The scalps were removed with special attention to expose the inion. The cranial vault (calvaria) above the axial plane across the nasion and the inion was removed by a surgical power device (Xishan, China). Then, the brain tissue around the straight sinus was en bloc resected, whereas the brain tissue around the junction between the GCV and straight sinus was removed underneath the pia mater in piecemeal fashion. The straight sinus was flushed with tap water to remove blood clots. Then, a high-definition rigid endoscope (Karl Storz, Germany) measuring 2.7 mm in diameter and optics (0 and 30°) was inserted into the lumen of the sinus from torcula herophili. Special attention was given to the morphology of the CW and junction between the GCV and straight sinus.

The GCV, partial inferior sagittal sinus, and straight sinus were carefully removed en bloc under a surgical microscope (OPMI6, Zeiss). The superior sagittal sinus (SSS) and right and left transverse sinus (TS) were transversely sectioned 3 cm from the torcula herophili. The observation area included the anterior SS, posterior straight sinus, and GCV.

Five venous sinus samples containing nodules at the junction between the GCV and straight sinus were prepared for microscopic assessment. Hematoxylin and eosin (H&E) staining, Masson’s trichrome staining (for detecting collagen fibers), and Victoria blue staining (for detecting elastic fibers) were applied to sectioned samples. To differentiate the nodules from the arachnoid granules, 10 arachnoid granulations in the SSS were also subjected to H&E, Masson’s trichrome, and Victoria blue staining. The histological sections were analyzed using a Zeiss Axioskop plus microscope (Carl Zeiss Microscopy) at ×50, ×100, and ×400 magnifications. Images were acquired and stored by the AxioVision software. For elastic fibers quantitation, a series of five equally spaced tissue sections (~0.5 mm apart) spanning the entire nodes and arachnoid granules was stained with Victoria blue. The elastic fibers quantification was determined as the % of blue staining area in three randomly chosen fields (×400) using the ImageJ software. All samples were conducted in a blind manner.

A specific antibody against the N-terminal portion of SMA-a protein (mouse anti-human SMA-a monoclonal antibody) was used in an immunohistochemical study on samples containing the nodule. Paraffin embedded samples were cut into 4 μm thick successive coronal sections. After deparaffinization and dehydration in serial alcohol concentrations (70–100%), the endogenous peroxidase activity was blocked for 30 min in methanol containing 0.3% hydrogen peroxide. Moreover, 5% bovine serum albumin (BSA) was used for avoiding nonspecific binding of antibodies. Antigen retrieval was performed in target retrieval solution for 30 min. Sections were then incubated with primary antibody diluted 1:100 for 30 min at room temperature. A horseradish peroxidase-labeled secondary antibody was applied for 30 min at room temperature after the primary antibody.

There were 135 patients enrolled in the study, including 59 male and 76 female patients. The mean age at diagnosis was 50.63 ± 15.23 years (range: 27–76 years). The indications for MRI-enhanced analysis were suspicion of an intracranial lesion (48 patients) and evaluation of cerebral vascular diseases (65 patients) or tumor (22 patients). Exclusion criteria were: (1) cerebral vascular diseases involved with the SS and GCV; (2) intracranial tumor involved with the pineal region, third ventricle, and corpus callosum; and (3) the GCV and straight sinus were unclear in the imaging. MRI images were obtained as previously described (Kim et al., 2019).

All statistical analyses were performed using SPSS 22 for Windows (SPSS Inc., Chicago, IL, USA). Categorical data, including numbers and percentages of CW, were summarized using descriptive statistics. Numerical data were expressed as means, SDs, minimums, and maximums. Independent-sample Student’s t-test was used to compare the distribution of elastic fibers in the nodules and that in the arachnoid granulation. One-way ANOVA with Tukey’s post hoc test was used to examine the difference in diameter among the GCV, GCV–SS, and SS.

Neurosurgeons have increasingly turned their attention to the CW that were recognized by early anatomists. Schmutz stated in detail about the morphological characteristics of CW in the SSS and divided the CW into three different forms: valve-like lamellae, longitudinal lamellae, and trabeculae (Schmutz, 1980). Some researchers confirmed cord-like structures in the lumen of cerebral veins ostia and networks of accessory cords in the venous lacunae (Sharifi et al., 2004; Shao et al., 2009). Similar to our data, valve-like lamellae was the most common type of CW in the straight sinus. Different from the bridge veins that emptied into the SSS, the bridge veins around the tentorium cerebelli, which collect venous drainage of the cerebellum, occipital lobe, and temporal lobe, flowed into the tentorial sinus and then entered into the straight sinus or TS (Matsushima et al., 1989; Muthukumar and Palaniappan, 1998; Rosenblum et al., 2019). This pattern of venous reflux revealed that there was no direct connection between the bridge vein and the straight sinus, and the valve-like lamellae might just be a drainage channel in the SS to guarantee a laminar venous blood confluence in the sinuses. The longitudinal chords were mostly detected at the opening of the straight sinus. They divided the SS into double or triple lumens (Kaplan et al., 1972; Kaplan and Browder, 1976), which increased flow velocity and led to a decrease of static pressure at the orifices.

The knowledge of intraluminal structures helps the interventional neuroradiologists to navigate microfilaments and catheters through the dura sinus (Farb, 2007; Iwanaga et al., 2020). The existence of valve-like lamellae and longitudinal lamellae in the straight sinus could cause the catheter moving into a blind cavity in the posterior portion of the straight sinus. A catheter could progress into the narrow channels surrounded by the trabeculae and the sinus wall and cause the stent to open unsuccessfully (Figures 7A–E). Therefore, understanding the detailed anatomy of the straight sinus may help surgeons to avoid procedural difficulties and to achieve higher success rate.

The physiological regulation of the deep cerebral venous system has not been well understood. What is more, few researchers have paid attention to the anatomy of the junction between the GCV and the straight sinus. Velut described a fold of the GCV and the origin of the junctions in the straight sinus and assumed that the fold would create a pouch between the venous wall and the dura (Velut, 1987). Ghali et al. (1989) reported the presence of a body and a small elevation at the junction, which had a smooth surface and convex upwards with dimensions ranging between 3.5–4.5 mm in length and 2.5–3.0 mm in width at the basal part. In the present study, under the endoscope, we categorized the junction into three types: type 1 junction has folds in the GCV and has an elevation on the floor of the SS; type 2 has folds and bulges at the junction; and type 3 has a nodule and extends on the horizontal plane of the cord.

MRI can clearly define normal intraluminal architecture within the dura venous sinuses. Arachnoid granulations were considered well-defined, hypointense, or isointense on T1- and hyperintense on T2-weighted sequences and the nature of the intra-sinus filling defects images (Farb, 2007). Haroun et al. (2007) showed a 75-year-old female patient with the presence of rounded, well-defined intraluminal structure at the junction of the straight sinus and the vein of Galen. In our study, the elevation, small bugle, or nodule was revealed by MRI-enhanced analysis. The fold was not presented for blood flow in the junction. Based on the anatomic types of junctions, type 1 was the most common in the general population.

A risk of hemorrhage existed if the microwire/microcatheter catches on the origin of veins while maneuvering along in the deep venous system. Based on the types of junction, the tip of the microcatheter can be steam shape as an arc or S-shape with different angles. Shaped catheters aid access to branches at angles and provide a stable position once placed in the venous cavity without contacting the venous wall. In addition, the GCV–SS was easy to tear in type 2 for thin organizational structure on the posterior wall of the GCV during thrombectomy in the SS. Moreover, the anatomic types of junction mean different physiological regulations of venous blood flow in the GCV–SS.

Traditionally, the arachnoid granulation tissue could form a nodule that appeared as a ball-shaped valve preventing the venous return from the third and lateral ventricles and effecting the secretion of the cerebrospinal fluid. Unlike an arachnoid granulation that contains collagen fibers and a few amounts of the elastic fiber, we found that the nodule contains elastic fiber and smooth muscles in addition to collagen fibers. Since smooth muscles were observed at the intersection between the GCV and the wall of the straight sinus, the smooth muscles of the nodule probably had the same histological origin as the GCV. The existence of elastic fibers and smooth muscles was to maintain the physiological regulation of venous blood flow in the straight sinus. The smooth muscle in the GCV could increase GCV pressure; meanwhile, the smooth muscle contracts to reduce the diameter of the nodule and thus reduce its volume, and finally, the venous blood flows into the straight sinus through their combination.

The nodule and SS have similar histological structure. All the nodule and SS exhibited an endothelial cell, a thin layer of subendothelial connective tissue containing collagen and elastic fibers, and the presence of a muscle layer. In our study, the collagen fibers in the nodule were connected with the inferior wall of the SS, with the same course. Elastic fibers were also demonstrated on the surface of the SS, which connected with that of the nodule.

In embryonic development, the intraluminal structures in the dura sinuses probably come from the cells of the primitive veins in the epidural space, which merges to form different sized venous sinuses. During the formation of the SS, folds may form since the brain tissue is compressed between the mesencephalon and rhombencephalon (Senders et al., 2018). Accompanying the formation of these folds, the anterior and posterior plexuses combine to form the various venous sinuses. Based on the differences in the forms of fusion among different individuals, many variations, including position in the TC, channel duplication, and CW, appeared in the subsequent sinuses (Kaplan and Browder, 1976).