2025 年 32 巻 8 号 p. 994-1005
Aim: Branch atheromatous disease (BAD), characterized by the occlusion of perforating branches near the orifice of a parent artery, often develops early neurological deterioration because the mechanisms underlying BAD remain unclear. Abnormal wall shear stress (WSS) is strongly associated with endothelial dysfunction and plaque growth or rupture. Therefore, we hypothesized that computational fluid dynamics (CFD) modeling could detect differences in WSS between BAD and small-vessel occlusion (SVO), both of which result from perforating artery occlusion/stenosis.
Methods: This cross-sectional observational study included consecutive patients admitted to our institution within 7 days after symptom onset who met the following criteria: absence of stenosis/occlusion in the intracranial major arteries on brain magnetic resonance angiography (MRA) or extracranial carotid arteries on carotid ultrasonography. The WSS and blood flow velocity in the M1 segment of the middle cerebral artery were analyzed using CFD based on MRA.
Results: The number of patients with a WSS ratio (ipsilesional/contralesional) of ＞1 was significantly higher in patients with BAD (n = 27) than in those with SVO (n = 27) [20 (74.1%) vs. 11 (40.7%), p = 0.013]. Higher WSS on ipsilesional M1 than on contralesional M1 was an independent risk factor for BAD (adjusted odds ratio 4.38, 95% confidence interval 1.29–14.82, p = 0.018). Blood flow velocity in the M1 segment was not associated with BAD.
Conclusions: In patients with BAD, higher M1 segment WSS on CFD can be a risk factor for the development of vulnerable plaques in branch orifices. Moreover, the use of CFD may contribute to the diagnosis of BAD.
As first described in 1989 by Louis Caplan1), branch atheromatous disease (BAD) is characterized by the occlusion of perforating branches that supply deep brain structures, including the internal capsule/basal ganglia or pons near the orifice of a parent artery due to microatheroma or junctional plaques2). The clinical manifestations of BAD can vary and often lead to progressive motor and sensory deficits. Notably, the high frequency of early neurological deterioration is one of the most important issues in patients with BAD2-4). However, BAD is difficult to distinguish from lacunar infarction, because both conditions usually present as classical lacunar syndrome in the hyperacute phase. Furthermore, there is no established effective therapy for BAD, likely because the precise mechanisms of BAD are not fully understood2-4) due to only few pathological studies have focused on microatheroma or junctional plaques5, 6).
Wall shear stress (WSS) is the tangential stress of blood flowing on the arterial wall and exerts a major force on the endothelial surface. Alterations in hemodynamic status, which can be reflected by WSS or flow velocity (FV), play a key role in endothelial dysfunction, plaque growth or rupture, and arterial remodeling7). Low WSS contributes to the initiation and progression of atherogenic plaques, whereas high WSS may enhance plaque vulnerability and increase the risk of plaque rupture8). In this regard, computational fluid dynamics (CFD) modeling has attracted increasing attention in cardiocerebrovascular medicine, given that it provides detailed patient-specific hemodynamic information and allows reconstruction of conventional imaging modalities, such as magnetic resonance angiography (MRA)9, 10) and computed tomography angiography11, 12), which are commonly available in clinical practice. CFD modeling simulates cerebral blood flow and quantifies cerebral hemodynamic metrics as our case report shows12). In this study, we aimed to investigate the pathophysiology of BAD by assessing WSS and FV in the middle cerebral artery using a CFD model based on brain MRA in patients with BAD.
We included patients admitted to the National Cerebral and Cardiovascular Center in Japan due to acute BAD or small-vessel occlusion (SVO) within 7 days after symptom onset. This cross-sectional study conformed to the principles outlined in the Declaration of Helsinki and was approved by the institutional research ethics committee (approval number: R20113). To respect patients’ autonomy and preserve the voluntary nature of participation, an opt-out consent was employed for this study. This approach meant that participants were included in the study unless they expressed their decision to be excluded. The reasons to adopt an opt-out consent were as follows: (1) it was not feasible to obtain express consent from each patient because the visit to our institution had already finished in some patients. (2) The data utilized in this study had already been anonymized, indicating minimal or no risk to the participants. In lieu of obtaining individual informed consent, information about this study was provided on the website of our institution (http://www.ncvc.go.jp/research/clinical-research/) and made it readily accessible to patients. This ensured that patients easily had the opportunity to decline participation in this study if they wished to do that.
The CFD parameters of patients with BAD and SVO were compared. Patients with SVO served as the control. Inclusion of patients was systematically performed based on our defined criteria. First, we collected data from 27 consecutive patients with acute BAD or SVO located in the territory of middle cerebral artery dating back from December 2022 (Supplementary Fig.1). Second, BAD and SVO were diagnosed through brain MR imaging within 7 days after admission. The definition of SVO was based on the Trial of ORG 10,172 in Acute Stroke Treatment (TOAST) classification. BAD was defined according to neuroradiological diagnostic criteria, as previously described13). Briefly, BAD of the lenticulostriate arteries was defined as infarcts ＞10 mm in diameter on axial slices and visible in ≥ 3 axial slices of 7 mm thick13). Third, patients were included if they did not have any stenotic lesions in the intracranial internal carotid and middle cerebral arteries on 3 Tesla brain MRA or in the extracranial carotid artery on carotid ultrasonography. Data on neurological severity, such as the National Institutes of Health Stroke Scale (NIHSS) score and early neurological deterioration, medical history, and blood tests, were collected from the patients’ medical records. Early neurological deterioration was defined as an increase in the NIHSS score by ≥ 2 within 72 h after stroke onset14). The primary study outcome was the association between WSS on M1 and BAD.
Abbreviations: NCVC, National Cerebral and Cardiovascular Center; SVO, small-vessel occlusion; MRA, magnetic resonance angiography; ICAS, intracranial major artery stenosis; BAD, branch atheromatous disease.
Hemodynamics in the intracranial arteries was simulated by CFD to obtain the WSS distribution in the M1 segment of the middle cerebral artery (Fig.1). A geometric model of the intracranial artery was created from brain MRA images and was then divided into numerical meshes, while carefully allocating finer meshes near the wall. The assumptions were a steady flow and that blood was an incompressible Newtonian fluid with a density of 1,050 kg/m3 and a dynamic viscosity of 0.004 Pa/s. As boundary conditions, flow rates were assigned to all edge surfaces, except for one outflow edge surface for which zero pressure was given. The flow rates were estimated based on the equivalent diameter of the edge surface15). The blood vessel was assumed to be rigid, and a no-slip condition was applied to the wall. CFD was implemented using scFLOW ver. 2024 (MSCsoftware, Japan). The WSS was calculated from the obtained velocity field, according to Newton’s law of viscosity. Finally, spatially averaged WSS values on the upper and lower walls of the M1 segment were calculated.
Abbreviations: CFD, computational fluid dynamics; MRA, magnetic resonance angiography; WSS, wall shear stress
The ABC/2 method is considered an effective way of measuring infarct volume with diffusion-weighted imaging16, 17). Specifically, infarct volume was calculated using the formula 0.5×A×B×C where (A) represented the largest diameter of the largest infarct area on the visually selected slice, (B) represented the largest diameter perpendicular to the line above, and (C) represented the total number of slices showing the infarct multiplied by slice thickness of 7 mm.
Statistical AnalysisContinuous variables were presented as mean±standard deviation and categorical variables as frequencies with percentages. Continuous variables were analyzed using Student’s t-test, and categorical variables were analyzed using the chi-square test. To identify the independent risk factor of BAD, we conducted multivariate logistic regression analyses, in which we adjusted for age and sex in model 2 and for age, sex, hypertension, diabetes mellitus, and dyslipidemia in model 3, with calculations of adjusted odds ratio (aOR) and 95% confidence interval (CI). The univariate logistic regression model was designated as model 1. The WSS and FV ratios were calculated by dividing the ipsilesional value by the contralesional value, respectively. An absolute value of WSS on M1 below the lower 95% confidence limit in patients with SVO was considered low. Therefore, a WSS on M1 of ＜1.2 Pa was considered low18).
All statistical tests were two-sided, and p＜0.05 was considered statistically significant. We used SPSS (IBM Corporation, Armonk, NY, USA) and GraphPad PRISM (GraphPad Software, Boston, MA, USA) for all statistical analyses.
This study included 54 patients. The mean age was 73.7±12.9 years in patients with SVO (n = 27) and 75.1±11.9 in patients with BAD (n = 27). The number of men was 22 (81.5%) in patients with SVO and 17 (63.0%) in patients with BAD. The vascular risk factors and blood test findings were comparable between the two groups. Compared with patients with SVO, those with BAD exhibited significantly higher mean NIHSS score on admission (2.0 vs. 4.3, p＜0.001) and discharge (1.4 vs. 5.3, p = 0.009) and mean worst NIHSS score during hospitalization (2.1 vs. 6.4, p＜0.001); significantly more frequent early neurological deterioration [1 (3.7%) vs. 10 (37.0%), p = 0.002]; and significantly larger mean infarct volume on brain diffusion-weighted imaging (3.1 mL vs. 30.7 mL, p＜0.001) (Table 1).
| SVO | BAD | p value | |
|---|---|---|---|
| Number | 27 | 27 | ― |
| Age | 73.7±12.9 | 75.1±11.9 | 0.66 |
| Men | 22 (81.5) | 17 (63.0) | 0.13 |
| Smoking | 17 (63.0) | 15 (55.6) | 0.58 |
| Alcohol consumption | 13 (48.1) | 15 (55.6) | 0.59 |
| Hypertension | 17 (63.0) | 14 (51.9) | 0.41 |
| Diabetes mellitus | 6 (22.2) | 4 (14.8) | 0.48 |
| Dyslipidemia | 7 (25.9) | 5 (41.7) | 0.51 |
| Systolic blood pressure (mmHg) | 170.7±28.1 | 173.4±26.2 | 0.72 |
| Diastolic blood pressure (mmHg) | 96.5±14.7 | 98.7±16.7 | 0.61 |
| Hemoglobin A1c (%) | 6.2±1.3 | 6.2±1.2 | 0.98 |
| Triglyceride (mg/dL) | 148.1±106.7 | 153.6±103.1 | 0.85 |
| Low-density lipoprotein cholesterol (mg/dL) | 112.7±39.3 | 133.6±41.8 | 0.063 |
| High-density lipoprotein cholesterol (mg/dL) | 49.9±12.7 | 53.2±15.2 | 0.39 |
| Estimated glomerular filtration rate (mL/min/1.73 m2) | 63.5±17.6 | 65.4±20.7 | 0.72 |
| C-reactive protein (mg/dL) | 0.17±0.20 | 0.24±0.46 | 0.43 |
| NIHSS score on admission | 2.0±2.1 | 4.3±2.9 | ＜0.001 |
| NIHSS score at discharge | 1.4±2.1 | 5.3±5.5 | 0.009 |
| Worst NIHSS score | 2.1±2.2 | 6.4±5.0 | ＜0.001 |
| Early neurological deterioration | 1 (3.7) | 10 (37.0) | 0.002 |
| Infarct volume (mL) | 3.1±6.4 | 30.7±28.1 | ＜0.001 |
Categorical variables are presented as frequencies (%) and continuous variables are presented as mean±standard deviation Abbreviations: SVO, small-vessel occlusion; BAD, branch atheromatous disease; NIHSS, National Institutes of Health Stroke Scale
Absolute WSS values on ipsilesional M1 tended to be higher in patients with BAD than in those with SVO (whole M1, p = 0.13; upper wall of M1, p = 0.14; lower wall of M1, p = 0.14) (Supplementary Fig.2). To reduce the influence of interpatient variations in cerebrovascular geometry, we compared absolute WSS values between ipsilesional and contralesional M1. Among patients with BAD, WSS on M1 was approximately 2-fold higher on the ipsilesional side than on the contralesional side, whereas among patients with SVO, the ipsilesional and contralesional WSS on M1 were comparable (Supplementary Table 1, Supplementary Fig.3). WSS ratio of ipsilesional M1 to contralesional M1 was higher in patients with BAD than in those with SVO (whole M1, p = 0.052; upper wall of M1, p = 0.076; lower wall of M1, p = 0.034) (Supplementary Table 1, Supplementary Fig.3). The number of patients with higher ipsilesional WSS than contralesional WSS was significantly higher in patients with BAD than in those with SVO on M1 [20 (74.1%) vs. 11 (40.7%), p = 0.013]; upper wall of M1 [20 (74.1%) vs. 12 (44.4%), p = 0.027]; and lower wall of M1 [20 (74.1%) vs. 10 (37.0%), p = 0.006] (Fig.2 and Table 2). We investigated whether alteration of ipsilesional WSS on M1 was associated with BAD. The independent risk factor for BAD was larger ipsilesional WSS than contralesional WSS on M1 (aOR 4.38, 95% CI 1.29–14.82, p = 0.018), including the upper wall (aOR 3.80, 95% CI 1.11–12.95, p = 0.033) and lower wall (aOR 4.38, 95% CI 1.29–14.82, p = 0.018) (Fig.3).
The bar graphs show the WSS (Pascal [Pa]) on (A) ipsilesional M1, (B) ispilesional upper wall of M1, and (C) ispilesional lower wall of M1 in patients with SVO and BAD.
| SVO | BAD | p | |
|---|---|---|---|
| WSS ratio (= ipsilesional/contralesional) | |||
| - M1 | 1.11±0.91 | 2.13±2.53 | 0.052 |
| - upper wall of M1 | 1.14±0.87 | 2.44±3.63 | 0.076 |
| - lower wall of M1 | 1.08±0.91 | 1.93±1.81 | 0.034 |
| FV ratio (= ipsilesional/contralesional) | |||
| - internal carotid artery | 1.10±0.34 | 1.26±0.64 | 0.24 |
| - M1 | 1.21±0.82 | 1.76±1.72 | 0.18 |
Continuous variables are presented as mean±standard deviation
The bar graphs show the WSS ratio on (A) M1, (B) the upper wall of M1, and (C) the lower wall of M1 in patients with SVO and BAD.
In the patient with small-vessel occlusion (A), WSSs on bilateral M1 were similar. In the patient with branch atheromatous disease (B), WSS on left M1 was higher than that on right M1.
| SVO | BAD | p value | |
|---|---|---|---|
| WSS ratio (ipsilesional/contralesional) ＞1 | |||
| - M1 | 11 (40.7) | 20 (74.1) | 0.013 |
| - upper wall of M1 | 12 (44.4) | 20 (74.1) | 0.027 |
| - lower wall of M1 | 10 (37.0) | 20 (74.1) | 0.006 |
| FV ratio ＞1 | |||
| - internal carotid artery | 15 (55.6) | 16 (59.3) | 0.78 |
| - M1 | 13 (48.1) | 18 (66.7) | 0.17 |
Categorical variables are presented as frequencies (%) and continuous variables are presented as mean±standard deviation
Abbreviations: WSS, wall shear stress; FV, flow velocity; SVO, small-vessel occlusion; BAD, branch atheromatous disease
Model 1: univariate
Model 2: adjusted for age and sex
Model 3: adjusted for model 2, hypertension, diabetes mellitus, and dyslipidemia
Abbreviations: WSS, wall shear stress; BAD, branch atheromatous disease; CI, confidence interval
The number of patients with a contralesional WSS of ＜1.2 Pa on M1 tended to be larger in patients with BAD (7 of 27, 25.9%) than in those with SVO (2 of 27, 7.4%) (p = 0.068). In model 2 of the multivariate logistic regression analysis, a contralesional WSS of ＜1.2 Pa on M1 was associated with BAD (aOR 5.37, 95% CI 0.92–31.24, p = 0.062).
FV in the Ipsilesional Internal Carotid Artery and M1 was not Associated with BADTo investigate the association between alteration of blood FV into M1 and BAD, the FV ratios in the internal carotid artery (ICA) and M1 were compared between BAD and SVO. The two groups had a comparable number of patients with ipsilesional/contralesional FV ratio of ＞1 in the ICA [16 (59.3%) vs. 15 (55.6%), p = 0.78] and M1 [18 (66.7%) vs. 13 (48.1%), p = 0.17]. On univariate logistic regression analysis, the ipsilesional/contralesional FV ratio in the ICA (OR 1.16, 95% CI 0.40–3.43, p = 0.78) and M1 (OR 1.86, 95% CI 0.62–5.58, p = 0.27) were not associated with BAD (Fig.4).
Abbreviations: FV, flow velocity; BAD, branch atheromatous disease; CI, confidence interval; ICA, internal carotid artery.
To investigate the associations of WSS with neurological severity and infarct volume in each stroke subtype, we compared the worst NIHSS scores and infarct volumes between patients with ipsilesional/contralesional M1 WSS ratios of ≤ 1 and ＞1. Among patients with SVO, those with ipsilesional/contralesional M1 WSS ratios of ≤ 1 and ＞1 had similar worst NIHSS scores (2.3 vs. 1.9, respectively, p = 0.69) and infarct volumes (4.0 mL vs. 1.9 mL, respectively, p = 0.42) (Supplementary Figs.4A and 4B). Similarly, among patients with BAD, the two subgroups had comparable worst NIHSS scores (7.7 vs. 6.2, respectively, p = 0.49) and infarct volumes (23.1 mL vs. 33.4 mL, respectively, p = 0.42) (Supplementary Figs.4C and 4D). Moreover, the frequency of early neurological deterioration among patients with BAD was similar between those with ipsilesional/contralesional M1 WSS ratios of ≤ 1 and ＞1 [12 (70.6%) vs. 8 (80.0%), respectively, p = 0.59].
The bar graphs show the worst NIHSS score and infarct volume according to the WSS ratio on M1 among patients with SVO (A, B) and those with BAD (C, D).
This study revealed that higher WSS on ipsilesional M1 than on contralesional M1 was an independent risk factor for BAD. Considering that CFD is available at institutions where brain MRA or computed tomography angiography is performed, a relatively high WSS on ipsilesional M1 may be useful for the diagnosis of BAD immediately after emergency room arrival. Furthermore, a low contralesional WSS on M1 was associated with BAD. Based on these results, a relatively high WSS on ipsilesional M1 at onset after a period of low WSS on M1 is a potential pathophysiology of BAD.
At present, there is no therapy with established efficacy for BAD, because the pathological mechanisms remain unclear4). Nevertheless, inflammation plays a pivotal role in the onset and worsening of BAD, based on previous reports showing inflamed atherosclerotic vulnerable plaques on gadolinium-enhanced brain MRA19). Two previous autopsy case reports on BAD demonstrated that the microatheroma from the branch orifices comprised almost entirely large macrophages and that the branch vessel lumen was obstructed and filled with a mixture of large macrophages, a platelet mass, and red blood cells5, 20), suggesting the vulnerable plaques coexisting marked inflammation and plaque rupture.
In general, further progression of plaques to a vulnerable phase was associated with a high WSS on CFD8). High WSS structurally represents thin21-23) and eroded22, 24, 25) plaque caps, plaque ulceration26), and plaque rupture25-28). Moreover, high WSS was reported to increase macrophage density in human atherosclerotic plaques29, 30). In in vitro studies, acutely high WSS was associated with the activation of proinflammatory factors in endothelial cells, mediated by the mitogen-activated protein kinase and nuclear factor-κB signaling pathways31) and with the early activation of extracellular signal-regulated kinase 1/2, p38 mitogen-activated protein kinase, and Akt in human umbilical vein endothelial cells32). Furthermore, a high WSS enhances platelet activation and subsequent adhesion and aggregation24, 25, 33). Therefore, a high WSS in perforating artery occlusion can lead to BAD through the formation of vulnerable plaques and local platelet activation and aggregation.
On the other hand, low WSS generally leads to plaque initiation and greater plaque progression8, 27). In this study, low WSS on contralesional M1 was associated with BAD. Presumably, patients with BAD initially have a low WSS on diffuse intracranial arteries, which favors the formation of atherosclerotic plaques on M1 before symptom onset. However, vulnerable plaques responsible for BAD may require a high WSS. Small atherosclerotic plaques that develop under low WSS conditions may have slight stenotic changes, which are invisible on conventional MRA but can increase WSS, based on Poiseille’s law, which implies that WSS is inversely proportional to the cube of the vessel diameter. Elevated WSS can exacerbate plaque vulnerability, even if the stenosis remains undetectable by standard imaging techniques, such as MRA.
We used the ipsilesional/contralesional WSS ratio on M1 rather than absolute WSS values to account for the effects of intersubject variations in intracranial arterial geometry on WSS values. In previous studies using CFD to investigate cerebrovascular WSS, WSS is commonly normalized to minimize the effects of varying vessel geometry and size among the patients. Normalized WSS is defined as the ratio of cerebral aneurysm WSS to parent artery WSS. WSS distributions are usually normalized to the average parent vessel WSS in the same patient to allow comparison of WSS among different patients34-36). Therefore, to minimize intersubject variations, we defined normalized WSS as the ratio of ipsilesional to contralesional M1 WSS because the branching patterns of the left and right M1 are known to be symmetrical in 91% of patients37).
This study had several limitations. First, its retrospective design comes with a risk of selection bias. Second, the sample size was relatively small. Third, vascular wall imaging (high-resolution MRI) to evaluate plaque phenotypes, including plaque rupture and inflamed plaques, on the M1 was not performed. In future studies, a combination of CFD techniques and high-resolution vascular wall imaging may reveal the effects of WSS on plaque components and vulnerability. Nevertheless, this study proposed that BAD was associated with the characteristics of hemodynamic forces on M1 despite of invisible plaques on 3 Tesla brain MRA, which can be verified in a future prospective study with larger sample sizes.
The higher ipsilesional than contralesional WSS on M1 demonstrated by CFD suggested that the pathological mechanisms in BAD could be a high WSS causing conversion into vulnerable plaques. Moreover, the use of CFD may contribute to the diagnosis of BAD.
The authors would like to thank Enago (www.enago.jp) for the English language review.
This study was supported by Takeda Science Foundation (YH). The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
The authors declare that they have no known competing financial interests or personal relationships that could have influenced the work reported in this study.
