The Dorello Canal

Historical Development, Controversies in Microsurgical Anatomy, and Clinical Implications

Varun R. Kshettry, M.D; Joung H. Lee, M.D; Mario Ammirati M.D., M.B.A


Neurosurg Focus. 2013;34(3):e4 

In This Article

Controversies in Microsurgical Anatomy

Relatively little was published on the anatomy of the petroclival region after Dorello's time until recent developments in the surgical treatment of petroclival tumors led to renewed interest in the anatomy of the region.[1,2,30,31] Additionally, reports of abducent palsy associated with blunt trauma, hydrocephalus, intracranial hypotension, and aneurysm further increased interest in studying the course of the abducent nerve as a means of explaining the etiology of these clinical entities. Table 1 summarizes the literature on microsurgical anatomical studies of the Dorello canal.

Description, Anatomical Boundaries, and Dimensions of the Dorello Canal

The introduction of the operating microscope in the latter half of the 20th century has allowed more detailed anatomical study. In 1991, Umansky et al.[36] performed the first anatomical study of the Dorello canal in human cadavers using a microscope. Their description of the canal was consistent with Primo Dorello's original definition (referred to as the classic definition in Table 1), as the space between the petrous apex and the superolateral portion of the clivus bounded superiorly by the petrosphenoidal ligament of Gruber (Fig. 1). Since 1991, there have been numerous anatomical studies of the Dorello canal,[11,19,20,26,27,35] and some authors have argued for a revised definition. In 1997, Destrieux et al.[11] found specimens in which the abducent nerve ran superior to the Gruber ligament; these authors therefore argued for a broader definition of the canal to accommodate the observed variations in the course of the nerve. Destrieux et al.[11] argued for the term "petroclival venous confluence" (PVC), rather than Dorello canal and defined this space more broadly as the area between the inner (meningeal) and outer (periosteal) layers of the dura mater between the petrous apex and the margin of the sphenoid bone just below the posterior clinoidal process. Beside Dorello's canal and other structures, the PVC contains this dural venous sinus, which is fed by the posterior cavernous sinus, lateral basilar sinus, and superior petrosal sinus and empties into the inferior petrosal sinus (Figs. 2 and 3). Dolenc and Yaşargil[12] argued that the Dorello canal is not even a canal in the true sense because it is not surrounded by bony walls like the optic canal, for example. He defined the boundaries as the larger space between the dural leaflets starting from the point where the abducent nerve pierces the dura to its entry into the cavernous sinus. Table 1 lists the dimensions of the Dorello canal in the published anatomical studies. Studies that utilized the term "petroclival venous confluence" generally did not provide any dimensions for the canal because, according to this concept, certain boundaries of this space consist of virtual planes.

Figure 1.

Rotated 3D computer graphic views of the Dorello canal as the space between the petrous apex and superolateral portion of the clivus bounded superiorly by the petrosphenoidal ligament of Gruber (asterisk, left). Also depicted are cranial nerves III (white arrowhead), IV (white arrow), V (black arrow), and VI (black arrowhead).

Figure 2.

Rotated 3D computer graphic views of the anatomy of the PVC (orange area), which is fed by the posterior cavernous sinus (CS), lateral basilar sinus (BS), and superior petrosal sinus (SPS) and which empties via the IPS.

Figure 3.

Rotated 3D computer graphic views of the PVC (orange area). The PVC is bounded superiorly by the posterior petroclinoidal dural fold, inferiorly just below the dural porus of the abducent nerve, medially by the IPS-BS junction, and laterally by the medial petrous apex. Superiorly, the PVC is bound by a single plane. Inferiorly, the PVC extends anterior to posterior from the posterior CS to the posterior wall of the IPS.

Anatomy of the Gruber Ligament

All studies describe the Gruber ligament as butterfly shaped and silver in color. Four studies provide a mean length of the ligament, varying from 11.5 to 13.3 mm.[19,20,35,36] Five studies with a total of 208 specimens describe variations in the ligament,[11,19,20,37,36] which was ossified in 13 specimens (6.3%) and hypoplastic in 7 (3.4%).[11,19,20,35,36] Only 2 studies mention duplication of the ligament in 5%–15% of specimens.[35,36] Iaconetta et al.[19] found that the Gruber ligament was in continuity with the petrolingual ligament. This ligament runs from the petrous apex to the lingual process of the sphenoid bone at the foramen lacerum[20,40,41] and covers the lacerum segment of the carotid artery. Iaconetta et al.[19] refer to the combined petrosphenoidal and petrolingual ligament as the falciform ligament. When the Gruber ligament was mentioned, all studies described it as being immersed in venous blood.

Course of the Abducent Nerve

Five studies describe the relation of the abducent nerve to the Gruber ligament, with the nerve coursing inferior to the ligament in 89 (98.9%) of 90 specimens.[11,26,27,35,36] Three studies provide the position of the abducent nerve in the canal.[27,35,36] It was located in the medial third in 2 (4.2%) of 48 specimens, the middle third in 15 (31.2%) of 48, and the lateral third in 31 (64.6%) of 48 specimens. Three studies describe duplication of the abducent nerve in 13 (8.8%) of 148 specimens.[11,19,36] In Iaconetta's sample of 100 specimens, the abducent nerve was duplicated at the origin in 2 specimens while in 6 specimens the nerve arose as 1 trunk, which then divided into 2 roots that separately pierced the dura to enter the Dorello canal. In all 8 cases, the split roots fused to enter the cavernous sinus as a single nerve.[19] This finding is consistent with a large study done by Jain[21] in 1964, which found that the abducent nerve was duplicated in 18 (6%) of 300 cases. In all of these aberrant cases, the 2 rootlets arose separately at their ventral pontine origin, generally as a larger medial root and a smaller lateral accessory root. In 17 of 18 cases, the rootlets merged in the cavernous sinus. Only in 1 case did the roots remain separated in their entire course, even penetrating the lateral rectus muscle separately. Another study by Nathan et al.[25] in 1974 found duplication of the abducent nerve in 9 (14.5%) of 62 specimens. In 4 cases (6.5%), the abducent nerve originated as a single trunk, and then immediately divided into 2 roots that pierced the dura separately. In 5 cases (8.1%), the nerve emerged from the pons as 2 roots that pierced the dura separately. In all 9 cases of duplication, 1 root ran above and 1 ran below the Gruber ligament, and in all cases the roots fused in the cavernous sinus. In summary, abducent nerve duplication was found in 35 (7.6%) of a total of 462 specimens described in the literature. When duplication occurred, 2 roots emerged at the pontine origin in 25 (71%) of 35 specimens, and a single root split into 2 rootlets within the pontine cistern in 10 (29%) of 35 specimens. Duplicated roots fused into a single nerve in the cavernous sinus in 34 (97%) of 35 cases.

The abducent nerve was covered by a dural sheath in all anatomical studies. Histological studies demonstrate that, at its entry at the dural porus, the nerve carries with it a dural sheath and an arachnoid membrane containing its blood supply from the cisternal segment of the nerve.[26,27] The dural sheath has multiple connective tissue attachments to the Gruber ligament, the endosteal dura of the PVC, the lateral wall of the cavernous ICA, and the medial portion of the Meckel cave.[11,19,20,26,27,35,36] These attachments limit the mobility of the nerve in compressive or stretch injuries. One study directly measured the length of the abducent nerve through the PVC at a mean of 11.5 ± 3.7 mm.[20]

Vasculature of the Dorello Canal

Arterial supply to the petroclival region is provided by branches from the posterior bend of the cavernous ICA. Studies generally describe a meningohypophysial trunk that branches into an inferior hypophysial artery, tentorial artery, and meningeal branches. The meningeal branches consist of a medial clival artery (DMA) and a lateral clival artery. All studies indicate that the medial clival artery travels to supply the dura of the dorsum sellae and upper clivus. The lateral clival artery supplies the lateral petroclival dura and the dural sleeve of the abducent nerve. Initial studies indicated that the lateral clival artery directly supplies the abducent nerve, but histological analysis done by Ozveren et al.[26,27] demonstrates that the lateral clival artery only supplies the dural sheath of the nerve. Blood supply for the actual nerve comes from tiny arteries arising from the subarachnoid space that travel in an invaginated arachnoid membrane where the nerve penetrates the dura. In some specimens, the clival arteries originate directly off the cavernous ICA. Four studies explicitly describe the relationship of the DMA to the Gruber ligament. The DMA traveled under the Gruber ligament in 151 (83.9%) of 180 specimens.[19,20,35,36]

Although Dorello included the IPS within the canal in his original description, all studies done using the operating microscope have demonstrated that the ostium of the IPS is outside of the canal. Three studies explicitly describe the relationship of the ostium of the IPS in relation to the abducent nerve.[20,35,36] The ostium of the IPS is lateral to the abducent nerve in 57 (71.3%) of 80 specimens and medial to it in 23 (28.8%) of 80 specimens.