A single-center retrospective study was conducted in 25 patients who were enrolled consecutively. All patients were evaluated in terms of the pre- and post-operative clinical status. We collected medical data related to demographics, preoperative symptoms, the offending vessel, operation time, surgical outcomes, and postoperative complications. Preoperative electrophysiological examinations were conducted to confirm the presence of abnormalities in the blink reflex and facial nerve conduction. Additionally, prior to surgery, magnetic resonance imaging (MRI) was performed using the volume isotropic turbo spin echo acquisition (VISTA) sequence with a focus on the brain stem to preoperatively identify the offending vessel.

All operations were performed by a single neurosurgeon with the fully E-MVD technique without any assistance of a microscope. Surgeries were performed under general anesthesia, with the patient in a supine position and ipsilateral head rotation at 90 degrees. A 3-pin head fixator was used for head stabilization. For surgery, a 4 mm 0-degree and 30-degree rigid endoscope were utilized (IMAGE1 S^TM RUBINA^TM; Karl Storz SE & Co. KG, Tuttlingen, Germany), and a manually operated endoscopic fixation device was used to secure the endoscope to the bed.

The surgical procedure was performed in a similar manner to that of microscopic MVD, but with some modifications. Instead of the typical retromastoid approach, a shorter linear skin incision was made, followed by a burr hole craniectomy to expose the sigmoid sinus and transverse sinus (Fig. 1). More than one third of the sigmoid sinus was exposed to ensure a clear view and angle for entry of the endoscope. A C-shaped dural incision was made along the boundary of the sinus, and cerebrospinal fluid (CSF) was drained with slow rate to gently separate the cerebellum from the dura.

For effective CSF drainage, several conditions were necessary. Firstly, the patient’s position must be adjusted to a classic 15-degree vertex down. Secondly, the craniectomy must be located above the low cranial nerves (CNs), making adjustments to allow a vertical view of the cerebellomedullary cistern by drilling the skull’s inner table. If CSF drainage was not conducted in a timely manner, it might be necessary to retract the cerebellum to open the cistern and secure an adequate view, which could be challenging. After ensuring enough CSF drainage and relaxing the cerebellum, the endoscope was carefully inserted. Once the endoscope was mounted on the endoscope holder, the arachnoid membrane was dissected to carefully expose and identify the CNs, especially the REZ of the facial nerve. It was crucial to completely dissect the arachnoid membrane around the low CNs to allow for a tension-free release of the low CNs and to prevent unnecessary damage. When instruments such as suction were introduced alongside the endoscope, they should enter ahead of the endoscope. Especially before becoming accustomed to the endoscope, introducing instruments after the endoscope increased the risk of causing damage to the cerebellum. Using a cottonoid patty at the access corridor (lateral to the cerebellar cortex), the instruments were safely inserted from the dura mater to the operative field, inspecting them with a macroscopic view. The 30-degree endoscope provided a more accurate anatomical position and panoramic view compared to the 0-degree endoscope, but caution was exercised due to spatial distortion. When entering with the endoscope, it would be safe to approach parallel to the petrous bone. Additionally, creating an angle that feels like looking down from above would be safe rather than directly facing the objectives.

When the offending vessel was identified, Teflon (TFE Polymer Pledget; Ethicon, Inc., Raritan, NJ, USA) was inserted using forceps and secured with fibrin glue. Microscopic MVD involves performing surgery while looking from the front. But, E-MVD is performed by looking down from above, and it facilitated that the extent of retraction is naturally decreased with direct visualization of the REZ. Moreover, using suction and curved instruments during endoscopic surgery is thought to offer slight improvements compared to microscopic surgery. The dura was sutured tightly to ensure it was watertight, and a burr hole-type miniplate was used to cover the bony defect. The operating time was defined as from the start of skin incision to the end of skin closure.

IONM included triggered CN electromyography, brainstem auditory evoked potentials (BAEPs), and disappearance of the lateral spread response (LSR). IONM was performed by using the NIM-Eclipse (Medtronic, Minneapolis, MN, USA) system.

We assessed the outcomes by evaluating patients’ clinical status immediately after surgery and at their last follow-up. We conducted a separate analysis for acute complications occurring before discharge and delayed complications observed during the outpatient follow-up period. We described the types of complications, such as facial paralysis and hearing loss, individually. Additionally, we examined the findings from the last follow-up examination.

Patient demographics are summarized in Table 1. There were nine men (36.0%) and 16 women (64.0%). The median age was 55 years (range, 35–72). Eleven cases showed right-sided disease (44.0%) and 14 cases showed left-sided disease (56.0%). Among these 25 patients, 10 patients (40.0%) had pre-operative facial palsy on the ipsilateral side, which was graded as House-Brackmann grade 2. The asymmetry of mouth corner, deep nasolabial fold caused by long-standing spasm of affected muscles was included as facial palsy.

The most common offending artery was the AICA in 15 cases (60.0%), followed by the PICA in eight cases (32.0%), the vertebral artery (VA) in one case (4.0%), and a tandem offender involving the AICA and VA in one case (4.0%).

Disappearance of the LSR was confirmed in all patients (100.0%). The median operating time was 105 minutes with a range from 65 to 230 minutes.

All patients were symptom free at the immediate post-operative neurologic examination. The follow-up period ranged from 3 to 49 months, with a median follow-up duration of 24 months. All patients remained symptom free at the last follow-up examination. However, in two patients (7.4%), there was a recurrence of symptoms 2 months after surgery. These individuals were successfully managed with carbamazepine, and at their most recent follow-up, they achieved stable symptom control without the need for medication.

In one patient, post-operative hearing impairment was observed, and high-frequency hearing loss was confirmed on audiometric testing. Another patient experienced CSF leakage after surgery, which was managed with a lumbar drain for 5 days, and the patient was discharged without any further complications. Delayed facial palsy occurred in a total of eight patients (32.0%) approximately 2 months after surgery. However, all these patients showed improvement following oral corticosteroid treatment, and at their last follow-up, they exhibited normal facial function. No other complications, such as intracranial hemorrhage, cerebral infarction, or CN injuries, were observed.

A 64-year-old man was admitted with a chief complaint of right-sided HFS. He had been experiencing HFS for 3 years, with worsening of symptoms over the past 1 month. Despite receiving oral medications and BOTOX^® (clostridium botulinum toxin type A; AbbVie, Inc., North Chicago, IL, USA) treatment, his symptoms did not improve. Consequently, we decided to proceed with surgical MVD due to the refractory nature of his condition to medical treatment. MRI with the VISTA sequence revealed that the right AICA and right VA were compressing the right facial nerve at the REZ (Fig. 2A). Dolichoectasia of the VA was also observed, and the VA diameter was approximately 8 mm. Preoperative electromyogram motor nerve conduction velocity (EMG-NCV) testing of the facial nerve was conducted to establish baseline data for comparison between before and during the operation. Prior to surgery, baseline data indicated the presence of HFS. Pure tone audiometry and speech audiometry were performed to assess hearing impairments, but no specific findings were noted.

After standard linear skin incision and burr hole craniectomy, we opened the cerebellomedullary cistern and cerebellar relaxation was achieved with CSF drainage. Following arachnoid dissection, we identified the CN VII & VIII complex and lower CNs. Simultaneously, we observed that the dolichoectatic right VA and right AICA were compressing CN 7 at the REZ (Fig. 2B). We attempted gentle mobilization of the VA for better visualization, and then interposed Teflon (TFE Polymer Pledget; Ethicon, Inc.) between the nerve and vessel. A total of four pieces of Teflon were embedded (Fig. 2C), resulting in the disappearance of LSR. However, the LSR recurred when the retracted VA returned to its original position. Therefore, we prepared for transposition of the VA. The VA diameter was measured, and a sling (Bard^® PTFE Felt Pledget; Bard Peripheral Vascular, Inc., Tempe, AZ, USA) was designed with dimensions of approximately 4 mm in width and 30 mm in length. The VA was encircled by the sling without interference by any small vessels during the passage. We then mobilized the sling towards the posterior petrous bone and sutured it to the dura of the petrous bone, transposing the VA away from its original position (Fig. 2D). After one suture, the transposed VA was abutting the lower CNs, prompting the creation of another suture in a more superolateral direction (Fig. 2E). This resulted in the release of compression on the lower CNs, and there was no obstruction at the REZ of CN 7 anymore. Ultimately, the LSR permanently disappeared after the transposition of the dolichoectatic VA using the sling technique. Throughout the entire operation, BAEP monitoring was conducted, and no significant findings were observed (Supplementary Video 1).