Quantitative evaluation of hemodynamics after partial embolization of brain arteriovenous malformations

BackgroundTo explore the hemodynamic changes after embolization of arteriovenous malformations (AVMs) using quantitative digital subtraction angiography (QDSA).MethodsWe reviewed 74 supratentorial AVMs that underwent endovascular embolization and performed a quantitative hemodynamic analysis comparing parameters in pre- and post-operative DSA in correlation with rupture. The AVMs were further divided into two subgroups based on the embolization degree: Group I: 0%–50%, Group II: 51%–100%. In the intergroup analysis, we examined the correlations between embolization degree and hemodynamic parameter changes.ResultsA longer time to peak (TTP) of the main feeding artery (OR 11.836; 95% CI 1.388 to 100.948; P=0.024) and shorter mean transit time (MTT) of the nidus (OR 0.174; 95% CI 0.039 to 0.766; P=0.021) were associated with AVM rupture. After embolization, all MTTs were significantly prolonged (P<0.05). The full width at half maximum (FWHM) duration of the main feeding artery was significantly shortened (P<0.001), and several hemodynamic parameters of the main draining vein changed significantly (TTP: prolonged, P=0.005; FWHM: prolonged, P=0.014; inflow gradient: decreased, P=0.004; outflow gradient: decreased, P=0.042). In the subgroup analysis, several MTT parameters were significantly prolonged in both groups (P<0.05), and the MTT increase rate in Group II was greater than in Group I (P<0.05).ConclusionsEmbolization can significantly change the hemodynamics of AVMs, especially when an embolization degree >50% is obtained. Partial embolization may reduce the AVM rupture risk in hemodynamics perspective.

Cotinine Levels Influence the Risk of Rupture of Brain Aneurysms

The relationship between nicotine levels in smokers and brain aneurysm has never been determined. To assess the levels of cotinine in smoker patients with ruptured and unruptured brain aneurysm and the risk of aneurysm rupture we quantified cotinine levels in smoker patients with ruptured or unruptured brain aneurysms. We identified a control group of smokers and nonsmokers without brain aneurysm. Out of 182 participants cerebral aneurysms were statistically significantly associated with smoking (P<0.001) and female sex (P=0.006). Cotinine levels were significantly correlated with both the presence (P=0.009) and the rupture (P=0.002) of brain aneurysms. Compared with nonsmokers, smokers had a 5-fold higher risk of having a brain aneurysm (OR, 5.72; 95% CI, 2.96–11.07; P<0.001). The risk of rupture of brain aneurysms increased by 50% with each cotinine unit and was 4-fold higher with cotinine levels between 4 and 6 (OR, 3.75; 95% CI, 1.48–9.53; P=0.005). With increasing age, the cotinine level decreased (P <0.001, rho= –0.28), declining by 2% with each year of age. In the whole population, the probability of a ruptured aneurysm in patients with cotinine levels between 4 and 6 was higher than in those with levels between 0 and 3 (OR, 5.55; 95% CI, 1.08–28.5; P=0.040). Our results suggest that high cotinine levels in smokers with brain aneurysm, rather than size, are significantly associated with high rupture risk. Cotinine levels decrease with age, possibly reducing the risk of formation and rupture of a brain aneurysm.

Interpretable Machine Learning Model to Predict Rupture of Small Intracranial Aneurysms and Facilitate Clinical Decision

Estimating the rupture risk of small intracranial aneurysms (IAs) to determine whether to treat is difficult but crucial. We aimed to construct and external validation a convenient machine learning (ML) model for assessing the rupture risk of small IAs.1004 patients with small IAs recruited from two hospitals were included in our retrospective research. The patients at hospital 1 were stratified into training (70%) and internal validation set (30%) randomly, and the patients at hospital 2 were used for external validation. We selected predictive features using the least absolute shrinkage and selection operator (LASSO) method, and constructed five ML models applying diverse algorithms including random forest classifier (RFC), categorical boosting (CatBoost), support vector machine (SVM) with linear kernel, light gradient boosting machine (LightGBM) and extreme gradient boosting (XGBoost). The Shapley Additive Explanations (SHAP) analysis provided interpretation for the best ML model.The training, internal and external validation cohorts included 658, 282, and 64 IAs, respectively. The best performance was presented by SVM as AUC of 0.817 in the internal [95% confidence interval (CI), 0.769-0.866] and 0.893 in the external (95% CI, 0.808-0.979) validation cohorts, overperformed than the PHASES score significantly (all P < 0.001). SHAP analysis showed maximum size, location and irregular shape were the top three important features to predict rupture. Our SVM model based on readily accessible features presented satisfying ability of discrimination in predicting the rupture IAs with small size. Morphological parameters made important contributions to prediction result.

Abstract 1122‐000015: Differences of Morphological and Hemodynamic Rupture Risk Factors in Cerebral Aneurysms Between Pre‐ and Post‐Rupture

Introduction : Morphological and hemodynamic characteristics have been reported to be involved in the rupture of cerebral aneurysms. Therefore, geometrical measurements of cerebral aneurysms and blood flow analysis using computational fluid dynamics (CFD) have been conducted. Some previous studies investigated the rupture risk factors from cerebral arterial geometries that were taken before the rupture (pre‐rupture), and the others used geometries taken after the rupture (post‐rupture). However, aneurysm rupture may alter arterial geometries and CFD simulation results. The aim of this study is to evaluate the morphological and hemodynamic alternations due to cerebral aneurysm rupture. Methods : We identified 21 cerebral aneurysms (ICA: 9, MCA: 3, ACA: 4, BA: 3, VA: 2) which had ruptured during the follow‐up terms. Each case had at longest two‐years term between the rupture date and the latest angiographic date before the aneurysm rupture (pre‐rupture). The post‐rupture arterial geometries were acquired preoperatively for subarachnoid hemorrhage. We used the arterial geometries reconstructed from computed tomography angiography or digital subtraction angiography images for conducting morphological measurements and CFD simulations. We performed transient blood flow simulations for two heart pulse cycles in the CFD simulations. We obtained five morphological parameters and 24 hemodynamic parameters considered as the rupture risk factors. Finally, we conducted Wilcoxon’s signed‐rank sum test between the parameters obtained from pre‐ and post‐rupture aneurysms to specify altered parameters due to the aneurysm rupture. We also calculated the change rate (CR) based on the value in pre‐rupture for parameters that had a statistical significance to investigate the alternation in detail. Results : The aneurysmal volume ( V ), height ( H ), aspect ratio ( AR ), and spatial averaged, maximum, and minimum wall shear stress of the aneurysm dome normalized by the spatially averaged wall shear stress of the parent vessel ( NWSSave , NWSSmax , and NWSSmin ) were significantly altered between pre‐ and post‐rupture. In particular, the morphological parameters increased after the rupture (average CR of V , H , and AR were 25.8 %, 13.4 %, and 15.9 %, respectively). These results indicate that the aneurysm shapes tended to increase lengthwise after the rupture. On the other hand, the NWSS tended to decrease (average CR of NWSSave , NWSSmax , and NWSSmin were ‐21.0 %, ‐13.7 %, and ‐22.7 %, respectively). These results imply that the aneurysm rupture altered the aneurysm to a more complicated shape, and thereby the blood flow became stagnated that introduced lower WSS . In contrast, there were some cases in which NWSS increased, and these cases had vasospasm at their parent arteries caused by the rupture (i.e., 5 of 21 cases had vasospasm, and the average CR of NWSSave was 14.1 %). The parent vessel proximal to the aneurysm was shrunk due to the vasospasm, resulting in increased flow velocity and thus increased NWSS . Conclusions : The cerebral aneurysm rupture deformed the aneurysms into longitudinal and led to increased volumes. The NWSSs in CFD simulations using post‐rupture geometries tended to decrease in comparison with pre‐rupture. When studying rupture factors of cerebral aneurysms using geometrical measurements and CFD simulations, special attention should be paid to the clinical image and rupture characteristics standardization criteria.

Abstract 1122‐000130: Prediction of Risk of Rupture of Intracranial Aneurysms in a Latin‐American Population: A Restrospective Study

Introduction : The Population, Hypertension, Age, Size, Earlier Subarachnoid Hemorrhage (SAH), Site (PHASES) score was developed in North America, Europe, and Japan and it is a widely used model in day‐to‐day clinical practice for intracranial aneurysm (IA) rupture risk stratification. Here, we aimed to determine the predictors of aneurysm rupture and assess the components of the PHASES score in a Latin American population. Methods : Four hundred eighty‐six Peruvian patients presented at our institution with ruptured IAs between 2010 and 2020. We retrospectively collected the following variables: age, sex, a hypertension or diabetes mellitus history, previous SAH, the aneurysm size in millimeters (<5, 5–6.9, 7–9.9, 10–19.9, and ≥20), aneurysm morphology (saccular or non‐saccular), neck diameter (≤4 and >4), presence of a pseudoaneurysm, and aneurysm location. We then performed two separate multivariate analysis. For the first one, we included variables using a stepwise approach with a cut‐off p‐value of 0.2 in univariate logistic regression. For the second one, we evaluated the PHASES score components. A p‐value of 0.05 was considered statistically significant. Results : The median age was 56 years old, and 114 females were included. One hundred seventy‐five patients had a hypertension history, 21 had a diabetes history, and 11 had a previous SAH. Seventy‐eight patients had an aneurysm with <5mm, 118 with 5–6.9mm, 125 with 7–9.9mm, 85 with 10–19.9, and 10 patients with an aneurysm >20mm. There were 372 patients with a saccular aneurysm and an associated pseudoaneurysm was found in 197 patients. The most common location was posterior communicating artery (n = 219), followed by the anterior cerebral artery (n = 125), the middle cerebral artery (MCA) (n = 58), branches from the posterior circulation (n = 33), and finally by a paraclinoid aneurysm (n = 33). In our initial multivariate analysis, only the presence of an associated pseudoaneurysm was an independent predictor for aneurysm rupture (OR 7.93; 95% CI 3.45 – 18.25). An age >70 years (OR 1.12; 95% CI 0.3 – 4.12), the male sex (OR 1.39; 95% CI 0.54 – 3.62), a hypertension history (OR 1.14; 95% CI 0.53 – 2.44), a size of 10–20mm (OR 1.46; 95% CI 0.46‐ 4.64), and location in the MCA (OR 1.07; 95% CI 0.25 – 4.57) also predicted a higher rupture risk but without statistical significance. When we performed a multivariate logistic regression of the factors making up the PHASES score, we found that only the age (OR 1.79; 95% CI 1.11‐ 2.88) and a hypertension history (OR 1.61; 95% CI 1.14 – 2.27) were independent predictors of aneurysm rupture. Conclusions : Based on our findings and its limitations, we observed that the presence of an associated pseudoaneurysm was a predictor for aneurysm rupture. Moreover, we found that only two of the five components of the PHASES score were predictors of the event in our population: the age and a hypertension history. Therefore, new research should be carried out in the Latin American population to establish predictors for the development of clinical predictive models in this field.

Difference in Rupture Risk Between Familial and Sporadic Intracranial Aneurysms: An Individual Patient Data Meta-analysis

Objective:We combined individual patient data (IPD) from prospective cohorts of patients with unruptured intracranial aneurysms (UIA) to assess to what extent patients with familial UIA have a higher rupture risk than those with sporadic UIA.Methods:For this IPD meta-analysis we performed an Embase and Pubmed search for studies published up to December 1, 2020. We included studies that 1) had a prospective study design; 2) included 50 or more patients with UIA; 3) studied the natural course of UIA and risk factors for aneurysm rupture including family history for aneurysmal subarachnoid haemorrhage and UIA; and 4) had aneurysm rupture as an outcome. Cohorts with available IPD were included. All studies included patients with newly diagnosed UIA visiting one of the study centers. The primary outcome was aneurysmal rupture. Patients with polycystic kidney disease and moyamoya disease were excluded. We compared rupture rates of familial versus sporadic UIA using a Cox proportional hazard regression model adjusted for the PHASES score and smoking. We performed two analyses: 1. only studies defining first-degree relatives as parents, children, and siblings and 2. all studies, including those in which first-degree relatives are defined as only parents and children, but not siblings.Results:We pooled IPD from eight cohorts with a low and moderate risk of bias. First-degree relatives were defined as parents, siblings and children in six cohorts (29% Dutch, 55% Finnish, 15% Japanese), totalling 2,297 patients (17% familial, 399 patients) with 3,089 UIA and 7,301 person-years follow-up. Rupture occurred in 10 familial patients (rupture rate: 0·89%/person-year; 95% CI:0·45-1·59) and 41 sporadic patients (0·66%/person-year; 95% CI:0·48-0·89); adjusted HR for familial patients 2·56 (95% CI: 1·18–5·56). After adding also the two cohorts excluding siblings as first-degree relatives resulting in 9,511 patients the adjusted HR was 1·44 (95% CI: 0·86–2·40).Conclusion:The risk of rupture of UIA is two and a half times higher, with a range from a 1.2 to 5 times higher risk, in familial than in sporadic UIA. When assessing the risk of rupture in UIA, family history should be taken into account.

Definitions of intracranial aneurysm size and morphology: A call for standardization

Background: Intracranial aneurysms (IAs) are classified based on size (maximal dome diameter) as well as additional parameters such as neck diameter and dome-to-neck ratio (DNR). The neurosurgical literature includes a wide variety of definitions for both IA size and neck classifications. Standardizing the definitions of IA size and wide-neck classifications would help eliminate inconsistencies and potential misunderstandings of aneurysm morphology and rupture risk. Methods: We queried the MEDLINE (EBSCO) database using the terms “unruptured IA” and (“small” or “medium” or “large”) and filtered based on publication date, language, and scholarly journals. The resulting articles and their references were further screened for eligibility. This identified 286 records, of which 104 were excluded, leaving 182 articles for analysis. The review found several different IA size classifications and neck classifications. Results: A review of the existing literature describing size and neck classifications revealed 13 size classifications for small aneurysms, four classifications for medium aneurysms, 15 classifications for large aneurysms, and one classification for giant aneurysms. There were also seven different wide-neck classifications found. Conclusion: It is imperative that a standardization in classification be implemented to help interventionalists make the most informed decisions regarding emerging treatment options as new endovascular technologies and devices are emerging with indications based around these classifications. Based on the database findings, this article recommends standardized quantitative measurement ranges for IA size and neck classifications.

TP5.2.13 Perceived risk of pre-operative abdominal aortic aneurysm rupture due to pulmonary function testing: all blown out of proportion?

Aim Though contraindicated in some respiratory laboratories due to a perceived increased risk of aneurysm rupture, pulmonary function tests (PFT) and cardiopulmonary exercise tests (CPET) are used commonly in preoperative assessment of patients with abdominal aortic aneurysm (AAA) considered for elective repair. We examine whether preoperative PFT/CPET are associated with increased AAA rupture risk. Methods A retrospective analysis of a prospective database of patients with AAA ≥50mm who had PFT/CPET/both as part of their preoperative assessment from November 2012 to December 2020 in a single vascular hub before elective AAA repair (open/endovascular). Patients with AAA &lt;50mm, or those who did not have PFT/CPET preoperatively were excluded from analysis. Baseline demographics, AAA size, type of repair, time from PFT/CPET to repair, rupture rate and 30-day post-operative mortality are presented. Results A total of 410 patients (mean age 74 ± 8 years, 361 (88%) male) were included in the analysis. The mean AAA size was 63 ±11mm. 350 (85%) patients underwent endovascular repair. 220 (53%) patients had PFT, 69 (17%) had CPET and the remaining 121 (30%) had both PFT/CPET preoperatively. The median time from PFT/CPET to repair was 36 days (IQR 48 days), with no preoperative AAA rupture in this cohort. 30-day post-operative mortality was 3%. Conclusions This study strongly suggests that PFT and CPET are not associated with increased risk of AAA rupture and are therefore safe for assessing respiratory function in patients considered for AAA repair without adding to pre-operative rupture risk.