurate and stable prediction model.

A retrospective longitudinal study of age-related shifts and deformations in the basilar artery bifurcation.

PURPOSE: Previous studies have indicated that cerebral arterial morphology is linked to aging and some cerebrovascular diseases. However, the mechanisms of morphological changes remain unclear. This study evaluated age-related positional changes in the basilar artery (BA) bifurcation based on longitudinal computed tomography angiography (CTA) data. METHODS: This retrospective study evaluated clinical and imaging data from 72 subjects who underwent two CTA scans between July 2011 and August 2019. Three-dimensional (3D) models were reconstructed for each subject based on the two CTA scans with the longest separating interval. Skull landmarks were used to fuse the two models, and the fused model was used to evaluate positional changes in the BA bifurcation. Univariable and multivariable analyses were used to identify variables that were correlated to BA bifurcation shifting. Pearson's correlation test was used to analyze the correlation between the shifting distance and change in the BA bifurcation angle. RESULTS: Significant differences between aneurysm and non-aneurysm cases were observed in terms of sex (p = 0.004), CTA scan interval (p = 0.023), and BA bifurcation shifting distance (p = 0.007). Multivariable linear regression analysis revealed that the BA bifurcation shifting distance was significantly correlated with the CTA scan interval (p = 0.038) and the presence of aneurysms (p < 0.001). Furthermore, the shifting distance was positively correlated with widening of the BA bifurcation angle (p = 0.002). CONCLUSIONS: Aging-related widening of the BA bifurcation angle may be related to distal shifting of the BA bifurcation's position, and larger distal shifting of the BA bifurcation may be associated with the risk of aneurysm formation.