Neurosurg Focus  Volume 38 • April 2015

neurosurgical

focus Neurosurg Focus 38 (4):E10, 2015

The surgical management of lesions ventral to the neuraxis at the craniovertebral junction (CVJ) and upper cervical spine is challenging because of the

limited access to this restricted anatomical and eloquent compartment. The extent of the surgical approach and its associated morbidity rates differ significantly depending on the precise localization, extent, and type of the lesion.

The majority of craniovertebral lesions are located in the dorsal or dorsolateral aspect of this region, which can be accessed easily by traditional posterior approaches.2,8,9,11 Access to lesions located ventral or ventrolateral to the neuraxis at the CVJ level is much more complicated and

usually involves extensive muscular mobilization and bone removal to get adequate exposure without neural retrac-tion. Many variants of ventral and dorsal approaches to the ventral aspect of the CVJ for accessing a variety of pathol-ogies, such as meningiomas, chordomas, neurinomas, me-tastasis, vascular lesions, and intramedullary lesions, have been reported. However, most of these reports focused on large lesions with corresponding broad approaches and significant exposures.8,9,15 Apart from time-consuming surgical preparation, these approaches are associated with relevant perioperative morbidities including nuchal pain, restricted mobility, and CSF leakage and with, eventually,

AbbreviAtioNs C-0 = occiput; CVJ = craniovertebral junction. submitted November 29, 2014.  Accepted February 3, 2015.iNclude wheN citiNg DOI: 10.3171/2015.2.FOCUS14799.disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. 

Minimally invasive approach for small ventrally located intradural lesions of the craniovertebral junctionsven o. eicker, md, Klaus christian mende, md, lasse dührsen, md, and Nils ole schmidt, md

Department of Neurosurgery, University Medical Center, Hamburg-Eppendorf, Germany

object The surgical management of lesions ventral to the neuraxis at the level of the craniovertebral junction (CVJ) and upper cervical spine is challenging. Here, the authors describe a minimally invasive dorsal approach for small ven-trally located intradural lesions at the CVJ as an alternative for the more extensive classic transoral approach or variants of suboccipital approaches.methods Between 2012 and 2014, 6 symptomatic patients with a small lesion of the ventral aspect at the CVJ level were treated using a minimally invasive dorsal approach at the University Medical Center in Hamburg-Eppendorf, Germany. The anatomical distance between the posterior atlantooccipital membrane and the posterior atlantoaxial liga-ment, as determined by CT images, was assessed in the treated patients and in 100 untreated persons.results The authors treated 6 patients (mean age 54.7 years) who had a clinical presentation of mild neurological symptoms that disappeared after resection. Minimally invasive surgical dorsal access was achieved by using tubular systems and using the natural space between the occiput (C-0) and C-1, and in 1 case between C-1 and C-2, without having to remove bony structures. The postoperative course in each of the 6 patients was uneventful. The neuropatho-logical findings confirmed a meningotheliomatous meningioma (WHO Grade I) in 5 cases and an extramedullary cavern-ous hemangioma in 1 case. MRI confirmed complete resection of all the lesions. The atlantooccipital distances ranged from 3 to 17 mm (mean 8.98 mm) in the supine neutral position, and the atlantoaxial distances ranged from 5 to 17 mm (mean 10.56 mm). There were no significant differences between women and men (atlantooccipital p = 0.14; atlantoaxial p = 0.72).coNclusioNs The results of this study demonstrate that the minimally invasive dorsal approach using the space be-tween C-0 and C-1 or C-1 and C-2 provides direct and sufficient exposure for the safe surgical resection of small ventral-ly located intradural lesions at the CVJ level while reducing the necessity for musculoskeletal preparation to a minimum.http://thejns.org/doi/abs/10.3171/2015.2.FOCUS14799Key words cervical spine; craniocervical junction; spinal tumor; minimally invasive; meningioma

1©AANS, 2015

Unauthenticated | Downloaded 08/03/20 10:45 PM UTC

s. o. eicker et al.

a need for additional stabilization in cases of instability. In contrast, minimally invasive approaches to small ventral lesions at the upper cervical spine and CVJ have rarely been reported.4,5,7

Here, we report a surgical technique for accessing small intradural lesions located ventral to the neuraxis at the CVJ. By using a minimally invasive dorsal approach between the occiput (C-0) and C-1 or C-1 and C-2, com-plete removal of ventrolateral lesions can be achieved while avoiding bone removal and extensive muscular mo-bilization.

methodsSix symptomatic patients with small ventrally located

intradural lesions at the CVJ level who were eligible for a minimally invasive surgical dorsal approach were in-cluded in this study. Tumors with a maximum diameter of > 25 mm, especially in the craniocaudal direction, and those that displayed signs of vascular or bone infiltration were not eligible for this approach and were excluded from this analysis. Each patient was treated in our institution between 2012 and 2014 and gave informed written con-sent. In accordance with local and institutional laws and data-protection regulations, no approval by the local eth-ics committee was necessary for this study. Patient data, including demographics, clinical presentation, imaging, surgical treatment, histology, and postoperative results, were analyzed retrospectively. In each case, MRI revealed lesions of the ventral aspect of the CVJ, which is generally defined as the region between the lower third of the clivus and the C-2 vertebral arch.11 Each lesion was suspected to be a meningioma with the exception of 1 case of an unclear cystic lesion. All the patients were evaluated at regular follow-up examinations in our neurosurgical out-patient clinic.

The distances between C-0 and C-1 (posterior atlan-tooccipital membrane) and between C-1 and C-2 (poste-rior atlantoaxial ligament) were analyzed using CT scans of 100 randomly selected patients from the radiological archive at our institution. Only CT scans of patients in a neutral supine position and those without a lesion of the musculoskeletal system at the CVJ level were included for analysis. The smallest paramedian distance was measured by using picture-archiving and communication system software (PACS) (GE Healthcare).

surgical techniqueAfter oral intubation and placement of the electrophysi-

ological monitoring electrodes, the patient’s head was fixed in a Mayfield skull clamp and the patient was turned to a prone position with slight inclination (Video 1).

video 1. In this video, we demonstrate the minimally invasive dorsal approach via the posterior atlantooccipital membrane for the surgical removal of a ventrally located intradural meningioma (maximum diameter 22 mm) at the CVJ level in a 77-year-old male patient. Copyright Nils Ole Schmidt. Published with permis-sion. Click here to view with Media Player. Click here to view with Quicktime.

The Mayfield skull clamp was connected to the operating table. The access was selected under fluoroscopic bipla-

nar guidance in the orthograde lateral and posterior an-terior views in cases of a transmuscular approach using a tubular (Spotlight; DePuy Spine) or a standard Caspar retractor (Spine Classics MLD; Aesculap) system. After selecting the point of entry, a 3- to 4-cm skin incision was made (3–5 cm lateral to the spinous process of C-1). The muscle fascia was sharply perforated, and the muscle was dissected with 2 dissectors until the C-1 arch was reached to insert the Caspar retractor or the dilator was inserted until it reached the depth of the C-1 arch. Then, the dila-tor was advanced a few degrees toward the posterior at-lantooccipital membrane (5 cases) or toward the posterior atlantoaxial ligament (1 case). After this muscle dilata-tive approach was completed using a second and a third sequential dilator, a tubular retractor with a diameter of 14–18 mm and a length of 50–70 mm was inserted (Fig. 1 left). Fixation was achieved by using a table-mounted flexible arm to secure the tubular retractor in place. The following steps were performed using a surgical micro-scope (NC 4; Carl Zeiss). The membrane or ligament was prepared, and a micro-Doppler catheter was inserted to identify or exclude the course of the vertebral artery within the surgical field. Then, the dura was opened by using a longitudinal paramedian incision and subsequent tack-up sutures for a better overview, which takes some time because of the limited space. The dorsolateral aspect of the spinal cord came into view. The denticulate liga-ment was uncovered (Fig. 2C) and cut to gain additional room while minimizing manipulation of the spinal cord. The layer between the spinal cord and the lesion was mi-crosurgically prepared while mobilizing the lesion in the field of view and reducing the tumor step by step. Unless it was already identified microscopically, we used the mi-cro-Doppler catheter during tumor preparation to locate and define the course of the vertebral artery. Figure 2D displays the ventral dura, which was coagulated to achieve a Simpson Grade II resection of the meningioma. After re-section of the lesion, the dura was closed water-tightly us-ing 6-0 polypropylene sutures; however, because this step of the surgery was somewhat challenging because of the narrow working canal, the dura was covered with medical sponge (TachoSil; Takeda). The tubular system was slowly removed with special attention paid to bleeds originating from the muscle tissue. There was no need for drains in any of the cases. The fascia was closed with 1 suture, and the wound was closed with 3 subcutaneous sutures.

resultsBetween 2011 and 2014, we treated 6 patients with

small intradural lesions (maximum diameter < 25 mm) located ventral to the neuraxis at the CVJ level. The pa-tient ages ranged from 41 to 64 years (mean 54.7 years); 2 women and 4 men were included. The clinical presentation in each case was mild neurological disturbance in terms of intermittent paresthesias in the lower extremities during inclination. No manifest focal neurological deficits were detected. Each patient underwent surgery in which a mini-mally invasive dorsal transmuscular approach with access to the intradural lesion via the space between C-0 and C-1 was used without a need for significant bony removal. In 1

Neurosurg Focus  Volume 38 • April 20152

Unauthenticated | Downloaded 08/03/20 10:45 PM UTC

minimally invasive dorsal approach to the cvj

case, the lesion was localized intradurally at the level of the dens axis basis, and therefore the space between C-1 and C-2 was used to access the meningioma. Throughout the procedures, somatosensory evoked potentials and motor evoked potentials remained unaltered. The postoperative course in each of the 6 patients was uneventful with no neurological deficits and no postoperative CSF leaks. The neuropathological findings confirmed meningothelioma-tous meningioma (WHO Grade I) in 5 cases and an extra-medullary cavernous hemangioma in 1 case. Postoperative MRI confirmed gross-total resection (Fig. 3A–D). During the follow-up time period (range 4–48 months; median 28 months), there were no recurrences and we detected no signs of clinical spinal instability. In each patient the pre-surgical intermittent paresthesia symptoms disappeared.

Because radiologically measured anatomical data on the distances between C-0 and C-1 and between C-1 and C-2 are lacking, we assessed these distances on the CT images of 100 randomly selected untreated individuals (mean age 52.4 years). The atlantooccipital distances ranged from 3 to 17 mm in the supine neutral position (mean 8.98 mm; Fig. 1 right). In the 6 treated patients, these distances ranged from 9 to 13 mm (mean 10.83 mm). The atlantoaxial dis-

Fig. 1. Illustration of the surgical access route and the space used between C-0 and C-1 or C-1 and C-2.  left: Measurements of the lowest distances between C-0 and C-1 and between C-1 and C-2 in the neutral position based on CT images of 100 untreated patients.  right: Displayed are the lowest, highest, and median C-0 to C-1 and C-1 to C-2 distances along with the upper (75%) and lower (25%) quartiles. Copyright Nils Ole Schmidt. Published with permission.

Fig. 2. Sagittal (A) and coronal (b) T1-weighted contrast-enhanced MRI studies of a small homogeneous enhancing lesion of the ventral aspect at the C0–C1 level. The images reveal the considerable subarachnoid space at the C0–C2 level.  c: After the dorsal dura is opened, the denticulate ligament is identified and resected to gain access to the tumor.  d: A Simpson Grade II resection of the meningioma is achieved by coagulation of the clearly visible dural tumor attachment.

Fig. 3. Preoperative axial (A) and sagittal (b) T1-weighted contrast-enhanced MRI studies of a ventrolaterally located intradural lesion at the C0–C1 level. Postoperative axial (c) and sagittal (d) T1-weighted contrast-enhanced MRI studies showing complete resection of the WHO Grade I meningioma without any removal of bony structures.

Neurosurg Focus  Volume 38 • April 2015 3

Unauthenticated | Downloaded 08/03/20 10:45 PM UTC

s. o. eicker et al.

tances in the untreated individuals ranged from 5 to 17 mm (mean 10.56 mm; Fig. 1 right) and from 7 to 14 mm in the 6 treated patients (mean 10.83 mm; in the 1 patient for whom the atlantoaxial approach was used, the distance was 14 mm). There were no significant radiologically documented anatomical differences between the women and the men in the group of 100 untreated individuals (atlantooccipital p = 0.14; atlantoaxial p = 0.72).

discussionMany variants of surgical techniques for accessing

ventrally located tumors at the CVJ have been discussed in the literature, and most of them require extensive mus-culoskeletal preparation to achieve minimal manipulation of neural tissue. Here, we demonstrate that minimally in-vasive dorsal access routes through the natural space be-tween C-0 and C-1 or C-1 and C-2 can be used for the removal of small ventrally located intradural lesions at the CVJ. This technique represents a natural adaptation of surgical access to the size of the lesion and prevents the need to use transoral or other extended dorsal approaches such as the far-lateral approach or classic variants of the suboccipital approach.

Several approaches for the treatment of lesions at the upper spine and CVJ levels have been described. The most direct approach to the ventral aspects of the CVJ, especially for extradural but also for intradural lesions, is the transoral route.3 The transoral route provides a direct and natural approach to the CVJ. However, the approach is sophisticated in terms of avoiding CSF leakage and sub-sequent infections and in the prevention of velopharyn-geal function impairment and loss of stability. Inadequate exposure of the lateral margins of the tumor and the lack of proximal control of the vertebral artery are other disad-vantages of this approach.12 Avoiding the transoral route, Banczerowski et al.1 used a conventional ventral approach with complete corpectomy and additional stabilization. This approach has been well described, is often performed for degenerative spinal pathology, C1–2 fractures, and ret-ropharyngeal lesions, and may provide the best overview of the complete ventral aspect of the spinal dura and col-umn. Complete corpectomy of the second vertebral body and removal of the ventral aspect of the first vertebral body to treat lesions in the ventral aspect of the spinal canal are not recommended, because fusion of at least 2 segments is required in any case. Approaching the upper cervical spine and the CVJ anterolaterally and laterally, George et al.6 and Yasuda et al.14 reported that the risk of lacerating the esophagus or trachea is minimized and the amount of bone loss is reduced in comparison with the results of a conventional anterior approach with corpec-tomy. However, the risk of accessory nerve or sympathetic chain injury is higher, and surgical control of the vertebral artery or venous plexus hemorrhage might be challenging.

Multiple variants of the suboccipital approach (dor-sal and dorsolateral) have been published. The far-lateral and extreme-lateral variants are very well-described ap-proaches that are clearly indicated for large lesions.8,9,15 Nevertheless, whether using a ventral or dorsal access route, all of these surgical approaches require extensive

musculoskeletal preparation, which is time-consuming and associated with the corresponding perioperative mor-bidity.

To avoid these extensive approaches for small ventral-ly located spinal lesions, the posterior approach through hemilaminectomy or laminectomy is the most commonly used approach in the upper cervical spine.7 Martin et al.10 described a posterolateral approach for vascular malfor-mations and tumors of the ventrolateral aspect of the spi-nal cord. They proposed a slight rotation of the spinal cord after multilevel division of the dentate ligament cranial and caudal to the lesion, which in our opinion is not pos-sible at the CVJ level because this is where the brainstem begins.

Watanabe et al.13 noticed the wide space between C-1 and C-2, and Zozulya et al.15 pointed out that at the C-1 level, the ratio of the squared spinal cord and dural sac cross-section areas is, on average, 1:3. Therefore, one-third is occupied by the spinal cord and two-thirds is occupied by subarachnoid space, which leaves considerable space for potential surgical access. We show here that the para-median spaces between C-0 and C-1 and between C-1 and C-2 are a mean of approximately 10 mm, which is in agreement with the Watanabe et al.13 report. With our approach, this space is extended as a result of the concord positioning during the operation and the slight inclination of the head (Video 1). Our radiological data of the ana-tomical distances between C-0 and C-1 and between C-1 and C-2 provide the basis for additional studies, because it is not clear at this point if there exists a minimal distance necessary for using this minimally invasive approach. Nevertheless, if the space is found intraoperatively to be too narrow, it is still possible to achieve additional space by removing bony structures (e.g., undercutting or partial hemilaminectomy of C-1 or C-2). In our experience, a du-ral opening of 10–15 mm can be achieved in most cases without bone removal. In combination with the consider-able subarachnoid space at the CVJ level, enough room is available for the surgical resection of small ventrally lo-cated lesions at the CVJ without additional bone removal. In line with the report of Martin et al.,10 additional space can be achieved by dividing or resecting the dentate liga-ment, which can provide an improved view of the ventral dura or lesion.

The surgical technique we present here is a minimally invasive approach for small ventrally located intradural le-sions at the CVJ. This approach does not require anterior or posterior fixation; therefore, physiological mobility is preserved. This can also be achieved by reconstruction after laminotomy, but muscle trauma is much greater if paravertebral muscles on both sides have to be divided. Furthermore, artifacts caused by metal implants have to be expected and may impair imaging at follow-up.7

Limitations of this minimally invasive approach are the limited space for manipulation and the restricted view. However, because of the widespread distribution of tubu-lar systems, more and more surgeons can get used to this kind of approach. Conversely, this technique is limited to the resection of small lesions of the ventral aspect at the CVJ. Lesions that extend to the contralateral side of the approach, larger lesions with a diameter of > 25 mm, es-

Neurosurg Focus  Volume 38 • April 20154

Unauthenticated | Downloaded 08/03/20 10:45 PM UTC

minimally invasive dorsal approach to the cvj

pecially in the craniocaudal direction or those that display signs of infiltration of the vertebral artery or infiltrative destruction of bone structures, still have to be addressed by extended surgical exposures to handle potential com-plications while minimizing manipulation of the neural axis.

conclusionsA minimally invasive dorsal approach involving a tu-

bular system and using the space between C-0 and C-1 or C-1 and C-2 provides direct and sufficient exposure for the surgical removal of small ventrally located intra-dural lesions at the CVJ level. Therefore, musculoskeletal preparation, with its potentially associated perioperative morbidity, is minimized by avoiding the classical anterior transoral and extensive dorsal approaches.

AcknowledgmentWe thank Mr. Arne Eicker for preparing the illustration in Fig.

1 left.

references 1. Banczerowski P, Lipóth L, Vajda J, Veres R: Surgery of ven-

tral intradural midline cervical spinal pathologies via ante-rior cervical approach: our experience. Ideggyogy Sz 56: 115–118, 2003

2. Bertalanffy H, Bozinov O, Sürücü O, Sure U, Benes L, Kappus C: Dorsolateral approach to the craniovertebral junction, in Cappabianca P, Iaconetta G, Califano L (eds): Cranial, Craniofacial and Skull Base Surgery. New York: Springer Verlag, 2010, pp 175–195

3. Crockard HA, Bradford R: Transoral transclival removal of a schwannoma anterior to the craniocervical junction. Case report. J Neurosurg 62:293–295, 1985

4. Eicker SO, Szelényi A, Mathys C, Steiger HJ, Hänggi D: Custom-tailored minimally invasive partial C2-corpectomy for ventrally located intramedullary cavernous malformation. Neurosurg Rev 36:487–491, 2013 (Erratum in Neurosurg Rev 36:493, 2013)

5. Fong S, DuPlessis SJ: Minimally invasive anterior approach to upper cervical spine: surgical technique. J Spinal Disord Tech 18:321–325, 2005

6. George B, Dematons C, Cophignon J: Lateral approach to the anterior portion of the foramen magnum. Application to surgical removal of 14 benign tumors: technical note. Surg Neurol 29:484–490, 1988

7. Ikuma H, Shinohara K, Maehara T, Yokoyama Y, Tanaka M: C2 lamina reconstruction using locking miniplate for the intradural tumor of the craniocervical junction (two case reports). Eur Spine J 21 (Suppl 4):S509–S512, 2012

8. Kawashima M, Tanriover N, Rhoton AL Jr, Ulm AJ,

Matsushima T: Comparison of the far lateral and extreme lat-eral variants of the atlanto-occipital transarticular approach to anterior extradural lesions of the craniovertebral junction. Neurosurgery 53:662–675, 2003

9. Lanzino G, Paolini S, Spetzler RF: Far-lateral approach to the craniocervical junction. Neurosurgery 57 (4 Suppl): 367–371, 2005

10. Martin NA, Khanna RK, Batzdorf U: Posterolateral cervical or thoracic approach with spinal cord rotation for vascular malformations or tumors of the ventrolateral spinal cord. J Neurosurg 83:254–261, 1995

11. Mei-Hua L, Geng-Sheng X, Zhi-Qun J, Yi-Yun L, Tao H: Supracondylar transjugular tubercle approach to intradural lesions anterior or anterolateral to the craniocervical junction without resection of the occipital condyle. Turk Neurosurg 23:202–207, 2013

12. Sen CN, Sekhar LN: Surgical management of anteriorly placed lesions at the craniocervical junction—an alternative approach. Acta Neurochir (Wien) 108:70–77, 1991

13. Watanabe M, Sakai D, Yamamoto Y, Iwashina T, Sato M, Mochida J: Upper cervical spinal cord tumors: review of 13 cases. J Orthop Sci 14:175–181, 2009

14. Yasuda M, Bresson D, Cornelius JF, George B: Anterolateral approach without fixation for resection of an intradural schwannoma of the cervical spinal canal: technical note. Neurosurgery 65:1178–1181, 2009

15. Zozulya YP, Slynko YI, Al-Qashqish II: Surgical treatment of ventral and ventrolateral intradural extramedullary tumors of craniovertebral and upper cervical localization. Asian J Neurosurg 6:18–25, 2011

Author contributionsConception and design: Schmidt, Eicker. Acquisition of data: all authors. Analysis and interpretation of data: all authors. Drafting the article: Schmidt, Eicker. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Schmidt. Statistical analysis: Schmidt, Eicker. Administrative/technical/material support: Schmidt, Eicker. Study supervision: Schmidt, Eicker.

supplemental information Videos 

Video 1, Media Player. http://mfile.akamai.com/21490/wmv/digitalwbc.download.akamai.com/21492/wm.digitalsource-na-regional/focus14-799_video_1.asx.

Video 1, Quicktime. http://mfile.akamai.com/21488/mov/digitalwbc.download.akamai.com/21492/qt.digitalsource-global/focus14-799_video_1.mov.

correspondenceNils Ole Schmidt, University Medical Center Hamburg-Ep pen-dorf, Neurosurgery, Martinistrasse 52, Hamburg 20246, Germany. email: nschmidt@uke.de.

Neurosurg Focus  Volume 38 • April 2015 5

Unauthenticated | Downloaded 08/03/20 10:45 PM UTC