Neuroendoscopy-Assisted Minimal Invasive Management of Chiari 1 Malformation

Abstract

Objective  The aim this study is to present the results of the minimal invasive neuroendoscopic-assisted system application as an alternative to traditional surgery in patients with Chiari malformation type 1 (CM type 1) with/without syringomyelia.

Design, Setting, and Participants  In the study, data of 22 symptomatic patients were prospectively collected. Before and after the operation, patient characteristics, computed tomography, magnetic resonance imaging (MRI), cerebrospinal fluid (CSF) flow dynamics MRI, and outcome scales scores were recorded. Foramen magnum decompression and C1 total laminectomy were performed. The fibrous band at the craniocervical junction was opened and a durotomy was performed. In patients with a syrinx, the pre- and postoperative axial and sagittal lengths of the syrinx were measured and compared.

Results  The mean age of the patients was 32 ± 5 years. There were eight male patients. Ten patients had syrinx. The mean visual analog scale (VAS) score before and after surgery was 8 ± 1.06 and 2.18 ± 1.13, respectively. When evaluated according to the Chicago Chiari Outcome Scale, there was improvement in 20 patients, while there was no change in 2 patients. Syrinx resolved completely in 3 of 10 (13.6%) patients with syringomyelia, and the syrinx volume decreased in 3 patients (13.6%). In 4 of 10 (18.1%) patients, there was no significant change in the syrinx volume. The average operation time was 105 minutes (80–150 minutes). The average blood loss was 40 mL (20–110 mL).

Conclusion  Although the study was limited due to the small number of patients with a short follow-up, endoscopic decompression was a safe and effective technique for surgery in CM type 1 patients.

Introduction

Cerebellar tonsillar herniation, also called Chiari malformation (CM), is divided into four types (I-IV): clinical and radiological. ¹ Among them, types I and II are the most common. Currently, the common surgical treatment of CM is suboccipital decompression including the foramen magnum and atlas laminectomy with removal of the extradural fibrous band. If necessary, other options include bony decompression with dural splitting, bony decompression with dural opening and with duraplasty, and bony decompression with intradural manipulation and duraplasty. ² Surgical interventions vary from patient to patient. The most common surgery is foramen magnum decompression with atlas laminectomy (FMD-AL). ^{3 4} Sometimes, intradural exploration, C2 laminectomy, excision of the tonsils, duraplasty, and shunt implantation for the syrinx may also be required. Postoperative complications of these surgeries including pseudomeningocele, meningitis, cerebrospinal fluid (CSF) leak, adhesion, cerebellar ptosis, and cervical instability are serious, life-threatening conditions. ⁵ Currently, surgeons are still searching for minimally invasive methods that will provide optimal benefits for patients. Neuroendoscopic-assisted FMD-AL was investigated in selected patients for this purpose. Management of Chiari I malformations with fewer postoperative complications and recurrence continues to pose great challenges for surgeons. ^{6 7} The authors also investigated the postoperative complication rates of neuroendoscopic-assisted FMD-AL. In this study, the pre- and postoperative clinical findings, radiological features, perioperative values, developing complications, and postoperative outcomes in patients who underwent neuroendoscopic-assisted FMD-AL were compiled.

Patients and Methods

During the study period (between 2020 and 2023), 31 Chiari type 1 patients who were surgically diagnosed and deemed suitable for neuroendoscopic-assisted FMD-AL were admitted to the hospital. Nine of them declined endoscopic surgery and opted for open surgery instead. The remaining 22 patients underwent the endoscopic procedure at the Ankara City Hospital Department of Neurosurgery. The study was approved by the Institutional Review Board (TUEK E1–22–2615), and written informed consent was obtained from all the patients, including a discussion of the potential risks associated with the surgery. The patients were given the option to choose open surgery if they preferred it. The preoperative, perioperative, and postoperative data of the patients were collected. All the patients underwent a thorough physical examination and a detailed history, and their visual analog scale (VAS), Karnofsky performance scale (KPS), and modified Japanese Orthopaedic Association (mJOA) scores were recorded pre- and postoperatively. The Chicago Chiari Outcome Scale (CCOS) was used for follow-up. ⁸ Computed tomography (CT), magnetic resonance imaging (MRI), and CSF flow studies were performed before and after surgery. The dimensions of the syrinx were measured, and the percentage decrease was calculated ⁹ ( Table 1 ).

Table 1. Patient demography, clinical status, radiological findings, and complications.

Patient	Age	Sex	Pre-op symptoms	Post-op outcome	Cerebellar tonsils descent below McRae line (mm)	Pre-op syrinx	Post-op syrinx volume	Outcome	Complications
1	38	M	H/A, neck/arm pain, paresis, paresthesia, dysphagia	Improved	10			Improved	Dural tear
2	52	F	H/A, neck/arm pain, paresis, paresthesia, cerebellar symptoms	Improved	16			Improved
3	30	F	H/A, neck/arm pain, ataxia	Improved	20	+	Decreased	Improved	Dural tear
4	42	M	H/A, neck/arm pain	Improved	8			Improved
5	39	F	H/A, neck/arm pain, ataxia	Improved	11	+	Decreased	Improved	Inadequate decompression
6	49	M	H/A, impaired hot/cold discrimination	Reduced pain, unchanged neurological findings	14			No change
7	36	F	H/A, paresthesia, impaired hot/cold discrimination, ataxia	Improved	15	+	Decreased	Improved	Pseudomeningocele
8	38	F	H/A, neck/arm pain, paresthesia	Improved	10			Improved
9	19	F	Neck/arm pain, ataxia, paresthesia	Resolved	13	+	Unchanged	Improved	Dural tear, inadequate decompression
10	48	M	H/A, paresthesia, paresis, impaired hot/cold discrimination	Improved	18			Improved
11	56	M	Neck/arm pain, sleep apnea	Improved	12	+	Unchanged	Improved
12	22	F	H/A, neck/arm pain, cerebellar symptoms	Improved	13			Improved
13	24	F	H/A, neck/arm pain, dysphagia	Improved	19			Improved
14	26	M	H/A, paresis, paresthesia, ataxia, CN dysfunction	Improved	7	+	Decreased	Improved	Dural tear (changed to open technique during the operation)
15	32	F	H/A, neck/arm pain, paresthesia	Reduced pain, unchanged neurological findings	9	+	Decreased	No change
16	44	F	H/A, neck/arm pain, CN dysfunction	Improved	13			Improved
17	20	F	H/A, neck/arm pain, paresthesia, ataxia	Improved	16			Improved
18	21	M	H/A, neck/arm pain, paresis	Reduced pain, unchanged neurological findings	12			Worsened	Inadequate decompression, reoperated
19	27	M	H/A, paresis, ataxia, sleep apnea	Improved	14	+	Decreased	Improved
20	51	F	H/A, neck/arm pain, paresthesia, impaired hot/cold discrimination	Reduced pain, unchanged neurological findings	10	+	Unchanged	Improved
21	23	F	H/A, paresthesia, impaired hot/cold discrimination	Improved	11	+	Unchanged	Improved
22	25	F	H/A, paresthesia, ataxia	Improved	9			Improved

Abbreviations: CN, cranial nerve; F, female; H/A, headache; M, male.

Patients with atlantoaxial instability, basilar invagination, tethered cord, and tethered cord–related myelomeningocele/hydrocephalus, craniocervical tumors, a history of previous cervical surgery, and patients with high intracranial pressure were excluded from the study. The patients included had CM type 1 with/without syringomyelia for whom FMD-AL decompression was sufficient. The duration of operation, blood loss, results, and complications were recorded.

In addition to preoperative MRI, control MRI was performed at 3, 6, 12, and 24 months postoperatively, and phase contrast MRI was performed for all patients for CSF flow assessment. The postoperative MR images were reviewed and compared with the preoperative ones, and the volume reduction of the syrinx was calculated. If the syrinx was not seen in the postoperative MRI study, the syrinx was considered resolved. The syrinx volume was roughly calculated using the axial, coronal, and sagittal diameter of the syrinx (the A × B × C/2 method). If the syrinx volume was reduced by more than 20% on postoperative MRI, the syrinx was considered to have reduced. ⁹ This classification was based on the most recent postoperative MRI ( Fig. 1 ).

Fig. 1.

( A ) Preoperative magnetic resonance imaging (MRI) showing Arnold–Chiari malformation with syringomyelia. ( B ) Postoperative MRI scans of the same patient showing reduction of the syrinx.

CCOS was used to evaluate the postoperative outcomes of the patients in four categories of pain, painless symptoms, functionality, and complications, specific for the follow-up of Chiari patients ( Table 2 ). ⁸

Table 2. Chicago Chiari outcome scale ⁸ .

Chicago Chiari Outcome Scale
Pain	Nonpain	Functionality	Complications	Total score
1: worse	1: worse	1: unable to attend	1: persistent complication, poorly controlled	4: incapacitated outcome
2: unchanged and refractory to medication	2: unchanged or improved but impaired	2: moderate impairment (<50% attendance)	2: persistent complication, well controlled	8: impaired outcome
3: improved or controlled with medication	3: improved and unimpaired	3: mild impairment (>50% attendance)	3: transient complication	12: functional outcome
4: resolved	4: resolved	4: fully functional	4: uncomplicated course	16: excellent outcome

Statistical Analysis

Data analysis was conducted using IBM SPSS 25.0 (IBM Corp., Armonk, New York, United States) and MedCalc 15.8 (MedCalc Software bvba, Ostend, Belgium) statistical software packages. Descriptive statistical methods, including frequency, percentage, mean, standard deviation, median, minimum–maximum, etc., were employed. Chi-squared tests (Pearson's chi-squared test, Yates's corrected chi-squared test, Fisher's exact test) were utilized to compare the qualitative data. The Smirnov test, skewness-kurtosis, and graphical methods (histogram, Q-Q plot, stem and leaf, boxplot) were evaluated. For the comparison of normally distributed quantitative data between groups, independent sample's t -test was employed, while the Mann–Whitney U test was used for non-normally distributed data. The relationship between variables was assessed using Spearman's rho correlation test. The statistical significance level was set at p  = 0.05.

Operative Technique

The patients were placed in the prone position under general anesthesia, with the head in a neutral position and limited neck flexion not exceeding 30 degrees to ensure that the foramen magnum and the lamina of the atlas were almost parallel to the ground for better placement of the neuroendoscope. Care was taken not to obstruct the patient's intubation tube and cerebral venous return during neck flexion. Prior to making the incision, the posterior tubercle of the atlas was located using fluoroscopy ( Fig. 2 ). In patients with preoperative CT showing C1 posterior tuberculum deviation, the entry point and angle were adjusted accordingly. The single skin incision was made starting 0.5 cm inferior to the posterior tuberculum of the atlas and extending 1.5 to 2 cm toward the superior midline intersection ( Fig. 2 ). All the patients were operated on by a single surgeon (GG).

Fig. 2.

( A ) Preoperative detection of the posterior tubercle of the atlas with fluoroscopy. ( B ) Image after placement of dilators. ( C ) Preoperative incision image. ( D ) Perioperative incision image. ( E ) Postoperative incision scar.

After the incision was made, the first dilator was used to target the posterior tuberculum of the atlas as the entry point, followed by the placement of the other dilators in sequence to obtain the surgical field at the craniocervical junction ( Fig. 2 ). At certain stages of the operation, decompression was performed by angling the dilator.

The surgical technique used in all patients was FMD-AL, which involved suboccipital decompression. The incision was made 2.5 to 3 cm superiorly from the foramen magnum and 1.5 to 2 cm lateral from the midline, covering an area between 9 and 12 cm ² ( Figs. 3 and 4 ). In one patient, the preoperative CSF flow studies revealed decreased flow between C1 and C2, and this patient also had a syrinx at the same level. Therefore, the patient underwent a superior C2 hemilaminectomy, which did not require a longer incision and was performed by angling the dilator.

Fig. 3.

( A,B ) Postoperative computed tomography sagittal reconstructed images and ( C,D ) axial sections showing bony decompression.

Fig. 4.

Postoperative three-dimensional (3D) computed tomography section.

A neuroendoscopic set was used (Karl Storz in Tuttlingen, Germany). Fixation arms were utilized to fix the neuroendoscope and water aspiration systems. These systems not only facilitated the intraoperative cleaning of the surgical field but also served to clean the endoscope lens. After making the skin incision, it was deepened using a scalpel, electrocautery, and bipolar electrocautery without deviating from the midline. The incision was then advanced with blunt dissection and neuroendoscope dilators until the foramen magnum and C1 lamina were visible. Using a blunt dissection technique, an area of 12 to 15 cm ² was exposed, with borders 3 cm above the foramen magnum, C2 lamina superior border, and 1.5 to 2 cm from both sides of the midline ( Fig. 5 ).

Fig. 5.

Endoscopic images showing ( A ) the foramen magnum (FM; arrow ), ( B ) dissection of the foramen magnum ( arrow ) and ( C,D ) fibrous bundle removal of the suboccipital bone ( arrows show the FM and fibrous bundle), ( E,F ) atlas laminectomy ( arrow shows C1).

The neuroendoscope exposes a small site, making it crucial to maintain anatomical orientation and avoid losing the midline. The suboccipital muscle raphe provides one landmark for this purpose, and the subocciput beneath it is marked or diathermied. Failure to maintain midline alignment can cause confusion and disorientation, as well as increase blood loss. A small hemorrhage can obstruct vision and make the operation more challenging in a limited area. Loss of midline alignment may result in insufficient decompression, as well as increase the risk of vertebral artery injury. When dissecting from both sides, the surgeon must reestablish the midline position, which may vary between the edges, or use a 30-degree endoscope to widen the visual field.

After exposing the occipital bone and the C1 posterior arch, C1 laminectomy was performed using a high-speed drill and Kerrison rongeurs. Then, suboccipital craniectomy was performed using a high-speed drill and Kerrison rongeurs. The reason for removing the C1 arch first was to facilitate the dilator manipulation necessary for suboccipital decompression. Otherwise, the C1 posterior tubercle made it difficult to approach the occipital bone. In this way, the musculus rectus capitis posterior major was not cut and was not damaged by retraction. Usually, at this stage, we can clearly see that the bony structures are causing dural matter entrapment, for example, the posterior border of the foramen magnum or obvious C1 posterior arch compression. Particular attention is recommended when approaching the occipital condyle to spot the vertebral artery. Due to the rich paravertebral venous plexus, venous bleeding occasionally occurred, and adequate hemostasis could be achieved through gentle compression, using a bipolar electrocautery, using a hemostatic matrix, or suture ligation. In our study, there were intraoperative venous plexus breaches that occurred in five patients. All bleeding was stopped with appropriate methods. There was no air embolism in any case.

The duratomy is made in a “ Y ” shape to avoid the occipital sinus. ¹⁰ The durotomy is typically started caudally at the “stem” of the “ Y ” at the C1 level, and then developed cranially into each “arm” of the “ Y ” laterally over each cerebellar hemisphere as the “ Y ” is completed. Parallel longitudinal incisions were made on the safe dura area to allow decompression for dural splitting. Only the superficial layer of the dura was cut; the deep layer and arachnoid mater were not damaged. A full-thickness incision should be avoided.

Results

Patients

Twenty-two adults underwent surgery for CM type 1 with or without syringomyelia during the studied time period. The mean age at the time of surgery was 32 ± 5 years, and eight (36.3%) patients were males. Ten (45.4%) patients had syringomyelia.

Symptoms were examined, and 20 patients (90.9%) reported headaches that were aggravated by exercise, straining, coughing, sneezing, laughing, and bending. Additionally, 15 patients (68.1%) experienced neck and arm pain. Six patients (27.2%) had muscle weakness, 13 patients (59%) had sensory impairment, 5 patients (22.7%) had impaired hot–cold discrimination, and 9 patients (40.9%) had gait instability. Six patients (27.2%) also had brain stem or cranial nerve symptoms, such as fullness in the ears, sleep apnea, and swallowing problems.

Syringomyelia involved the cervical spinal cord in eight cases (36.3%) and the thoracic spinal cord in five cases (22.7%). Three patients (13.6%) had syringomyelia in both the cervical and thoracic spinal cords.

Dural tears developed during the operation in four patients, and CSF flow was observed. In one of these patients, since dural tear repair could not be performed, the endoscopic FMD-AL method was changed to the traditional open surgical technique during the operation. Duraplasty was performed with primary suture in one patient, while in the other two patients, duraplasty was performed using muscle and fibrin tissue glue. In three patients, adequate decompression was not ensured on postoperative control CT. As the symptoms worsened in one patient, the patient was reoperated using the traditional open surgical technique. However, two patients were followed up and not operated, as their symptoms improved.

During the study period (between 2020 and 2023), the patients' preoperative mean VAS values were 8 ± 1.06. In the postoperative 12th-month controls, the mean VAS values were 2.18 ± 1.13. The preoperative mean mJOA values were 14.5 ± 1.6, and the postoperative mean mJOA was 15.6 ± 0.9. The preoperative mean KPS values were 81.8 ± 10.52, and the postoperative mean KPS was 86.81 ± 6.46 ( Table 3 ). When CCOS was evaluated, 19 patients (86.3%) showed an improved outcome, 2 patients (9.1%) had an unchanged outcome, and 1 patient had a worse outcome, requiring reoperation with the open surgical technique. Syrinx volume decreased in six patients (27.2%), while four patients (18.1%) showed no significant change in syrinx volume. None of the patients had an increase in the syrinx volume.

Table 3. Comparison of pre- and postoperative pain and outcome scores.

	Preoperative	Postoperative	p a
KPS (mean ± SD)	81.8 ± 10.5	86.8 ± 6.5	0.005
VAS (mean ± SD)	8.0 ± 1.1	2.2 ± 1.1	<0.001
JOA (mean ± SD)	14.5 ± 1.6	15.6 ± 1.0	0.002

Abbreviations: KPS, Karnofsky performance scale; mJOA, modified Japanese Orthopaedic Associations; SD, standard deviation; VAS, visual analog scale.

^a Paired sample t -test.

CSF flow dynamic MRI was performed in all the patients, which showed a decrease (ranging from total obstruction to slight slowdown) in CSF flow due to subarachnoid space compression at the craniocervical junction. After surgery, restoration of the subarachnoid CSF flow was observed in the postoperative CSF flow dynamic MRI in all patients. One patient developed a pseudomeningocele at the 1-year follow-up, although the postoperative follow-ups at 3 and 6 months were normal.

The average operation time was 105 minutes (range: 80–150 minutes) with an average blood loss of 40 mL (range: 20–110 mL). No postoperative complications were observed in the patients, who were hospitalized for 1 day and then discharged. Although severe headache decreased in for patients, occasional headache (1–2 headache episodes per month) persisted. Neck pain resolved in all the other patients. Mild sensory impairment (according to mJOA scale) persisted in six patients. Lumbar drain insertion, meningitis, and wound infection did not occur in any of the patients. The mean postoperative follow-up was 13.3 months (range: 6–24 months).

Discussion

Patients with CM type 1 with or without syringomyelia undergo different procedures in different hospitals. ¹¹ There are different approaches regarding whether the dura mater should be opened or not. ¹² Traditional open surgery can cause maximum surgical trauma with increased injury to the posterior muscles and ligaments. Excessive damage to the posterior muscles and ligaments may increase the craniocervical junction instability that is usually present in these patients. ¹³ However, there is no proven literature on this subject. As a result of traditional surgery, it reduces the quality of life of patients by causing head and neck pain for many years. In the study by Mugge et al, headache persisted in 53% and neck pain in 50% of patients operated on with traditional open surgery in the postoperative period. ¹⁴ Also, postoperative recovery time is long and collar support is usually required for up to 3 months. ⁶ It is thought that the main predisposing factor in patients with CM type 1 with or without syringomyelia is the compression of the bone structures and fibrous band structure at the craniocervical junction, or the disruption of CSF circulation due to herniation of the cerebellar tonsils. ⁶ The main purpose of the traditional open surgery is to eliminate the cause of compression, to widen the posterior area of the craniocervical junction, and thus to restore the CSF flow.

Neuroendoscopy is increasingly being used in spinal surgery, cranial surgery, and skull base surgery. ^{15 16} The reason why we chose the endoscopic method instead of the microscopic tubular method in our study is that it is necessary to frequently change the angles with dilators. The field of view of the microscope remains limited at varying angles, and the surgeon's position and comfort during the case are poor. The most important of these reasons is the inability to fully control the operation area in the craniocervical region, where vascular and neuroanatomically important structures are complex. Additionally, the neuroendoscopy-assisted system enables free-handed application and manual maneuvering at different time points of the procedure. It is also much cheaper than the full endoscopic system.

Our results of neuroendoscopy-assisted FMD-AL surgery in patients with CM type 1 with or without syringomyelia were quite good, and these results are in agreement with those of authors with similar series. ^{2 6} Contrary to other publications, we chose the posterior tubercle of the atlas, not the foramen magnum, as the entry point because a bony prominence was targeted, and our point of entry could be pinpointed by fluoroscopy. Thus, cord injury, C1 and C2 root injury, and vascular injury were completely avoided. In neuroendoscopy-assisted FMD-AL, decompression is possible within the limits used in conventional microsurgery. Also, if dura splitting is necessary, it can be safely done with the endoscopic method.

Caffo et al observed in their study that it is sufficient to resect the median one-third of the C1 posterior arc. In our study, we kept the decompression wider by seeing and localizing the venous plexus well. ¹⁷

There is still debate about whether it is necessary to open the dura mater, which depends on the examination for bundle syndrome. ³ If a thickening is present on the dura mater surface, creating a bandlike bundle, a small longitudinal parallel superficial incision can be made. In our study, we left the arachnoid intact using the Logue and Edwards technique. ¹⁸ This technique prevents CSF leakage, pseudomeningocele formation, subarachnoid hemorrhage, and postoperative malaise caused by exposure of the subarachnoid space to surrounding tissues. ^{10 18} However, pseudomeningocele developed in one patient. If CSF flow has increased sufficiently, a dural incision may not be required. Intraoperative Doppler ultrasound can be used to evaluate CSF flow after bone compression is removed. However, in this study, in addition to FMD-AL, the fibrous band was excised, and an incision was made in the superficial layer of the dura mater in all patients. We believe that measuring the CSF flow by intraoperative Doppler ultrasonography may be misleading. Therefore, dural splitting was performed in all 22 patients in addition to FMD-AL.

Patients with suspected syringomyelia on cervical MRI underwent thoracic MRI. Pre- and postoperative CSF flow MRI was evaluated. In the postoperative period, controls were made at 3, 6, 12, and 24 months.

The changes in the patients' living standards were evaluated and compared by considering the mJOA, CCOS, and KPS scores, and changes in pain were evaluated and compared using the VAS. In conclusion, the pre- and postoperative imaging, mJOA scores, CCOS scores, KPS scores, VAS scores, and clinical examinations of the patients were compared. Although there was an improvement in the KPS, CCOS, and mJOA scores, no statistically significant change was observed ( p  > 0.05). Due to limited muscle dissection, neck pain resolved in all patients, and the rate of analgesic use was lower than that in patients who underwent traditional surgery. Additionally, the decrease in VAS scores also supports this theory. The most significant of these comparisons is the results of the CSF flow MRI since the purpose of the surgery is to restore CSF flow.

Although the complication rate was low in our series, there are still risks such as pseudomeningocele, hydrocephalus, infection, instability, and inadequate decompression. Additionally, even though the mean blood loss was low in our study, significant bleeding from the bone or venous plexus may occur. To minimize bone bleeding, it may be more advantageous to use a high-speed drill. Bipolar coagulation and hemostatic matrix can be used for other bleedings.

The limitations of our study are the small number of patients and the short follow-up period. To better understand the results of this technique, a larger patient series that includes open, endoscopic, and microscopic techniques should be conducted and the results compared.

Conclusion

The neuroendoscopic FMD-AL approach has various advantages, such as reduced blood loss, short operation time, short hospital stay, fast recovery time, and small skin incision. It provides sufficient decompression and maintains craniocervical stability, causing minimal damage to the posterior muscles and ligaments. As a minimally invasive technique, it has the potential to evolve and develop further. However, the learning process for this technique is relatively long. The learning process for the minimally invasive FMD-AL technique remains intricate and multifaceted, making it difficult to determine a precise learning curve. Nevertheless, as surgeons gain expertise and familiarity, improvements in operative times and a smoother transition to open procedures are observed. It is important, however, to consider the limitations inherent in this approach, including the initial absence of tactile feedback, the constraints of a confined surgical space, and the reliance on two-dimensional imaging provided by the endoscope. Nonetheless, endoscopy appears to be a safe and feasible option, leading to favorable outcomes in patients with CM type 1.