Indication for surgery is based on imaging. High-resolution MRI allows to predict neurovascular compression (NVC) and, if MVD is decided, to help surgical planning and design microsurgical approach that should be tailored according to individual features of every patient. To achieve high sensibility and high specificity, MRI exploration should associate the three following high-resolution sequences: 3D T2 to finely depict the neural structures and vessels and degree of compression, 3D T1 with gadolinium to show both the arteries and veins, and 3D TOF angiography to differentiate the arteries from veins. As a matter of fact, the later sequence preferentially, if not exclusively, shows vessels with high velocity, i.e., the arteries, in almost an exclusive way when a superior band of presaturation filter is put to mask veins.

The main complication of MVD after surgery is hearing loss. As demonstrated by intraoperative BAEP recordings, principal cause is stretching of cochlear nerve due to lateral-to-medial retraction of the cerebellar hemisphere, Therefore, approach of the IXth–Xth REZ should be from below, passing inferolaterally to the cerebellar hemisphere and tonsil, i.e., following an infrafloccular trajectory along the cerebellomedullary fissure. The lesser the exposure of the cochleovestibular nerve complex, the better for hearing preservation. Figure 11.3 shows a schematic anatomic drawing of the lower part of the cerebellopontine angle, approached via an infrafloccular trajectory. Our current approach is as follows.

Under general anesthesia, the patient is placed in the contralateral decubitus position, the head in a three-pin holder, slightly flexed and rotated 15° toward the contralateral side. The neck is flexed to the contralateral side to access the retrocondylar region without view obstructed by the shoulder, but not too much to avoid stretching of the brachial plexus especially in patients with a gracile neck. The ipsilateral shoulder is taped and moderately pulled caudally and posteriorly. Hair is shaved in the retromastoid region. Then mastoid borders are identified by palpation with the index finger and landmarks of craniectomy drawn posteriorly to the tip of the mastoid process (Fig. 11.4).

The skin incision – 5 cm in length – is made obliquely, 1 cm medial to the bisector of the angle formed by the nuchal line and the posterior aspect of the mastoid process. The underlying subcutaneous tissue and muscles are divided using electrocautery, not too extensively to avoid damaging the occipital nerve. If the occipital artery is encountered, which is frequent with this approach, it is divided between two silk ligatures. The posterior aspect of the mastoid process together with the digastric groove and the retrocondylar fossa are cleared of soft tissue. The mastoid emissary bony vein is obliterated first with a small pledget of Surgicel and then waxed.

Then, a retromastoid craniectomy of the keyhole type is performed, posterior to the tip of the mastoid process, next to the retrocondylar fossa, with a semilunar shape of 2 cm in length and 1.5 cm in width, just posterior to the sigmoid sinus. Goal is to expose the inferolateral aspect of the cerebellum and in the depth cistern of the IXth and Xth nerves. The burr hole must not be turned too laterally onto the sigmoid sinus as this could endanger the external wall of the sinus, which is often reduced to a thin endothelial layer adhesive to the bone. If bleeding occurs, suturing would not be possible owing to its friable texture. Use of a Doppler microprobe may help in detecting the posterior border of the sigmoid sinus. If mastoid cells are opened – which is common – they are occluded by affixing a piece of subcutaneous tissue, e.g., fat plus aponeurosis (fascia lata harvested from the thigh). The dura is opened by making a small flap retracted along the sigmoid sinus. A self-retaining retractor – of the Yasargil type – mounted with a very thin blade (Sugita-Fukushima type) is placed on the inferolateral aspect of the cerebellum. No vein is on the way, excepted sometimes a (tiny) inferior petrosal vein that can be coagulated and divided without consequence, at least in our experience.

Then microscope is installed; and microsurgical steps start (Fig. 11.5). After incision of the arachnoid from XIth up to VIIIth nerve, the nerves are approached inferolaterally to the tonsil, to reach their root entry/exit zones (REZ) at the ventrolateral aspect of the medulla. REZ are often covered by the choroid plexus emerging from the lateral foramen of Luschka that must be gently retracted to expose REZ and vessels at the brainstem.

Not infrequently arachnoid is found thickened and strongly adhesive to the IXth–Xth rootlets. Dissection of rootlets needs to be meticulous, sometimes one by one, and as atraumatic as possible. Arachnoid chordae have to be divided using fine microscissors to free rootlets from vessels. Their sharp dissection should be completed before their mobilization.

The compressive vessels are the posterior inferior cerebellar artery (PICA), the vertebrobasilar artery (VBA), and frequently the association of both. Compression is most often ventral to the rootlets, which implies that maneuvers on the conflicting vessels be done passing in between rootlets at several interspaces. During mobilization of the compressive artery(ies), care should be taken to respect their tiny perforating collaterals and not to generate vasospasm.

Throughout vascular manipulations, gentle irrigation with saline and, time to time, application of a few drops of papaverine in solution (1 ml in 10 ml saline) are important measures to limit the mechanical vasospastic reactions. However, not too much of papaverine should be used because of its very acid PH.

After the offending artery(ies) has(have) been dislodged from its(their) conflicting situation, often marked by an invagination into the ventrolateral surface of the medulla, the vessel(s) has(have) to be maintained apart in such a way not to go back to the previous compressive location.

In cases when arterial loops have a sufficient laxity, transposition is relatively easy. The loop is kept apart by means of sling(s), approximately 3–4 cm in length and 2–3 mm in width, made of shredded fibers of Teflon Felt, passed around the artery to exert a pulling effect, and blocked to avoid recurrence of malposition. Conversely, in cases where the compressive vessel is rigid, transposition is usually rather difficult. The vessel cannot simply be pulled away. A small piece of semirigid prosthesis (Dacron or Teflon) and/or a small balled cushion of Teflon fibers is interposed between the REZ of the brainstem and the artery. Care must be taken not to exert any “neurocompressive effect”.

Before the microscope is taken out, the surgeon must verify that arteries have no kinking or twisting. Irrigation of the vessels with warm saline and a few droplets of papaverine in solution is a wise precaution to suppress possible spasms due to surgical manipulations. Venous hemostasis is checked by asking the anesthesiologist to perform sustained digital compression at the neck of both jugular veins or, if this is not possible, by carrying out a Valsalva maneuver with the ventilation machine.

Then the dura is closed either with single stitches or to better achieve watertight closure with a small patch of fascia lata. Additional fatty tissue is affixed onto the mastoid cells if opened. Bone chips are packed over the craniectomy defect only if no mastoid cells were opened. Finally, the muscular, subcutaneous, and cutaneous layers are closed with interrupted sutures and a compressive dressing is applied to avoid pseudomeningocele and decrease the risk of cerebrospinal fluid fistula or of rhinorrhea through the Eustachian tube if mastoid cells were opened.

In our series of 36 MVD for VGPN, there was complete pain relief in the long-term follow-up, i.e., with more than 2 years of follow-up, in more than 94.4 % of the cases. There was no mortality, no general morbidity, and no wound complication. As regards to neurological outcome, there was no postoperative permanent deficits, except two cases with a disabling IXth–Xth sensory and motor deficits.

The first attempted procedures for treating VGPN were the extracranial section of the glossopharyngeal nerve (Sicard and Robineau 1920), soon after the avulsion of the nerve at the jugular foramen via a high cervical approach (Adson 1924). On account of a high morbidity and a high incidence of recurrence of pain, these methods were not long propagated. It was Dandy in 1927 who gave the explanation of the low efficacy of the extracranial procedure and popularized the intracranial section of the glossopharyngeal nerve (Dandy 1927). Table 11.2 includes the series of patients who underwent rhizotomy of the IXth and/or Xth cranial nerves.