Anatomical study of variations in the configurations of the circle of Willis in relation to age, sex, and diameters of the components

Abstract

The circle of Willis (COW) refers to the anastomotic arterial network found on the brain base, tasked with provision of collateral circulation aimed at prevention of ischemia. The COW is of immense clinical importance especially with regard to the assessment of neurovascular diseases. Individuals portray significant variations in the COW’s anatomical configuration. The present study seeks to evaluate the existing anatomical variations of the COW and within the anterior and posterior segments of the COW. Thus, the study seeks to evaluate the different anatomical variations of the COW and its segments and components within the study population. To attain the set objectives, the present study has utilized the angiographic images for studying the COW variants in patients who underwent cerebral angiography during assessment of different types of cerebral anomalies and conditions. Therefore, this study used conventional angiography as an important tool in the evaluation of the different variations in the COW, and is most appropriate for evaluation of smaller anatomical variations owing to its perfect spatial resolution and portrayal of COW anatomy. The study findings indicated the existence between age and sex, and anatomical variations of the COW, particularly with regard to diameters of COW components like basilar artery (BA), P1, and internal carotid arterys (ICAs). Males had bigger BA, P1 and ICA diameters than females, while individuals aged below 40 years had bigger BA, A1, posterior communicating artery, and ICA diameters than those aged above 40 years.

Introduction

The circle of Willis (COW) refers to the anastomotic arterial network found on the brain base. The COW’s key role entails the provision of adequate collateral circulation to the cerebral and cerebellar tissues to prevent ischemia and subsequent transient ischemic attack and stroke [1]. Initially described in 1664 by Thomas Willis in his landmark publication “Cerebre Anatome”, the COW has been characteristically defined as being a symmetrical polygon, drawn from the anastomoses found between the branches of the vertebral and internal carotid arteries. Contemporary anatomy books and publications have referred to the COW as an unevenly pentagonal circle comprising vessels on the brain’s ventral surface [2]. COW comprises the posterior and anterior cerebral arteries, offering arterial supplies to different cerebellum and cerebrum lobes. The two posterior communicating arteries (PComs) and the anterior communicating artery (ACom) connect with the cerebral arteries and assist in the formation of the collateral arterial network [2]. The PCom refers to a part of a group of arteries found in the brain’s COW, which connects to the posterior cerebral arteries and the internal carotid, and its function entails the blood supply to the brain.

Since the initial definition by Willis, there has been increased evaluation of the COW in cadaveric studies and live patient imaging analyses that include magnetic resonance angiography (MRA) and computer tomography (CT). Despite the classical description of the COW as symmetrical with bilaterally almost equal diameter vessels, the COW is subject to considerable anatomical variations. However, it is unfortunate that the large number of anatomical studies focusing on COW has been unhelpful in clarifying the anatomical variations and their prevalence, mainly due to the inconsistency between the results of the studies. For instance, existent studies have reported that classical COW is present in between 4.8% and 85.4% of the population [3]. Such a more comprehensive range of variations in reported rates has been attributed to divergence in methodology and nomenclature between the researches [1], population and ethnic discrepancies [4], and if the studies focused on neurologically diseased or healthy populations. Nevertheless, there is a general agreement in the existing literature with regard to the observation that the variations are more common in PCom [5, 6]. Moreover, several variations have been reported and classified based on numerous factors. For instance, different types of variations of the COW have been observed and documented. This includes type A, considered as the normal anatomical morphology of the COW in an adult individual. Type B variation is mainly found in COWs with over one AComs, while type C variation entails the callosomarginal artery that arises from the ACom. Consequently, type D variation of the COW entails the fusion of the anterior cerebral artery (ACA), which refers to the artery that arises originates in the internal carotid at the lateral cerebral fissure’s medial extremity and whose function entails supplying the corpus callosum, the parietal, frontal and cingulate cortex, over a shorter distance, even as type E variation has been noted to entails the ACAs’ A2 segments that occur distally after the common trunk, and type F involved the middle cerebral artery (MCA) that originates from the internal carotid artery (ICA) as two distinctive trunks. Type G variation of the COW involved the absence of ACom or hypoplasia. Similarly, type H entails one of the ACAs having absent or hypoplastic A1 segments, and a contralateral anterior cerebral artery, which gives rise to the A2 segments. Consequently, type I variation of COW entails the absence or hypoplasia of the ICA. Thus, in type I variation, the ACA mainly gives rise to the A2 segments while also supplying retrograde blood flow to an ipsilateral A1 segment. Lastly, type J variation involves the absence or hypoplasia of ACom, even as the MCA originates as two distinctive trunks, including ICA and ACom.

Materials and Methods

For this study, the researchers interpreted and re-evaluated 85 cerebral MRA images of patients who underwent magnetic resonance (MR) assessment. The patients had undergone the cerebral imaging procedure between May 2022 and June 2023 for various reasons, including severe migraines/headaches, decrement in motor strength, suspected stroke, decrement and absence in sensations, deficient memories, neurological examination findings and referrals, dizziness, and suspected transient ischemic attacks, among others. The patients were examined using the 1.5-Tesla MR scanner (Siemens Healthcare).

The researchers further scanned the medical records of the patients included in the study to avert any potential effect related to vascular diseases found on the vessel calibres of the COW. As a result, 7 patients were excluded from the study, including 2 patients who were found to have left ICA occlusion, 1 patient who had significant (above 50%) right side ICA stenosis, 1 with ACom aneurism, 1 with subdural effusion, 1 with considerable cerebral hematoma, and 1 patient with subdural effusion. Consequently, the researchers excluded 8 MRA images that had artefacts that were capable of hindering effective and conclusive interpretation, as a result of motion due to instability during assessment or in instances where the patient presented dyskinesia, alongside other imaging artefacts resulting from clips of ferromagnetic intracerebral aneurism. Following the exclusion, the researchers included 12 MRA images drawn from the hospital database, each from different patients, without any cerebrovascular diseases. Further, in addition to using the 16-channel standard head coil on the patients, various technical parameters were employed in handling the 3D-TOF (3D-time of flight) images, including echo time: 7.16 ms, slice thickness of 0.58 mm, matrix of 0.4 × 0.4 × 0.6 mm³, repetition time of 25 ms, flip angle: 25°, field of view: 243 × 256 mm, and the imaging time was approximately 7 minutes and 18 seconds.

Furthermore, to ensure accuracy in the identification of the variations of the COW, the researchers employed the services of a radiologist with over 15 years of experience alongside a resident radiologist with over 5 years of experience to assess the selected MRA images jointly with regard to the existence of every vascular component of the COW, alongside every vessel’s diameter measurements, as well as the categorization of the variants of the COW, based on the system of classification developed by Chen et al. [7]. The various decisions on the evaluations and measurements were arrived at through consensus between the radiologists, even as a magnification of ×400 was employed to avert potential measurement errors and to discover the precise vessel measurements. Further, to evaluate the various cross-sectional images presented in the standard Digital Imaging and Communications in Medicine (DICOM) formats along with the 3D-TOF images acquired using the maximum intensity project algorithms, the researchers used Akgün PACS viewer v7.5 software developed by Akgün software for photo archiving and communications. An adjusted version of the measurement methodology developed by Shatri et al. [8] was then deployed to establish the calibre of the vessels in the COW, even as an axial plane was also employed in the measurement of the COW’s segments vascular diameters, at 3.5 mm from the origin point of the vessel, vertical to the artery’s elongation from the interior walls. Whenever the technique could not be employed due to the vascular segment being too short, a measurement of the artery’s middle part was conducted. While vascular segments that visualized 0.8 mm diameters and above were considered as present, vessel segments below 0.8 mm diameters were considered hypoplastic [7, 9]. Further, arteries lacking segment visualizations and those had non-continuous segments were marked as absent.

The COW was evaluated distinctly as posterior and anterior segments, and the results were noted. Further, the COW’s integrity was categorized into three main groups, namely: the complete COW, which did not have any variants and had all COW components present in the participants; the partially complete COW, which included participants with anatomical variations in either the posterior or the anterior segment, and incomplete COW, which includes a variation in both the posterior and anterior segments of the COW. Every anatomical variation’s prevalence was also reported, and the participant/patient populace was presented with sex difference, age distribution, and COW configuration. Moreover, every anatomical COW variant classification was indicated using percentages, even as the findings with regard to the vessel diameters of every COW component were mainly reported using mean and standard deviations. The measurement and presentation of the diameters of the different vessels were done based on sex and age differences. Thus, regarding the sex and age differences, the researchers conducted an independent sample t-test to calculate the significance of the vessel diameter differences. ANOVA analysis was effectively utilized to establish the study findings’ significance, particularly the integrity between the vessel diameters and the COW. Before evaluation of the age and sex differences in relation to the vessel calibres, the Shapiro-Wilk test was used to control data distribution.

No normal data distribution that would enable the comparison of the age and sex differences and for the comparison and the comparison of divergent COW configurations was found for both P1 (the segment found between the basilar artery [BA] top and the PCom artery origin) and A1 (the segment that is horizontal and with a proximal portion, which extends from the ACA origin to its connection to the contralateral ACA through the ACom way), as well as for BA and ACom. As such, to effectively determine the age and sex differences, the researchers utilized Mann Whitney U-test. Further, to measure the diameters of the vessels based on the COW configurations, the researchers used the Kruskall-Wallis test to identify any potential significant variations for the vascular segments. Also, for every COW vascular segment, the researchers have analysed the existing correlations between vessel diameter and age using Spearman’s Rho test. Lastly, a chi-square test was employed to analyse of the distribution of the incomplete, partly complete, and complete COWs in the study population in relation to age and sex differences. Throughout the study, P-values <0.05 indicated statistically significant variation/difference.

Ethical considerations

The study has followed ethical guidelines for primary data analysis, ensuring no individual could be identified. The Institutional Review Board approvals and informed consent was obtained from the institutional ethics committee as this study is retrospective study (IRB No. SP23/417/J).

Results

The researchers included MRA images of 85 patients in the study. The study populace’s mean age was 48 years, even as there was a higher female predominance, with patients 20 years and below constituting a comparatively smaller proportion of the study population. Moreover, more than half (59%) of the studied patient population presented at least one posterior or anterior variation of the COW. This has been aptly captured in Table 1 of Appendix 1.

Based on the analysis of the images, the type A variant was found to be the most dominant variation of COW with regard to anterior circulation, even as type G and type H were observed to be the second and third most dominant variants of the COW in the study population. The researchers also noted a predominance of type A variation in the posterior segment of COW; nonetheless, a higher proportion of the study participants (n=40) (47%) had variants mainly found in the posterior segments of the COW, compared to the anterior segments. Regarding the variants of the COW in the posterior segment, type D and E variants were found to be more widespread than the type A variants as indicated in Table 2 of Appendix 1 below. Nevertheless, it is worth noting that types C, H, and J, as well as types F and J, were not found in the posterior and anterior segments of any study participants. The prevalence of the anterior and posterior variants has been aptly captured in Table 2 of Appendix 1.

Furthermore, different COW variants were recorded throughout the present study, even though it was impossible to classify them based on the extant classification system employed in this study. These have been aptly captured in Table 3 of Appendix 1 below. For instance, 6 participants (7.06%) presented unclassified variations in the study population. Moreover, the evaluation of the COW MRA images indicated that no study participant had hypoglossal artery and persistent trigeminal artery, and no participants had carotid-vertebrobasilar anastomoses. The other types of variants found in the participants are indicated in Table 3 of Appendix 1.

Additionally, the researchers have noted an increased number of smaller calibrations for BA, P1s, and ICAs in female participants compared to their male counterparts. In study participants aged 40 years and above, the diameter of the vessels showed that the left PCom and the ACom diameters were significantly smaller, even as they presented higher ICA and BA calibres compared to the younger study participants, as indicated in Table 4 of Appendix 1.

The researchers also found a negative and weaker correlation between age and vessel calibres for the COW’s left PCom, A1, and ACom segments. However, they also observed a positive but weaker correlation between BA and left ICA calibrations in relation to the age of the participants. Still, the researchers did not observe any correlations between the COW’s right PCom, ICA, and P1 segments. The correlation between vessel diameter and age (P-value and r-correlation coefficient) has been captured in Table 5 of Appendix 1.

The researchers observed that while the P1 and BA segments’ diameters were lower, the diameter of the left PCom was considerably higher, particularly in patients who presented complete COWs. Additionally, it was observed that study participants who presented incomplete COWs had lower ACom values, even as considerably lower calibres were observed in the right side P1 segment, particularly in participants who presented partly complete COWs in comparison to other COW configurations. The vessel diameters according to the COW configuration have been captured in Table 6 of Appendix 1.

Lastly, the researchers have additionally observed considerable differences in age and sex in relation to the distribution of incomplete, partly complete, and complete COW groups. For instance, females presented a higher proportion of complete circles, and a higher proportion of complete COW was also noted in study participants who were below 40 years of age. The sex distribution according to the configuration of the COW has been captured in Table 7 of Appendix 1 (Figs. 1–7 of Appendix 2).

Discussion

The morphological configuration of the COW has been studied by many anatomical and clinical studies in the past. Studies on larger populations undoubtedly will be of interest in the literature since the larger the sample size, the more accurate and valuable information will be obtained. As one of the most extensive series in the literature, this study aimed to be descriptive and informative about the anatomical variances that exist in the COW. A total of six branchial arteries pairs play important roles in the formation of COW. The ICAs are developed during the initial stages of the embryonic period through the third branchial artery contributions. Moreover, common carotid arteries are formed through the fusion of the ICAs and the ventral pharyngeal arteries. At the age of 28 days, during development, the division of the ICA branches into the posterior and anterior segments occurs. Further, the anterior ICA forms the anterior choroidal artery, the MCA, and the ACA. Further, the hindbrain’s vascular supply is provided by the carotid-vertebrobasilar connections, which include proatlantal arteries, trigeminal artery, hypoglossal artery, and the otic artery. Further, commencing with the development of PCom arteries, which joins with the BA’s distal segment, the hypogossal artery, otic artery, and trigeminal artery tend to regress, even as the proatlantal artery does not develop until such a time that the development of vertebral arteries has occurred. The development of the MCA commences at about 35 days, and this occurs from ICA’s anterior divisions. During this time, the growth of ACA occur medially and concurrently with the development of ACom. Normally, both the ACom and ACA can be seen in the COW at between 6 and 7 weeks, following commencement of development [10].

As initially noted in earlier literature, to be considered complete, a COW must have these components/vessels; two A1 segments, two ICAs, one ACom, one BA, two P1 segments, and two PComs. With regard to the other anatomical variants of the anterior segment of the COW, several previous studies have reported increased dominance of type A variation, and similar findings have been made in the present study. The complete anterior COW segment had a prevalence that ranged between 75% and 91%, even as this prevalence was found to have been much lower when contrasted against the autopsy series [11].

The present study findings indicate that 76% of the participants (n=65) had complete COWs, comparable to the findings of similar earlier studies. Three key categories of the Posterior COW segment have been acknowledged in the literature, namely, fetal type, transitional type, and adult type. In the fetal/embryonic type, the P1 segment diameter is smaller compared to the P2 and PCom diameters [12], while in the transition type the diameters of both PCom and P1 are both equal and equally contribute to the development of posterior cortical atrophy of the P2 segment. Further, in the adult type, the P1 segment diameter is bigger compared to the diameter of the PCom [12]. These categories have played significant roles in the determination of variations and variants found in the posterior part of the COW. A number of earlier studies reported an increased prevalence of type E, particularly in instances where the absence of PComs was noted in relation to the posterior segment of the COW [13-15]. The findings of the present study have additionally indicated similarities to those of earlier studies with regard to the posterior COW segment, even as type A was found to be the most predominant variant and type E the second most predominant variant [16].

In various previous studies, a number of COW variants were not reported/observed. For instance, in the study conducted by Keeranghat et al. [16], type I and type F variants of the anterior COW segments were not observed. Further, no participant with a type I variant of the posterior COW segment was reported in the study conducted by Maaly and Ismail [14]. Still, an additional study conducted by Naveen et al. [11] reported the absence of type I in both the posterior and anterior COW segments, as well as the absence of type J variant in the anterior segment of the COW. An additional study conducted by Ravikanth and Philip [13] also disclosed the absence of any participant with a type I variant in both the posterior and anterior COW segments, as well as the absence of any participants with type F and type J variants in the anterior COW segment. Type J, I, and F variants were not reported in relation to the anterior COW segment in the study conducted by Shaikh and Sohail [15]. Nevertheless, in the present study, types J and F variants of the anterior COW segments were absent in the study population, even as types J, H, and C variants of the posterior COW segment were absent.

Similar to the findings of the study conducted by Keeranghat et al. [16], the findings of the present study have disclosed that participants with complete COW formed the highest proportion while those with incomplete COWs formed the lowest proportion. However, in a number of earlier studies, it was established that patients with complete COWs formed the highest percentage of the COW configuration [13, 14]. On the contrary to the findings of the study conducted by Keeranghat et al. [16], the findings of the present study have disclosed that there were no significant differences between the different age groups in relation to the distribution of the divergent configuration types; nonetheless, a significant difference was observed between the configuration types in males and females. In their study, Maaly and Ismail [14] disclosed an increasingly higher proportion of wholly complete COWs in males and in participants aged 40 years and above, dissimilar to the findings of the present study. Additionally, the present study has disclosed that, in male participants, the BA and PI segments, along with the ICAs had bigger diameters compared to those of the female participants. The study conducted by Keeranghat et al. [16] disclosed that, in males, the left ICA and left PCom had bigger diameters. Similarly, a study by Maaly and Ismail [14] found BA, ACom, A1, and ICA to have significantly bigger diameters in males compared to females, even as BA, PCom, A1, and ICA were found to have significantly bigger diameters in participants aged below 40 years. Still, the study has disclosed that, in individuals aged below 40 years, the BA, A1, PCom, and ICA had considerably bigger diameters [14]. Further, the study conducted on participants aged above 40 years disclosed that the right PCom diameter measurements had bigger values while the right P1 diameter measurement had lower values [16]. With regard to the present study’s participants, the left PCom and the ACom’s diameters were found to be considerably smaller, even as the left ICA and BA calibrations were considerably bigger in individuals aged above 40 years, in comparison to those aged below 40. Thus, old age has been linked to an increment in the vessel wall thickness and the subsequent decrement in the lumen diameter. As such, we anticipated that a negative correlation existed between the vessel caliber and the age. Nevertheless, the correlation analysis conducted by the researchers has shown the existence of a negative relationship only for the left PCom, ACom, and A1 segments of the COW. In the present study, both P1 and BA segments of the COW were observed to have considerably smaller diameters, even as the left PCom had significantly bigger diameters in participants who had wholly complete COW. Consequently, the study findings have indicated that in participants with incomplete COWs in comparison to those with partial and complete COWs, the ACom calibrations were significantly smaller. The P1 segment measurement in the group of participants with COWs that were partially complete disclosed lower/smaller calibres.

Additionally, the present study also disclosed the existence of considerable age and sex differences in relation to the COW configuration distributions. For instance, the researchers noted that the female participants had higher proportions of complete COWs in comparison to their male counterparts. Further, the researchers observed that the study participants aged 40 years and below had higher proportions of complete COWs in comparison to individuals aged above 40 years. In earlier studies, with regard to the COW configurations, significant differences were observed in both PComs, right P1, both A1s, and right ICA segment diameters [10, 14, 17]. Comparable to the findings of the present study, the findings of the research conducted by Keeranghat et al. [16] disclosed the existence of considerable age and sex differences in relation to the diverse COW configuration distributions; nonetheless, in the present study, the group of participants with partially complete COWs was noted to comprise the biggest proportion in all groups. Consequently, a number of researchers have acknowledged the existence of correlations between age and variations in the COW, particularly with regard to completeness and incompleteness of the COW. For instance, a recent study conducted by Zaninovich et al. [18] has supported the emergent hypothesis on sex and age having considerable impacts with regard to the rates of complete COWs, in addition to impacting the states of the disease. Still, regarding the existence of a correlation between age and COW variations, it has been noted that MCA aneurysms are increasingly infrequent in individuals aged >55 years, even as women are highly prone to present incomplete COW at later years in life [19]. Consequently, it has been noted that individuals aged <40, 40–69, and ≥70 years are highly likely to present complete COWs, at 67.93%, 31.78%, and 18.76%, correspondingly [20]. Further, a recent univariate analysis disclosed that increasing age was correlated with total hypoplastic and aplastic vessels [20]. Still, a multivariate analysis conducted by Eaton et al. [20] disclosed that age was a primary mechanistic driver that affected the anatomy of COW vessels, as opposed to the various confounding comorbidities.

Regardless of the above findings, several limitations were noted in this study, and given the need for cautious interpretation of the descriptive analysis and the data, it is vital that the observed limitations are stated alongside the limiting factors. Among the notable limitations is the observation that the 3D-TOF sequencing is not accurate technique for imaging with regard to the analysis and measurement of various vascular calibrations. Despite the 3D-TOF technique being extensively employed and increasingly successful with regard to imaging the intracerebral circulation, it faces challenges when used to image minute and smaller vascular collateral channels, given the slow velocity and turbulent flow of blood adjacent to the walls as a result of laminar flow [8]. To avert potential measurement-related errors and ensure precise vessel diameters, the researchers used a ×400 magnification for the different COW components; nonetheless, despite the use of increasingly elaborate and careful techniques, it is important to consider the human error factor as a key limitation in the present study. However, a number of studies have considered vessel calibres of 1 mm as determinants of vascular segments’ hypoplasia [16, 21], even as other studies have considered 0.8 mm vessel calibres as indicators of hypoplasia [14], similar to the present study. This, therefore, necessitates the cautious comparison of the findings of the present study with those of earlier studies, particularly in relation to the variations of the COW. Still, other researchers have found indifferent outcomes from the comparable topics they studied on the subject of variations of the COW in literature. This may mainly be due to genetic differences existing between diverse study populations. Therefore, it is important to categorize the outcomes of such studies, especially in relation to the definite demographic features of every study. For the present study, the researchers did not express the potential effects of the different vascular conditions in relation to the statistical findings, even as patients who did not have a positive cerebrovascular disease history or cerebrovascular lesions were included in the study. The ICA status has been acknowledged to impact the COW vessel calibres [22-24]. In instances where the study participants presented known ICA status, the measurements are highly prone to reflect accurate results. Thus, for this study, numerous statistical analyses were conducted using non-parametric tests owing to the observed properties of data distribution. This necessitates the careful evaluation of the findings of this study. Furthermore, regardless of the significant size of the study population, it is important that an increased number of patients are retrospectively re-interpreted to evaluate and compute increasingly precise values that will enable improved comprehension of the COW variations’ incidences.

In conclusion, the study has disclosed a reduced proportion of complete COW in males compared to females. The observed statistically significant difference between the male and female participants with regard to the distribution of incomplete, complete, and partially complete COW can be attributed to the existence of an increased number of COW variations in males. Moreover, in the entire study population, the observed variations in the vessel diameters of incomplete, partly complete, and complete COWs are prone to impact the blood flow dynamics within the COW. Particularly, the variations in the vessel calibrations shown in this study are comparable to the findings of other studies and meta-analyses with bigger study populations, and this is prone to help in guiding better comprehension of the anatomical variations in the COW, as well as studying the variances between complete, partially complete, and non-complete COWs.

Funding Statement

Funding None.