35.1 History and Overview of the Retrosigmoid Approach

During the early 1900s, the retrosigmoid approach was developed as a versatile technique to provide access to the posterior fossa. First described in detail by Fraenkel in 1903,s. Literatur the retrosigmoid approach went through many iterations before reaching its current form. Early versions were associated with high morbidity and mortality resulting from poor control of intracerebral pressure, injury associated with cerebellar retraction, and blood loss. This provided impetus for development of the translabyrinthine approach in 1904, which necessarily sacrificed hearing and was thought less useful for resection of larger tumors.s. Literatur Cushing addressed many of the problems associated with early unilateral retrosigmoid approaches by: (1) using a large bilateral suboccipital craniotomy to reduce the need for retraction and thereby limit injury to the cerebellum and medulla, (2) placing the patient in the prone position, and (3) releasing cerebrospinal fluid (CSF) from the cisterna magna or lateral ventricle to relieve intracerebral pressure. These innovations greatly reduced the mortality previously associated with the procedure.

Facial nerve injury was and is a serious complication of vestibular schwannoma (VS) resection. To avoid facial nerve dysfunction, Cushing advocated for subtotal resection and it was not until 1931 that Dandy first reported complete resection of a VS with preservation of the facial nerve. In contrast with Cushing’s bilateral suboccipital approach, Dandy resurrected the small unilateral suboccipital approach, exploiting many of Cushing’s morbidity-reducing innovations, but also reducing the risk to the facial nerve by carefully developing the dissection plane between the facial nerve and the tumor.

Subsequent modifications have included abandoning the prone position, use of the intraoperative microscope, and preoperative diagnostic imaging to identify patients with tumors best suited for the retrosigmoid approach as well as revealing critical anatomic details. More recent modifications of the approach include the extended retrosigmoid approach, which improves cerebellopontine angle (CPA) exposure by skeletonizing the sigmoid sinus (SS), improved intraoperative facial nerve monitoring with transcranial motor evoked potentials, and improved intraoperative monitoring of the cochlear nerve as well.

35.2 Advantages and Limitations

The retrosigmoid approach offers several advantages for appropriately selected cases (Fig. 35‑1 ). Preservation of the inner ear structures underlies the main advantage of the approach: hearing preservation remains an attainable goal of surgery even in tumors that are too large for the middle fossa approach or are predominantly extracanalicular. Whether the retrosigmoid approach is better suited for large tumors remains an area of controversy. Neurosurgeons traditionally receive more training in the retrosigmoid approach, but those trained in skull base techniques understand that the translabyrinthine approach in many ways epitomizes the tenets of skull base surgery—removing bone to obviate the need for moving or retracting the brain and removing a large tumor through a relatively small opening. Preservation of facial nerve function is possible with careful dissection techniques but may be more difficult given the more posterior approach to the anteriorly located facial nerve. The retrosigmoid approach is associated with a lower risk of CSF leak and reduced risk to vascular and nervous structures (SS, descending facial nerve). Finally, the retrosigmoid approach provides limited access to tumors that extend far into the internal auditory canal (IAC), so these lesions may be more appropriately accessed via translabyrinthine approach, especially after realistic consideration of the possibility of preserving hearing. Chapter 36 focuses on the retrosigmoid approach for small VS with the goal of hearing preservation. The current chapter primarily focuses on the retrosigmoid approach for medium or large-sized VS, where hearing preservation is less commonly a consideration. The interested reader is encouraged to read both chapters and review both summary tables for outcomes according to tumor size.

35.3 Patient Selection (Indications and Contraindications)

Traditionally, the retrosigmoid approach is considered best suited to tumors with a larger CPA component that do not reach to the fundus of the IAC and for whom hearing preservation is a possible goal of surgery. In our practice, however, we consider the retrosigmoid approach as primarily a hearing preservation approach and especially as a hearing preservation approach for tumors not suitable for a middle fossa approach. We prefer the middle fossa approach for hearing preservation in completely intracanalicular tumors, those within the lateral end of the canal, and tumors with less than a 10-mm extension into the CPA. For any tumor where hearing is poor or hearing preservation is very unlikely, we prefer the translabyrinthine approach, regardless of size. Finally, cases with a high-riding jugular bulb or a very anterior SS that might otherwise be suited for a translabyrinthine approach are better handled through a retrosigmoid approach. While there are no absolute contraindications to the retrosigmoid approach (save for unusual anatomic circumstances such as an obstructing developmental venous anomaly or other anatomic complication), it bears mentioning that the translabyrinthine approach can be easier in patients with an obese body habitus, since it allows access to the CSF early on without retraction of the cerebellar hemisphere.

35.4 Surgical Technique

Following intubation and Foley placement, the patient can be positioned either supine with the head turned as much as possible to the contralateral side or in full lateral decubitus position. In some locales, depending on training, neurosurgeons prefer the sitting position despite the added anesthetic complexity and risk because of the slackness of the cerebellar hemisphere. Continuous electrophysiological monitoring of the ipsilateral facial electromyogram (EMG), the ipsilateral facial motor evoked potentials, bilateral auditory brainstem responses (ABRs), and bilateral somatosensory evoked potentials should be considered. The Mayfield head holder is placed with two pins on the contralateral occiput and one pin on the ipsilateral frontal bone, with care taken to ensure adequate venous drainage with the neck rotated. This requires confirmation that the contralateral jugular vein is not compressed by the angle of the mandible. Many neurosurgeons have had the unfortunate experience of opening tight posterior fossa dura from complete occlusion of the contralateral venous drainage. We have gotten away from using the head holder altogether in routine retrosigmoid approaches for VS. In patients with a supple neck, it is easy enough to place the head on a foam donut and tape the head firmly to the bed after contralateral rotation. The tape has some natural give to it so that there is no issue with excessive neck rotation or occlusion of venous drainage. In patients lacking good neck rotation because of cervical spondylopathy or a short, thick neck, we do not hesitate to use the full lateral decubitus position with the head in pins. This position eliminates neck rotation altogether, resulting in a maximally relaxed cerebellar hemisphere even in morbidly obese patients. The supine patient should be positioned as far up and to the side of the table as possible to ensure a short reach for the surgeon. A C-shaped, anteriorly based, curvilinear incision is marked 4 cm posterior to the posterior edge of the mastoid, extending from the level of the mastoid tip to 2 cm above the pinna (Fig. 35‑2 ). Hair removal may be limited to 0.5 cm on either side of the planned incision. Perioperative antibiotics are administered and the site antiseptically prepped and draped. We prepare every patient for harvest of an autologous abdominal fat graft to be used in the closure for obstructing exposed mastoid air cells. The abdominal harvest site, either periumbilical or in the left lower quadrant, is similarly antiseptically prepped and draped.

The incision should be carried down to the periosteum covering the asterion, posterior mastoid, and the inferior surface of the occipital bone. The superficial skin flap is reflected anteriorly off the periosteum, and a periosteal flap, the so-called Palva flap, is cut with the same incision, also reflected anteriorly, and secured with fish hook retractors on rubber bands. At this point, there are many ways to expose the posterior fossa dura. For neurosurgeons, a craniectomy remains more common than a craniotomy but requires cranioplasty with foreign materials for closure. We prefer to avoid foreign materials as much as possible by performing a craniotomy. Rather than using a burr hole, we instead use the technique of exposing the SS and 2 mm of dura behind it to allow access for the craniotome footplate. Using cutting and diamond burrs, the SS is skeletonized caudally from the sigmoid–transverse junction (found deep to the asterion) to the jugular bulb, leaving only a thin layer of bone covering the posterior portion of the SS and exposing 2 to 3 mm of retrosigmoid dura. Hemostasis should be obtained by waxing air cells and with careful and definitive control of the emissary vein when it is encountered using bipolar cautery (Fig. 35‑3 ). The exposed retrosigmoid dura is now used as a starting point to free the dura from the overlying bone. Similarly, the footplate of the craniotome can now be inserted in the epidural space under direct vision without the need for additional burr holes, craniectomy, or bone loss. The craniotome is used to create a 2- to 3-cm retrosigmoid craniotomy in the occipital bone, and the bone flap is removed from the field. Some authors have touted endoscopic approaches to the CPA, claiming that the burr hole used is only 1 inch (2.54 cm) in diameter. We commonly use 2-cm craniotomies and seldom create craniotomies greater than 3 cm for the resection of VS, using the technique described.

Intracranial pressure is reduced through the usual maneuvers of low-dose mannitol (0.25–0.5 g/kg dose), temporary hyperventilation, and reverse Trendelenburg position until the dura is relaxed. The dura is then opened with a posteriorly curved incision 5 mm posterior and parallel to the sigmoid and following along the transverse sinus (Fig. 35‑4 ). The sinus is retracted anteriorly with a tack-up suture through the attached dural flap and tack-up holes drilled in the mastoid bone anteriorly. Under the operating microscope, the inferior arachnoid cistern either over the lower cranial nerves or over the lateral medullary cistern is opened sharply to allow CSF egress. The cerebellum will reliably relax posteriorly, eliminating the need for any retractor use after opening the cistern. The key is to rely on the microscope and to protect the inferior cerebellar hemisphere as it is lifted superiorly and posteriorly with the bipolar and microsuction. The lower cranial nerves can be identified at this point.

With the cerebellum relaxed and the sigmoid displaced slightly anteriorly, the exposure of the posterior petrous bone is exquisite and complete, and the tumor is now visible in the CPA, all without the use of a fixed retractor of any type. Even in the presence of a large tumor, opening the arachnoid cisterns and patiently draining CSF will allow the cerebellum to fall out of the way by gravity only. The cerebellar hemisphere is protected with a single Telfa strip so as to maximize the view angle and avoid any entanglement in the drills used to remove the posterior lip of the IAC. Care is taken to confirm that the facial nerve is not running over the posterior face of the tumor using direct stimulation of the nerve. Similarly, if hearing preservation is a goal of the surgery, the eighth nerve is now identified on the inferior pole of the tumor. The free-running facial nerve EMG and transcranial facial motor evoked potential are used throughout the ensuing dissection to judge the health of the facial nerve. It is always useful to note the flocculus and the relationship of the bulk of the tumor to the flocculus and the eighth cranial nerve emerging just deep to it. If, for example, the entire tumor appears to be superior to the flocculus, the surgeon can anticipate that this is a superior vestibular nerve tumor and expect the inferior vestibular nerve to protect the cochlear nerve with no involvement of the lower cranial nerves. The tumor is now internally debulked using the ultrasonic aspirator, so that the walls of the tumor can now be collapsed inward and the tumor dissected away in turn from the cerebellar hemisphere, the flocculus, and finally the brainstem and cranial nerves. Superiorly, the tumor is dissected free of the ala of the cerebellum, the trigeminal nerve, and the petrosal vein (Dandy’s vein). We make it a point to try to preserve the petrosal vein whenever possible, just to have the security of additional intact venous drainage. If, however, the petrosal vein has a short trunk and is under stretch, we coagulate and divide it sharply before it can be pulled free of the petrosal sinus with potentially voluminous bleeding. Senior residents learning the retrosigmoid removal of VS become obsessed with finding the facial nerve early and distressed at their failure to do so. No matter the size of the tumor, the eighth nerve and seventh nerve root entry zones remain in a relatively fixed relationship. Find the eighth nerve trunk and the seventh nerve will emerge from the brainstem just inferior and medial to it, separated from the eighth nerve by a small but constant recurrent branch of anterior inferior cerebellar artery (AICA; Fig. 35‑5 ). The eighth nerve trunk will often find the inexperienced VS surgeon, rather than vice versa. More than one young surgeon has said that in removing a large VS, they elected to stop the dissection when they encountered significant venous bleeding at the brainstem. That bleeding more often than not stems from the veins of the eighth nerve trunk, which expand and dilate as the tumor is removed offering a clumsy yet effective confirmation of where the surgeon is in the dissection. On large tumors where hearing preservation is not a true goal, controlling the veins around the eighth nerve trunk allows the surgeon to coagulate and divide the thinned-out nerve, revealing the facial nerve just inferior and medial, emerging directly from the pontomedullary sulcus. Alternatively, it can be a very effective strategy to identify and maintain the flattened and splayed eighth nerve during the dissection, even if hearing preservation is not a consideration. The almost translucent eighth nerve can be kept as a veil to protect the facial nerve deep to it during the dissection. The facial nerve can now be identified by direct stimulation at the brainstem, even through the intact but splayed eighth nerve. The tumor can now be dissected off the facial nerve at the brainstem, finally freeing the tumor from all other brainstem attachments. As the tumor is now dissected off the facial nerve and eighth nerve veil along the brainstem, a point comes where the facial nerve leaves the brainstem to ascend to the superior lip of the porus acusticus. This is an area in large tumors where the nerve can be especially vulnerable. At this point, rather than lifting the tumor blindly off the nerve, it can be effective instead to find the facial nerve medially in the IAC and dissect in the other direction from lateral to medial.

The intracanalicular extension of the tumor, which is almost always present, requires that the posterior wall of the porus be drilled out; between 6 and 9 mm can usually be removed without compromising inner ear structures. An H-incision should be made in the petrous dura overlying the IAC, creating “saloon doors” which can be used for the closure. The IAC will be drilled laterally (Fig. 35‑6 ) with the limiting structures being the vestibule and the semicircular canals laterally and the jugular bulb inferiorly. The dura of the canal is opened longitudinally and the vestibular nerves are identified. This is a critical point of the operation in avoiding recurrences from small tumor residuals left at the fundus. The vestibular nerves are divided sharply laterally, revealing the lateral dome of the tumor. Inability to identify a smooth lateral dome suggests that the lateral tumor may have been transected with a portion left behind. A combination of tiny right-angled curettes and dissectors can be used to probe the lateral end of the canal. It is at this point in hearing preservation operations where the risk of recurrence rises because of a hesitance to dissect too roughly and blindly at the fundus of the canal in a retrosigmoid approach. Larger tumors require debulking in every phase of the operation with the ultrasonic aspirator to ease dissection off the facial and cochlear nerves, and dissection should always proceed from both ends of the facial nerve to meet at the cisternal segment (Fig. 35‑7 ). The very smallest tip of the ultrasonic aspirator is advantageous, especially when used on the IAC portion of the tumor. The superior and inferior vestibular nerves must be transected to avoid prolonged postoperative vertigo. This is another point in a hearing preservation case where the desire to protect the cochlear nerve and ABR can lead the surgeon to leave intact vestibular fibers with the possible consequence of recalcitrant dizziness. The last critical point of the dissection comes just outside or at the porus acusticus, where the facial nerve is thinnest. The experienced surgeon will preserve cochlear nerve fibers in every case at least temporarily in order to prevent the facial nerve from bearing the full weight of the remaining tumor. Once the tumor is completely removed, the field should be copiously irrigated and inspected for any residual tumor pieces. Bleeding near the facial nerve or an intact cochlear nerve with a preserved ABR should be controlled with Gelfoam and not with bipolar cautery in order to prevent inadvertent nerve injury and ruining the outcome of an otherwise exquisite dissection.

After closing the “saloon doors” of the posterior petrous dura with a single 4–0 suture, an abdominal fat graft should be harvested and placed in the IAC to seal the canal and prevent CSF leak from perimeatal air cells. Tissue glue can be used to hold the fat graft in place during healing. Hemostasis is confirmed with clear irrigation and should be meticulously maintained during closure. After closing the retrosigmoid dura and reinforcing the suture line with tissue glue, a second fat graft is placed over the SS and the bone flap is replaced with three small titanium plates. The pericranial flap is reapproximated to the surrounding periosteum. The nuchal muscles are reapproximated to the surrounding musculature and nuchal line. The skin is closed in two layers in the usual fashion. Facial nerve function should be assessed shortly after extubation whenever possible so that delayed facial paresis can be identified easily and early. Postoperative hearing assessment generally is delayed until a formal audiogram can be done and until the mastoid and middle ear are fully aerated.

In addition to standard postoperative care, patients should be monitored closely for CSF leak, which may manifest as clear fluid discharge from the incision or as rhinorrhea. Serial neurological exams should be used to monitor for cerebellar or brainstem edema, hydrocephalus, and cranial nerve dysfunction. Hearing preservation after surgery is typically durable, although studies continue in attempts to identify medications or maneuvers that can prevent postoperative hearing deterioration. We continue to use nimodipine postoperatively in patients with preserved hearing who tolerate the medication without blood pressure compromise. Depending on how much manipulation of the facial nerve was required in the dissection and how compromised the nerve was by the tumor, some facial dysfunction may be present during the immediate postoperative period. Patients should be counseled, however, to expect improvement if the nerve was intact and simulatable at a low threshold (<0.1 mA) at the root entry zone at the end of the dissection. Additionally, a minimal change in the stimulation threshold for the facial motor evoked potential is also a predictor of intact facial nerve function postoperatively.