We selected the following anatomical structures as references for measuring the junction points of the superficial veins: central sulcus, frontomarginal sulcus, calcarine sulcus, and temporal pole. The selection was based on the following criteria: easy identification on the specimen, relative interindividual consistency of the structure, intraoperative applicability, favorable characteristics on imaging.

We used millimeter paper to measure the distances to account for the convexity of the brain. Anatomical structures define the location as the endpoints of the measured sections. Since the size of a brain varied between individual, we also documented the location of the central sulcus with respect to the frontomarginal and calcarine sulci. We measured the following structures: central sulcus midpoint to calcarine sulcus (Figure 1, line 1); frontomarginal sulcus to central sulcus midpoint (Figure 1, line 2); frontomarginal sulcus to confluence of the vein of Trolard into the superior sagittal sinus (Figure 1, line 3); confluence of the vein of Trolard into the superior sagittal sinus to calcarine sulcus (Figure 1, line 4); central sulcus midpoint to confluence of the vein of Trolard into the superior sagittal sinus (Figure 1, line 5); calcarine sulcus to confluence of the vein of Labbé into the transverse sinus (Figure 1, line 6); confluence of the vein of Labbé into the transverse sinus to the temporal pole (Figure 1, line 7).

We used Microsoft Excel® for the statistical analysis. We calculated the mean, standard deviation, minimum and maximum distances for all the measurements.

We classified the superficial veins according to five types listed in Table 1. For the right hemispheres the distribution was as following: Type I: six hemispheres (14.3%); Type II: three hemispheres (7.1%); Type III: eight hemispheres (19.1%); Type IV: three hemispheres (7.1%); Type V: one hemisphere (2.4%).

For the left hemispheres the distribution was: Type I: three hemispheres (7.1%); Type II: four hemispheres (9.6%); Type III: 10 hemispheres (23.8%); Type IV: three hemispheres (7.1%); Type V: one hemisphere (2.4%).

In summary, the 42 hemispheres are classified as such: Type I: nine hemispheres (21.4%); Type II: seven hemispheres (16.7%); Type III: 18 hemispheres (42.9%); Type IV: six hemispheres (14.3%); Type V: two hemispheres (4.7%).

Regarding the presence of the veins of Trolard and Labbé, there were no significant differences between the left and the right side in the present study. Due to the lack of data on handedness, no reliable assignment to hemispheric dominance was possible. If one postulates the left hemisphere as the dominant one, there appears to be a preference for both the vein of Trolard and the vein of Labbé on that side. Such relationship did not appear in studies by either Di Chiro (20) or von Lanz (21).

Di Chiro (20) described that the vein of Trolard occurred more frequently in the non-dominant hemisphere. These results were obtained using angiographic imaging; in the present microanatomical study those observations could not be confirmed. Instead, we found a more frequent occurrence on the left hemisphere for both the vein of Trolard and the vein of Labbé.

In his book (21), von Lanz described the superior anastomotic vein in the area of the central sulcus or slightly dorsal to it, noting that a course toward the anterior pole of the frontal lobe can also occur. We were able to differentiate this more precisely, describing the localization of the vein of Trolard in relation to the central sulcus. When viewed from both sides, the comparison between left and right showed no significant difference, the vein of Trolard was in front of the central sulcus in 8 hemispheres and in only four cases behind it. A confluence with the central sulcus was recorded in 6 cases. When measuring the distance of the contact point with the superior sagittal sinus, it appears that the frontal span of this location is significantly larger at 6 and 5 cm on the right and left, respectively.

Stephens and Stilwell (22) outlined the veins flowing into the superior sagittal sinus into a frontal and parietal group, with no information regarding frequency and exact position. However, they note that the vein of Trolard is the most variable of all superficial veins. According to Stephens and Stilwell (22), our data show a greater variability of the vein of Trolard than the vein Labbé when the veins flow into the superior sagittal sinus or the transverse sinus.

Thus far, there is no clear information in the literature regarding the confluence of the vein of Labbé into the transverse sinus. A sketch in Seeger's book Atlas of Topographical Anatomy of the Brain and Surrounding Structures (23) shows this region, and the confluence is 10.5 cm from the temporal pole. In our measurements, we had an average of 9.9 (± 0.9) cm for the right hemisphere and 9.7 (± 1.3) cm for the left (Figure 1, line 6).

The location of the confluence of the vein of Labbé into the transverse sinus, either via the tentorium or direct, is of great importance for the planning of subtemporal and occipito-basal approaches. Like the confluence angles of the bridging veins of the cortical convexity in the direction of the superior sagittal sinus, the vein of Labbé forms a very acute angle with the temporal base or with the occipito-basal tentorium. Within this acute angle, the surgeon need to pay attention moving toward the confluence of the Vein of Labbé: if the spatula retraction is forced too much, the compression may induce venous stasis.

In a paper by Oka et al. (24), 20 cerebral hemispheres were examined for the distribution of the superficial veins. With regards to the large anastomosing veins, the vein of Trolard, the vein of Labbé, and the Sylvian venous system, the paper proposed a classification into four types: Type I: the vein of Trolard is dominant; Type II: the vein of Labbé is dominant; Type III: the Sylvian venous system is dominant; Type IV: all three anastomosing vessels are present; the veins of Trolard and of Labbé are co-dominant. The authors made no statements about the frequency of each of the types. In addition to the types described by Oka, we found another type in which there is a direct connection between the vein of Trolard and the vein of Labbé without any connection to the Sylvian venous system (Type V in Table 1 and Figure 2).

Stephens and Stilwell (22) describe significant end-to-end anastomoses between superficial veins but do not propose any precise typing. They describe that the connection of the veins of Trolard and Labbé with the Sylvian veins is already present in the fetus; these connections remain during life, but the diameter of these vessels' middle segments increases.

In the present study we performed a systematic description and a numerical distribution of the superior anastomotic vein (vein of Trolard) and inferior anastomotic vein (vein of Labbé) for the first time. Furthermore, we proposed a new classification of the large anastomotic veins, describing five types of possible venous distributions for the vein of Trolard, the vein of Labbé, and the Sylvian venous system. The first four types reflect the classification proposed by Oka et al. (24), but a fifth type, a new anatomical pattern in the variability of the superficial vein distribution, was added.

Our classification, based on the distribution of the superficial veins showed the following five types: type I with a dominant vein of Trolard, type II with a dominant vein of Labbé, type III with a dominat Sylvian vein group, type IV is characterized by the presence of all three anastomosing veins (codominant veins of Trolard and Labbé with weak a Sylvian vein), type V shows a direct connection between the vein of Trolard and the vein of Labbé without a connection with the Sylvian vein group.