Anatomical Congurations of Dominant Anastomotic Veins of the Supercial Cortical Venous System

Understanding the anatomy of the anastomotic veins (AV) of the supercial cortical venous system (SCVS), viz. supercial Sylvian vein (SSV) - also known as the supercial middle cerebral vein; vein of Labbe (VL) and vein of Trolard (VT), are imperative for neurosurgical procedures. This study aimed to investigate variant anatomical patterns of dominance of the AV, to elucidate the haemodynamically balanced SCVS, by reporting variations between the presence, diameter and dominant patterns of the AV. Two hundred lateral angiograms were included, depicting left and right cerebral hemispheres of the same patient (n = 100 patients). Angiograms were analysed and variations recorded. Results were statistically compared against laterality, age, sex and ethnicity. Presence of the VL had the highest occurrence (96.5%), whereas the SSV and VT had an occurrence of 75.5% and 64.5%, respectively. This study reports presence of double veins of the AV: SSV (12.0%), VL (22.0%) and VT (19.5%). Furthermore, presence of a triple vein for each AV is reported. Diameters for the SSV, VL and VT were 1.99 ± 0.500mm, 2.18 ± 0.579mm and 2.14 ± 0.472mm, respectively. Statistically signicant relationships were established between diameters and the SSV, VL, VT and VT2 (double VT). Seven types of dominant patterns were recorded: Equilibrium; singular dominance of SSV, VL and VT; co-dominance of SSV/VL, SSV/VT and VL/VT. The Equilibrium dominant pattern of drainage had the highest occurrence (54.5%). Patterns of dominance of these AV can aid the neurosurgeon in curbing the risk of iatrogenic injury and postoperative infarcts even after an otherwise successful surgery. . Knowledge of the variations of the AV of the SCVS would then elucidate the system 9 . Therefore, this study aimed to investigate the anatomical patterns of dominance of the AV of the SCVS. The objectives were to document the presence of the AV, determine the diameter, as well as record patterns of dominance and equidominance, and compare these factors on the basis of laterality as well as age, sex and ethnicity. of these measurements was then used for statistical analysis. Dominance of the AV was ascertained by the territory of which the AV drained. A dominant AV was determined if one of the AV, viz. SSV, VL or VT, was visualised as being the main venous drainage of the territory of another AV. A haemodynamically balanced SCVS with an equilibrium dominance pattern was determined if all AV presents drain their respective territories. These dominant patterns were then recorded.


Introduction
The super cial cortical venous system (SCVS) is a complex system consisting of dural venous sinuses and cortical veins 1 . Cortical veins are large subarachnoid venous collectors that receive venous drainage from a network of subpial veins, which lie beneath the arterial circulation of the cerebral cortex, and drains the super cial surface of the cerebrum 2,3,4,5 . These cortical veins are divided into three collecting groups and are linked via anastomotic veins (AV). These AV are large venous channels present in each of the collecting groups, and drain into their respective dural venous sinuses, viz: (i) the super cial middle cerebral vein of the anterior group, which drains into the cavernous or sphenoparietal sinus, (ii) inferior AV of the posteroinferior group, which drains into the transverse sinus and (iii) superior AV of the mediodorsal group, which drains into the superior sagittal sinus 1,4,6 . The AV, which provide an alternate pathway of drainage, are commonly referred to by eponyms; therefore, the super cial middle cerebral vein is termed as the super cial Sylvian vein (SSV), inferior AV as the vein of Labbe (VL), and superior AV as the vein of Trolard (VT) 4,6,7,8 .
The SSV originates in the posterior aspect of the lateral ssure, drains the region of the opercula and courses anteroinferiorly to drain into the cavernous sinus or sphenoparietal sinus 4,9,10 . The posterior end of the SSV anastomoses with the VL inferiorly and VT superiorly 7,11 . The largest anastomotic channel connecting the SSV and transverse sinus is the VL 11,12 . The VL originates in the posterior aspect of the lateral ssure and courses obliquely over the occipitotemporal sulcus, as well as the inferolateral aspect of the temporal lobe, to project posteroinferiorly into the transverse sinus 9,12 . The VT is the largest vein that crosses the lateral surface of the frontal or parietal lobes between the SSV and superior sagittal sinus 5,8 . VT has the most variable course, and while reported to be located in the area of the central sulcus, this AV may be present as anteriorly as the anterior frontal vein or as posteriorly as the anterior parietal vein 4,8,9 . Due to the presence of these AV, the SCVS is haemodynamically balanced since the SSV anastomoses with the VL inferiorly and VT superiorly 4,7 .
Regarding presence of the AV, the literature reports the presence of double AV of the SSV, VL and VT, respectively 4,5,8,13,14,15 . While the SCVS is known to be a haemodynamically balanced system, in which drainage of the cortical hemispheres into the venous sinuses was noted as being equally distributed, the anatomical structure of the veins of the SCVS may differ 7,8 . The AV may range from interconnected or separated AV, to a single dominant AV while the other two AV are hypoplastic 4 . Furthermore, Tanriverdi et al. reported variations of combinations of AV as the dominant drainage pattern for the SCVS, in which two of the three AV are co-dominant 8 . Therefore, although the SSV, VL and VT have been noted to have equal sizes, it is not uncommon for one or two of these AV to predominate and the others to be hypoplastic 4,14 .
The variant patterns of anastomosis of these AV have important surgical implications 16 . In well-developed anastomoses of the AV, particular cortical veins could be safely sacri ced; however, in cases of poor collaterals, preservation of the AV is essential 7 . AV are regarded as 'dangerous veins' since the risk of iatrogenic injury would result in severe complications such as oedema, swelling, raised intracranial pressure and haemorrhagic infarcts 16,17 . Neglecting to recognise an SCVS in which a singular AV is dominant may result in a venous infarct, even after an otherwise successful surgical procedure, due to the drainage route being compromised 4,5,8 . Since the AV are generally encountered during pterional craniotomies, which is a common neurosurgical approach, the risk of iatrogenic injury to the AV must be reduced; therefore, the anatomy regarding the AV in terms of their presence and dominance becomes imperative for neurosurgical perspectives 4,8,12,18 .
However, while cortical veins contain 70% of the cerebral blood volume, few studies have focused on the venous drainage of the cerebral cortex in comparison to the arterial supply 9 . Knowledge of the variations of the AV of the SCVS would then elucidate the system 9 . Therefore, this study aimed to investigate the anatomical patterns of dominance of the AV of the SCVS. The objectives were to document the presence of the AV, determine the diameter, as well as record patterns of dominance and equidominance, and compare these factors on the basis of laterality as well as age, sex and ethnicity. Data processing Angiograms were observed and analysed by use of the RadiAnt DICOM Viewer software. Angiograms with pathology affecting the venous anatomy, e.g arteriovenous malformations (AVMs) and dural arteriovenous stulae (DAVF), were excluded. Only angiograms depicting both cerebral hemispheres of the same patient were included.

Data collection
In the event of a double or multiple occurrences of an AV, the respective veins were numbered sequentially in an anterior to posterior (VL and VT) or superior to inferior (SSV) direction, e.g. SSV, SSV2, SSV3, of which SSV is the most anterosuperior of the three veins and SSV3 the most posteroinferior.
Diameter of the vein was documented at the widest observable point of the vessel, as well as 5.00mm proximal and distal to the said point. All measurements were taken three times to ensure reliability, interclass correlation coe cients were calculated to determine the reliability of the measurements (con dence intervals: 95%). An average of these measurements was then used for statistical analysis.
Dominance of the AV was ascertained by the territory of which the AV drained. A dominant AV was determined if one of the AV, viz. SSV, VL or VT, was visualised as being the main venous drainage of the territory of another AV. A haemodynamically balanced SCVS with an equilibrium dominance pattern was determined if all AV presents drain their respective territories. These dominant patterns were then recorded.

Statistical analysis
Statistical analysis was performed by using R statistical computing software, version 3.63 (R Studio, Boston, MA, USA). A statistically signi cant value of p < 0.05 was set. The Chi-square and Fisher's exact tests were employed for association between categorical variables. The t-test and Rank-sum test were employed to determine the (± standard deviation) and median (inter-quartile) differences of numerical variables between categorical variables for normally and non-normally distributed data, respectively, when comparing two groups. Whereas the Kruskal-Wallis test and ANOVA F-test was used to determine mean (± standard deviation) and median (inter-quartile) differences of numerical variables between categorical variables for normally and non-normally distributed data, respectively, for comparing more than two groups. A Cramers V correlation was performed to determine a relationship between the dominant patterns and laterality.

Results
The following results were obtained after analysis of the lateral angiograms included in the retrospective chart review.
Results of an intra-rater reliability test infer that the average measurements used are highly reliable (p < 0.001), interclass coe cients for all intra-rater reliability tests were (ICC = 1). The inter-rater reliability tests yielded the following interclass coe cients (con dence interval:   (Fig.1).
Furthermore, we report ndings of triple presence of each AV. A triple SSV was present in the right hemisphere of a 51-year-old Indian female. The triple VL was also recorded in the left hemisphere of a 16-year-old African female and 45-year-old White female. Triple VT occurred in the right hemisphere of a 25year-old African male (Fig.1). Table 3 Mean diameters of the anastomotic veins The mean diameter recorded for the SSV was 1.99±0.500mm. The SSV2 had a mean diameter of 1.92±0.477mm. A mean diameter of 2.18±0.579mm was recorded for the VL, whereas VL2 had a mean diameter of 1.96±0.388mm. The recorded mean diameter for VT (2.14±0.472mm) was smaller than that of the more posterior VT2 (2.19±0.604mm). The diameters of SSV3, VL3 and VT3 were 2.68mm, 1.52±0.0424mm and 1.63mm, respectively (  Table 3). Table 4 Dominant patterns of the anastomotic veins  We report seven types of dominant patterns of the AV of the SCVS (Table 4, Fig.2). The pattern with the highest incidence is the Equilibrium pattern (54.5%), followed by the VL/VT pattern (28.0%).  Fig.2).  (Table 5).

Discussion
The anatomy of the veins of the SCVS is of interest to clinicians. Special focus lies in the anatomy of the AV, since these veins are regarded as 'dangerous veins' in the neurosurgical eld, due to the high risk of postoperative infarction after an otherwise successful surgery 7 . However, there is a signi cant paucity of literature focusing on the AV, especially in the context of the SCVS entirely. This study aimed to ll the gap in the literature regarding the anatomy and anatomical variations of the AV to assist clinicians in preoperative planning. Therefore, familiarity with the variant anatomy of these AV will elucidate the SCVS since precise information regarding these veins would be able to form a network of reference points that are relevant during surgical planning and surgical procedures 9 Table 2). The high occurrence of the VL must be noted by clinicians since a sacri ce of this vein in a poorly anastomosed SCVS would lead to a swollen temporal lobe and brainstem compression, as well as posterior haemorrhagic infarct 5 .
We report a presence of the VT in 64.5% of the total sample size, with 61.0% and 68.0% presence in the left and right cerebral hemispheres, respectively (Table 2). These results are higher than those reported by Kawamata et al., with 59.0% in the left hemisphere and 54.9% in the right hemisphere 20 . In the absence of the VT, blood is drained by the superior cortical veins into the superior sagittal sinus or drained by an adjacent AV. Kilic and Akakin stated that on lateral angiograms, 8-12 superior cortical veins can be visualised, while Appaji et al. stated that there are 10-15 superior cortical veins; however, in the present study, a minimum of three and maximum of seven superior cortical veins draining the superolateral cerebral convexity in the absence of the VT was identi ed 11,18 . This study reports a low occurrence of the presence of VT in the 30-39 year age group (47.5%), a higher presence of VT in males (71.2%) in comparison to females (60.0%), and the lowest occurrence in the African ethnic group (62.8%) ( Table 2).
Each of the three AV have been reported to have double veins present, in a study conducted by Tanriverdi et al. it was reported that 9.9% of patients had a duplication of any of the three AV 8 . A vein was considered as duplicate if the AV had more than one distinct vein that drained the respective territory of the AV into its' respective dural venous sinuses, as seen in Figs. 1C-1F. We report a presence of 12% of double SSV (left: 13.0%; right: 11.0%); this is lower than the results of the study conducted by Bisaria, who reported a double bilateral SSV in 37.5% of their sample 10 Fig. 2).
Dominant patterns of the AV were ascertained by the territorial drainage of the veins, and in some cases the absence of a vein played little to no role in the dominance of the AV. For example, in the absence of the VT, in some cases the territory of the VT was drained by the superior cortical veins and not the accompanying AV, i.e SSV and VL. The system was determined to be an SSV/VL dominant drainage pattern if the SSV and/or VL crossed into the territory of the VT.
The Equilibrium dominant pattern of drainage had the highest occurrence (54.5%) in the present study and is aligned with the theory of haemodynamic balance of the SCVS, due to each AV draining their respective territories in a well-anastomosed system, as seen in Fig. 2A  to determine the correlation of the types of dominant patterns of the AV, based on laterality ( Table 5). The results of the present study indicate a weak correlation between the VL dominant pattern on the left hemisphere and VT dominance on the right (0.03), as well as a VL dominant AV on the right and a VT dominant AV on the left hemisphere (0.04). Results from the Cramers V correlation reported the highest correlation between dominant patterns of the AV, based on laterality, with a VL dominant AV on the left and right hemisphere having a medium correlation of 0.51 (Table 5).

Conclusion
Studies investigating the anatomy of cortical veins are imperative due to the course of these veins differing from cerebral arteries, as well as the high variability of these vessels. Since the AV are encountered during a common neurosurgical procedure, viz. pterional craniotomies, and due to the increased risk of iatrogenic injury to these AV, highlighting the anatomical variants regarding the presence, size and drainage patterns of these veins becomes important for neurosurgical perspectives. This study supports the statements of haemodynamic balance of the SCVS due to the seven types of dominant drainage patterns of the AV reported, as well as the high occurrence of the Equilibrium dominant pattern. The factor of laterality was not seen to be statistically signi cant amongst several results reported in this study, except for the diameter of the SSV; however, data re ects proportions of presence, diameter and dominant patterns based on laterality that must be noted by clinicians to aid in the planning of surgical corridors. Furthermore, the results from this study report occurrences of triple AV presence, which adds to the knowledge surrounding the AV.
Declarations SSV (SSV and SSV2) and Double VL (VL and VL2) E and F: Double VT (VT and VT2) G and H: Triple SSV (SSV, SSV2 and SSV3 -in ascending order, in an anteroposterior direction) I and J: Triple VL (VL, VL2 and VL3 -in ascending order, in an anteroposterior direction) K and L: Triple VT (VT, VT2 and VT3 -in ascending order, in an anteroposterior direction)

Figure 2
Angiographic images and schematic depictions of the dominant patterns of the anastomotic veins 2A and 2B represent an equilibrium or equidominant pattern 2C and 2D represent a singular dominant SSV pattern 2E and 2F represent a singular VL dominant pattern 2G and 2H represent a singular VT dominant pattern 2I and 2J represent a co-dominant SSV/VL pattern 2K and 2L represent a co-dominant SSV/VT pattern 2M and 2N represent a codominant VL/VT pattern