INTRODUCTION
The outcomes of Woven EndoBridge (WEB) embolization are improving, with adequate occlusion rates reaching up to 80% at the 5-year follow-up [1-3]. However, approximately 20% of embolized aneurysms still exhibit inadequate occlusion [1,3]. Additionally, the retreatment rate exceeds 10% during the follow-up period, although some studies have reported lower retreatment rates [1,2,4,5].
The oversizing strategy, which involves increasing lateral compression of WEB against the aneurysm's side wall, is considered a key factor in selecting the appropriate WEB size for aneurysm treatment [3,6,7]. Despite this strategy, specific criteria have not yet been established to ensure adequate occlusion of aneurysms with varying sizes and shapes. Therefore, a more precise and practical selection tool must be developed to identify the most suitable WEB size. This improvement aims to enhance the final outcomes of aneurysm treatment and reduce WEB exchange rate, thereby minimizing the risk of selecting undersized or oversized WEBs.
The correlation of volumetric parameters to specific aneurysm configurations is essential because WEB device did not completely fill the entire volume of the lobulated aneurysm. Therefore, we categorized 5 shape types of aneurysms based on symmetric versus asymmetric shapes and developed a WEB index, defined as the ratio of WEB volume to aneurysm volume. We also present how this index is associated with effective aneurysm occlusion.
MATERIALS AND METHODS
Sixty consecutive aneurysms in 52 patients (72% female; mean age 60, range 37–84) with unruptured cerebral aneurysms who underwent endovascular WEB treatment were included from a prospectively collected hospital neurovascular database from December 2022 to March 2025. Inclusion criteria encompassed patients with aneurysms who underwent treatment with WEB. Patients with multiple aneurysms were also included if at least 1 aneurysm was treated with WEB.
Patients diagnosed with an aneurysm who were not treated with WEB and those who did not complete at least 3 months of follow-up were excluded from the study. The cohort in this research has not been included in any other studies.
This study cohort did not include any ruptured aneurysms, as there were no cases indicating the need for treatment during the study period. The embolization procedure followed the same protocol as previously described [8-11]. Flow diverters can be utilized as an alternative treatment option; however, their use is not permitted in Korea because the Health Insurance Review and Assessment Service has not approved them for aneurysms included in this study.
We measured the largest diameter, mean width along 2 different axes, height, neck, ASPECT ratio, and dome-to-neck ratio from the 2- and 3-dimensional angiograms of aneurysms. Angiographic results were evaluated using a 3-grade scale based on magnetic resonance angiography (MRA) conducted 1 day after embolization, as well as the post-procedural angiogram. The grading scale included complete obliteration, presence of a residual neck, presence of a residual sac, or evidence of recurrence. The post-procedural angiogram was not solely used to assess the results of embolization, as the contrast agent rapidly filled the aneurysm and WEB, making it difficult to determine the exact status of aneurysm occlusion based on the post-procedural angiogram alone.
Outcome evaluation was assessed through post-embolization angiographic results, any events as previously described [9], and the final neurological status following the procedure. Additionally, follow-up MRA or additional catheter angiography were conducted if any significant recurrence was suspected on the MRA.
Residual and/or recurrence of the treated aneurysms were assessed using MRA, as it is recognized as an effective screening tool with low sensitivity, high specificity, and a high positive predictive value for detecting aneurysm remnants [12,13]. We opted for MRA instead of catheter angiography because the latter often fails to accurately reveal aneurysm status, particularly when the contrast agent fills WEB [14]. Follow-up MRA scans were conducted the day after the procedure, at 3 months, 9 months, and then annually.
We categorized abnormal signal intensities in WEB-embolized aneurysms as either major or minor signal intensity abnormalities. Minor signal intensities were regarded as benign MRA features indicative of a stable condition, whereas major signal intensity abnormalities were associated with recurrence as evidenced by catheter angiography and/or the need for retreatment.
The final neurological status of patients was evaluated using the modified Rankin scale, which ranges from 0 (normal) to 6 (death). This assessment was conducted at the time of discharge following the procedure and during of the clinical follow-up.
We presented the technical details, clinical status before and after the procedure, angiographic results, and the incidence of recurrence and retreatment during the follow-up period.
Categorization of Aneurysm Shape Types
Based on the contour of symmetric versus asymmetric aneurysms, 5 types of aneurysm shapes were categorized along with their incidence in this study (Figs. 1, 2). Aneurysm shapes were defined based on 3 angled views (tangential, headon, and down the barrel) observed in 2-dimensional and/or 3-dimensional angiograms (Figs. 3, 4). The spherical shape features a round contour with a pronounced acute shoulder at the neck margin. The column shape presents as an elongated tubular contour with a wide neck. The submarine or sweet potato-shaped aneurysm lies between symmetric and asymmetric classifications; it appears symmetric along 1 axis but is fundamentally asymmetric due to a significant difference in width along the other axis (Fig. 3). The boots or Beosun (traditional Korean socks) shaped aneurysm exhibits a deviated dome (Fig. 4). The mitten-shaped aneurysm has a lobulated contour with an asymmetrically widened and uneven sac at the neck and/or body (Figs. 1, 2E, 5).
Volume Measurement of Each Aneurysm
The column volume is calculated using 3 volume parameters: 2 widths and 1 height. We consider the aneurysm volume to be 33% larger than the sphere volume. In symmetric aneurysms, WEB volume corresponds directly to the aneurysm volume. In asymmetric aneurysms, it is necessary to trim parameters to eliminate the lobulated portions of the aneurysm that are not filled by WEB. Since exact measurements cannot be obtained in a single assessment, multiple calculations are required to determine the most suitable aneurysm volume. The type of filed aneurysm can also serve as a reference for measuring subsequent aneurysms of the same type, allowing cumulative experience with WEB sizing to improve fitting results for the target aneurysm.
We developed a volume measurement formula based on a sphere and a column (cylinder), which maintains a proportional volumetric relationship with a volume ratio of 2:3. To estimate the appropriate aneurysm volume, we compared 2 volume parameters (sphere and column) derived from the size and shape of the aneurysm as observed in 2- and/or 3-dimensional angiograms. The diameters in 2 axial directions of the aneurysm body, along with the height (or depth) of the aneurysm, were utilized for volume calculation. We created a volume calculation process coded in Python and executed on Google Colab to expedite the measurement process and enable real-time responses.
To determine the core volume—the actual aneurysm volume that the WEB device would occupy—it is necessary to exclude the lobulated portions of the aneurysm. For example, this may involve removing 2 submarine-shaped sections, 1 boot-shaped section, or the thumb-shaped section in a mitten-shaped aneurysm.
Mitten-shaped aneurysms can exhibit several different patterns depending on where the thumb portion connects to the finger portion (the aneurysm body) (Fig. 5). In our study, the thumb part was usually connected to the finger part near the aneurysm neck, making neck coverage a concern when measuring aneurysm volume. The core volume measurement of mitten-shaped aneurysms is essentially the same as that of column-shaped aneurysms. If the inflow zone of the aneurysm abuts the central part of the WEB device, where the metal density is very compact, the aneurysm neck remnant at the thumb portion can regress after properly positioned WEB embolization. This differs from coil embolization, where compact packing is typically mandatory.
WEB Index Measurement
When the most suitable WEB was selected and utilized for aneurysm treatment, WEB index was calculated using the following formula.
WEB Index is a volume ratio that compares the volumes of WEB and the aneurysm. When WEB Index exceeds 1, it indicates that WEB is larger than the aneurysm (a condition known as over-fitting or over-sizing). Conversely, when WEB Index is less than 1, it signifies that WEB is smaller than the aneurysm (referred to as under-fitting or under-sizing).
Statistical Analysis
A statistical analysis using box plots was conducted to compare WEB index, providing the median and interquartile range for both symmetric and asymmetric shape types. The data were organized and analyzed using the Statistical Package (KESS; version V1.25.06). Descriptive statistics were performed to evaluate the frequency, mean, median, standard deviation, and range of the studied variables. Bivariate analysis was conducted using Pearson’s chi-square test or Fisher’s exact test to compare the association between aneurysm shape type and outcomes, including stent-assisted WEB embolization, sac remnant, and/or recurrence. P-values less than 0.05 were considered statistically significant.
RESULTS
The incidence of symmetric aneurysms was 65% (n=39), while asymmetric aneurysms accounted for 35% (n=21). The locations of the aneurysms included the M1 bifurcation (n=26), the anterior communicating artery (n=24), the basilar artery (n=8), and the terminal internal carotid artery (n=2). The largest diameter of the aneurysms had a mean of 4.7 mm (range: 2.4–12.0 mm). The dome-to-neck ratio had a mean of 1.4 (range: 0.8–3.9), and the ASPECT ratio had a mean of 1.3 (range: 0.7–2.8). Among the 5 shape types, the column shape was the most prevalent (n=32, 53%), followed by submarine (n=9, 15%), sphere (n=7, 12%), boots (n=6, 10%), and mitten (n=6, 10%) shapes (Fig. 1).
The box plot of the median WEB index was 0.96, with an interquartile range of 0.23, indicating that WEB volume was nearly equal to the aneurysm volume (Fig. 6A). There was a difference in the interquartile ranges between symmetric and asymmetric aneurysm shapes (Fig. 6B, C). Specifically, the interquartile range was wider in the upper part of asymmetric aneurysms than in symmetric aneurysms, suggesting that size is significant in symmetric aneurysms, while shape is more important in asymmetric aneurysms.
The procedural results included the following: WEB only (n=36), exchange into smaller size (n=3), exchange into larger size (n=5), stent-assisted WEB (n=15), and guide-wire and/or microcatheter protection (n=5).
Asymmetric shape was associated with stent-assisted procedures (oversizing) (P=0.029) and/or recurrence (undersizing) (P=0.033). The largest diameter, mean width measured along 2 different axes, height, neck, ASPECT ratio or dome-to-neck ratio did not show any statistical significance with any dependent variables.
Abnormal signal intensities in WEB-embolized aneurysms were identified in 11 patients. There were 3 major and 8 minor signal intensity abnormalities observed during a mean follow-up of 8 months for MRA (range: 3–21 months) and a mean of 12 months for clinical evaluations (range: 3–30 months). Major recurrences included one case of compression (n=1) and 2 cases of deviation (n=2) of WEB. Two deviated WEBs with sac recurrence were retreated using coiling. Minor recurrences included in-WEB filling (n=4), small neck filling (n=3), and small sac filling (n=1). We classified these 8 minor signal intensity abnormalities as benign MRA features in stable conditions. Three major signal intensity abnormalities were confirmed to have recurrences (n=3, 5%) and required retreatment (n=2, 3.3%).
There was no deterioration in neurological status among all patients following the procedure. One death at 7 months was attributed to a cardiac event related to pre-existing coronary vessel disease.
DISCUSSION
The classification of aneurysm shape types in this study indicates that analyzing aneurysm shape is crucial for WEB embolization. Determining the mean width of an aneurysm after measuring its volume is particularly important, especially for symmetric shape types. Conversely, additional shape-type analysis is essential for asymmetric aneurysms, necessitating modifications to the volume measurement based on the specific shape type. Therefore, it is advisable to determine WEB size based on the measured volume and then select WEB size that offers the maximum mean width. In this regard, WEB index, which is based on column-based volume measurement, proved to be valuable for evaluating aneurysm volume and selecting the appropriate WEB size.
The selection of an appropriate WEB size is the most critical aspect of the decision-making process in aneurysm WEB embolization, particularly given the various shapes of aneurysms. Research suggests that oversizing WEB width in relation to the average width of the aneurysm is highly predictive of improved occlusion rates [3]. Our study revealed that volume-based oversizing of WEB compared to the aneurysm volume tends to result in a higher occlusion rate and a reduced retreatment rate, even though stent-assisted WEB embolization is more commonly performed.
WEB position within the aneurysm is particularly important, especially in cases of asymmetric aneurysms. The longitudinal axis of WEB can be adjusted, particularly in boot-shaped aneurysms, as WEB can be elongated in the longitudinal direction when compressed by a narrow width.
In this study, aneurysm volume was measured using a self-calculation method based on a specific formula. To facilitate the process, a coded Python class function was embedded in a smartphone, allowing for measurements to be taken whenever necessary. This approach made the measurement process faster, easier, and more adaptable than using traditional software programs. Volume measurements can also be obtained through software programs such as Sim&Size (Sim&Cure) or MPNeuro (Medipixel), or by utilizing volume measurement programs integrated into angiographic equipment, such as Syngo Aneurysm Guidance (Siemens Healthineers) or SmarktCT (Philips). However, it remains uncertain what range of error should be permitted and/or compensated for in the decision-making process regarding WEB size selection.
Sim&Size and MPNeuro are simulation software programs based on large datasets. They can predict not only how the WEB device will be deployed in aneurysms with irregular shapes but also estimate the likelihood of aneurysm occlusion [15]. Artificial intelligence-based simulation software represents a distinct and increasingly relevant category of tools for aneurysm treatment planning in the future.
There are several factors to consider when determining the optimal choice of a WEB for the complete obliteration of aneurysms. Among these factors, the relationship between aneurysm volume and WEB volume, known as WEB index, has emerged as the most significant variable. However, it is also essential to consider aneurysm configuration and the physical properties of WEB, as these elements may mutually influence each other during the healing process [16,17]. The deformation of WEB within the aneurysm is particularly dependent on the aneurysm‘s configuration, especially in specific directions [14]. Prolate (column-shaped) and oblate (submarine-shaped) ellipsoids are applied differently, indicating that such adaptation may require more experience rather than relying on a general rule.
The shape classification in this study was based on the cohort of cases encountered by the authors. Therefore, some aneurysms may not fit into this classification; for example, mushroom-shaped, clover-shaped, or heart-shaped aneurysms. The study by Tanabe et al. [18] showed that the median ideal Woven EndoBridge-aneurysm volume (iWAVe) ratio, which is similar to the WEB index used in this study, was 1.0 (range 0.76–1.31) in the successful group and 1.27 (0.58–1.89) in the unsuccessful group. However, these aneurysm volumes were retrospectively calculated in cases of successful and unsuccessful deployment without any follow-up study or outcome confirmation.
Although we used MRA as the primary imaging modality during the follow-up period, MRA alone might not provide sufficient resolution to assess residual filling at the neck or to accurately differentiate between WEB Occlusion Scale type B (WOS-B, indicative of a concavity at the WEB base) and type C (WOS-C, indicative of a neck remnant) [19] or The Beaujon Occlusion Scale Score Grade 1 (BOSS Grade 1 which signifies opacification inside the WEB) and Grade 2 which denotes a neck remnant [20,21]. In cases suggestive of major recurrence, further evaluation may require catheter angiography and, preferably, high-resolution cone-beam computed tomography.
This study incorporated a decision-making process to determine the correct answer regarding volume measurement and WEB size selection. The complex relationship between various physical and biological factors involving the WEB device and the aneurysm wall means that a clear understanding remains uncertain until long-term follow-up can confirm the final outcome. Decision-making has likely improved as experience accumulated during the study period. Therefore, this conceptual methodology made it difficult to define strict dependent and independent variables, which represents a limitation of this study.
CONCLUSION
This study aims to identify 5 representative types of aneurysm shapes and to assess the significance of WEB index in relation to symmetric and asymmetric aneurysm shapes. Aneurysm size, calculated from column volume, was significant for symmetric aneurysms, while the modification of aneurysm volume based on shape was important for asymmetric aneurysms. It appears necessary to select a range of WEB sizes based on calculated aneurysm volume using an experience-based conceptual approach before choosing WEB with the maximum mean width.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Print
