Aneurysmal subarachnoid hemorrhage, resulting from the rupture of cerebral aneurysms, carries a poor prognosis exceeding 30% despite advancements in treatment options (1). Direct surgery to prevent rupture is recommended under specific conditions, with neck clipping as an established surgical technique. Intraoperative rupture, the most significant complication of neck clipping, occurs at rates ranging from 3 to 50% (2, 3). This variability depends on the complexity of the aneurysm’s vascular anatomy and the challenges of the surgical technique.

The risk of this complication is exceptionally high during the dissection phase of aneurysm exposure. Factors contributing to this include the technical difficulty of exposing the aneurysm, brain swelling, and the extent of manipulation during clipping (2, 3). The presence of a bleb in a cerebral aneurysm significantly increases the risk of intraoperative rupture during dissection (4). Furthermore, thinning of the vascular wall near the aneurysm neck, especially with microaneurysms, has been recognized to heighten intraoperative risk. These changes often elude detection in preoperative three-dimensional computed tomography angiography (3D-CTA) and become apparent intraoperatively, potentially confusing the surgeon and increasing procedural risk.

The number of reports related to computational fluid dynamics (CFD) analysis in the field of cerebral aneurysms has been increasing in recent years, and it is gaining more attention. There has been a substantial accumulation of research regarding CFD analysis specifically targeting cerebral aneurysms. These studies have primarily focused on aspects such as rupture risk, growth risk, hemodynamic changes related to treatment procedures, and post-treatment recurrence. Consensus has been gradually reached regarding parameters such as wall shear stress, wall pressure, and the vector of streamlines in aneurysms (5, 6). We have also conducted CFD analyses focusing on cerebral aneurysms themselves. Through our experience, we discovered abnormalities in CFD findings at sites of thinning of the parent artery wall and at locations where microaneurysms form. Based on these findings, we retrospectively analyzed the parent artery adjacent to cerebral aneurysms using preoperative 3D-CTA (7, 8) and defined this unique CFD finding as the “parent artery radiation sign” (PARS) (9). This study is significant because it focuses on morphological changes in the parent artery adjacent to the cerebral aneurysm.

However, previous studies with limited case numbers have not thoroughly examined the relationship between individual case factors and the significance of PARS in predicting microaneurysms preoperatively. Therefore, this study aims to investigate these aspects in detail, using a more extensive case series to evaluate the sensitivity and specificity of PARS for predicting microaneurysm formation in the parent artery near cerebral aneurysms.

This single-center, retrospective observational study reviewed 89 cases of unruptured cerebral aneurysms treated with direct clipping between May 2020 and April 2022. Background factors compared included age, gender, aneurysm diameter, presence of multiple aneurysms, prior treated aneurysms, history of subarachnoid hemorrhage (SAH), family history of cerebral aneurysms, underlying conditions (hypertension, diabetes, dyslipidemia), statin use, and lifestyle factors (alcohol consumption and smoking habits). As in our previous report, aneurysm locations were limited to the bifurcation of the middle cerebral artery (MCA) to facilitate observation and ensure compatibility with computational fluid dynamics (CFD) analysis.

The CFD analysis method has been previously reported. In summary, blood flow impingement on cerebral aneurysm walls is associated with wall thinning (10). Streamlines visualizing blood flow pathways can identify impingement and turbulence sites, which may promote vessel wall thinning and aneurysm formation (11). PARS was defined by three CFD findings indicating blood flow impingement and elevated vessel wall pressure relative to surrounding areas: (1) streamline collisions with the vessel wall adjacent to the aneurysm; (2) radial dispersion of wall shear stress vectors at the same site; and (3) increased wall pressure. We compared intraoperative findings between the PARS-positive group (PARS group) and the PARS-negative group (control group).

Three-dimensional vascular images for CFD analysis were generated from CTA scans (TOSHIBA Aquilion One 320 columns) using the Ziostation2 image-processing workstation (Ziosoft, Tokyo, Japan) to extract blood vessels around the aneurysm. CFD images were then created using Hemoscope Project Manager 2015 (EBM/AMIN, Tokyo, Japan). Hemoscope has the advantage of enabling CFD analysis for target cases in a relatively short amount of time in clinical settings. There have been prior studies conducted by our research group as well as reports from other groups (12–14). Specifically, this software consolidates essential workflows for CFD analysis—including imaging, rendering, modeling, meshing, blood properties, boundary conditions, computation, visualization, and analysis—facilitating the full automation of CFD blood flow simulations. Prior to analysis, preparation involves extracting the region surrounding the target aneurysm from vascular images obtained via CTA. Upon launching the software and selecting either the steady flow or pulsatile flow mode, processes such as mesh generation, input of blood properties, setting of boundary conditions, selection of discretization schemes, and control of iterative computations are all automatically managed within the software. In this study, all CFD analyses were conducted in steady flow mode. Cases were excluded if they involved suspected dissecting aneurysms, incomplete CFD data, or non-circumferentially observed aneurysms. Four neurosurgeons familiar with CFD analysis and unruptured aneurysm surgery were randomly assigned to evaluate CFD findings and intraoperative outcomes related to wall shear stress vectors, with agreement rates calculated. Microaneurysms in the parent artery adjacent to the aneurysm were defined as microaneurysms located within 5 mm proximal or distal to the primary cerebral aneurysm. Discrepancies in PARS status or intraoperative microaneurysm findings were resolved through secondary CFD evaluation by a different assessor and an independent, blinded review of intraoperative findings. All procedures in this study adhered to the l9atest Declaration of Helsinki guidelines and were approved by the institutional ethics committee (No. HM24–061).

During the study period, 89 patients with 94 unruptured middle cerebral artery (MCA) aneurysms underwent open clipping surgery. After applying the exclusion criteria, 82 patients with 87 aneurysms were included in the analysis, as depicted in the patient selection flowchart (Figure 1).

Background factors for the 82 patients are presented in Table 1.

According to the CFD analysis results, 25 aneurysms (28.7%) were categorized as PARS-positive, while 62 aneurysms (71.3%) were classified as PARS-negative (control group).

Intraoperative findings identify microaneurysms on the parent artery adjacent to the aneurysm in nine of 87 MCA aneurysms, corresponding to an incidence rate of approximately 10% for microaneurysm formation on adjacent vessel walls. All cases with microaneurysm formation underwent additional clipping. Of the nine aneurysms with microaneurysm formation, eight were in the PARS group, while the remaining one was in the control group.

Table 2 summarizes the relationship between PARS positivity and the presence of microaneurysms on vessel walls adjacent to aneurysms. The sensitivity and specificity analysis of PARS positivity for microaneurysm detection yielded a sensitivity of 89% and a specificity of 78%. The positive likelihood ratio for PARS positivity in detecting microaneurysm formation on vessel walls was 4.1, while the negative likelihood ratio was 0.142.

Table 3 presents a univariate logistic regression analysis comparing background factors with the presence of PARS in the nine aneurysms with microaneurysms. PARS positivity was statistically significant for microaneurysms, while no other factors showed statistical significance. Multivariate analysis could not be performed due to minimal intergroup differences for factors other than PARS and insufficient cases. For patients with multiple MCA aneurysms, the aneurysm size shown in the table corresponds to the larger aneurysm. However, separate analyses of the smaller aneurysm showed no statistically significant differences.

Previous studies analyzing cerebral aneurysms using CFD have reported findings predicting wall irregularities of the aneurysm (15–17). In contrast, this study focuses on the vessel wall outside the aneurysm, demonstrating the potential of CFD analysis to predict wall thinning and microaneurysm formation in the parent artery near the aneurysm, which were not detected by preoperative vascular imaging. The radial dispersion of wall shear stress vectors indicates localized force application, particularly susceptible to microaneurysm formation (18). Thus, PARS may enhance sensitivity by highlighting areas where relatively strong forces are applied to the vessel wall. As defined in our previous research, PARS’s ability to predict unpredictable morphological changes in vessel walls with a sensitivity and specificity of 89 and 78%, respectively, is highly valuable for developing treatment strategies for unruptured cerebral aneurysms.

Assessing this method, alongside routine preoperative imaging, could be valuable in planning safe brain retraction and approaches to access the aneurysm, potentially improving surgical outcomes for unruptured cerebral aneurysms.

In recent years, endovascular treatments such as coil embolization, intrasaccular flow disruption, or flow diverter treatment have become more popular than surgical clipping for managing unruptured cerebral aneurysms (19–22)—however, coil embolization and intrasaccular flow disruption target only the aneurysm for occlusion. Flow diverter treatment, a device placed in the parent artery, may aid in repairing the parent artery near the aneurysm. Nonetheless, the use of this treatment for bifurcation aneurysms, such as the middle cerebral artery aneurysms in our study, remains debated (23, 24). Direct clipping may be preferable in cases like ours, where microaneurysms are present in the parent artery near the aneurysm neck (25). Thus, positive PARS findings from pre-treatment CFD analysis could aid in selecting appropriate treatment techniques.

This study demonstrates that computational fluid dynamics (CFD) analysis, mainly using the parent artery radiation sign (PARS), is valuable for predicting microaneurysm formation in the parent artery near unruptured cerebral aneurysms. With a sensitivity of 89% and specificity of 78%, PARS can guide preoperative planning for neck clipping, potentially reducing intraoperative rupture risk and improving surgical outcomes. Despite its promise, the study’s retrospective, single-center design presents limitations, necessitating further research with more extensive multicenter cohorts. Integrating CFD analysis into routine preoperative assessments may improve the safety and effectiveness of surgical interventions for cerebral aneurysms.