Journal of Clinical Neuroscience 21 (2014) 250–255

Contents lists available at SciVerse ScienceDirect

Journal of Clinical Neuroscience

journal homepage: www.elsevier .com/ locate/ jocn

Clinical Study

Medullary hemangioblastoma: 34 patients at a single institution

0967-5868/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.jocn.2013.03.037

⇑ Corresponding author. Tel.: +86 1067098431; fax: +86 1067051377.E-mail address: yinluxin530@163.com (L. Zhang).

Luxin Yin a,b, Liwei Zhang a,⇑, Shuyu Hao a, Junting Zhang a, Zhen Wu a

a Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Tiantan Xili 6, Chongwen District, Beijing 100050, People’s Republic of Chinab Department of Neurosurgery, The Affiliated Hospital of Xu Zhou Medical College, Huaihai Xilu 99, Quanshan District, Xuzhou 221000, People’s Republic of China

a r t i c l e i n f o

Article history:Received 17 October 2012Accepted 30 March 2013

Keywords:HemangioblastomaMedullaryTreatment

a b s t r a c t

This study aimed to elucidate the surgical experience of medullary hemangioblastoma (MH) at a singleinstitution. We reviewed 34 consecutive patients with MH operated on between January 2005 and June2012 in the neurosurgery department of the Beijing Tiantan Hospital. There were 14 men and 20 women.The patients were aged from 17 to 60 years with an average age of 38 years. Tumors were cystic in 12patients (Type A), and solid in 22 patients. The solid tumors were of a small size in six patients(<3 cm, Type B), large in 12 (3.1–5 cm, Type C), and giant in four (>5 cm, Type D). Radical tumor removalwas achieved in all patients. Tracheotomy was performed in 10 patients (one Type B patient, seven TypeC, two Type D) postoperatively. Pneumonia secondary to lower cranial nerve palsy occurred in sixpatients (all Type C). Complications including intracranial infection (n = 5), gastrointestinal bleeding(n = 2), and intracranial hematoma (n = 1) also occurred in this group. Follow-up (range, 2–82 months;mean, 30 months) was available in all patients. At follow-up, 29 patients (85.3%) had a good outcome.Twenty-eight of these (82.4%) had an excellent outcome postoperatively (Karnofsky Performance StatusP80). Although transient surgical complications are possible especially for large solid tumors, total sur-gical resection can be performed with favorable long-term outcomes with meticulous microsurgical tech-nique and understanding of the vascular pattern of the tumor. Postoperative management of MH is asimportant as the operation.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction Beijing Tiantan Hospital, Capital Medical University, China. All 34

Intracranial hemangioblastomas account for 3.7% of all intracra-nial tumors at our hospital. These tumors are most commonlyfound in the cerebellar hemispheres (65%), followed by the vermis(15%), and cerebellopontine angle and brain stem (8%). The major-ity of hemangioblastomas are sporadic; about 20% are associatedwith von Hippel–Lindau (VHL) disease.1–4 Although hemangioblas-tomas are histologically benign tumors, their operative manage-ment can be challenging, especially with brain stem involvement.To elucidate the impact of neurosurgical progress on the manage-ment of these particular tumors, we retrospectively reviewed ourdata from a consecutive series of 34 patients without VHL diseaseor multiple hemangioblastomas.

2. Patients and methods

2.1. Patient population

From January 2005 to June 2012 a total of 34 consecutive adultmedullary hemangioblastoma (MH) patients underwent surgeryby the senior author (L.Z.) in the neurosurgery department of the

patients had a solitary MH and did not have a family history ofhemangioblastoma or VHL disease.

All patients or their legal representative gave written informedconsent to the study protocol, which was approved by the ethicscommittee of Beijing Tiantan Hospital.

2.2. Neuroimaging studies

All patients were evaluated pre-operatively with a CT scan andcontrast-enhanced MRI. MRI is the diagnostic test of choice, as itclearly demonstrates the location of the tumor and its relationshipto surrounding tissue.

We further divided the 34 patients with MH into four groupsaccording to tumor size. Type A: predominantly cystic tumors(the tumor nodule varied from 0.1 cm to 1.5 cm); Type B: solidtumors of a small size (<3 cm); Type C: solid tumors of a largesize (3.1–5 cm); and Type D: solid tumors of a giant size (>5 cm;Fig. 1).

2.3. Surgical procedure

All patients were operated on in the park-bench position.Electrophysiologic monitoring of the lower cranial nerves (LCN),somatosensory evoked potentials, and brainstem auditory evoked

Fig. 1. Sagittal T1-weighted MRI with gadolinium contrast showing (A) a predominantly cystic tumor (Type A; tumor nodule 0.1–1.5 cm), (B) a small solid tumor (Type B;<3 cm), (C) a large solid tumor (Type C; 3.1–5 cm), and (D) a giant solid tumor (Type D; >5 cm). (This figure is available in colour at www.sciencedirect.com.)

L. Yin et al. / Journal of Clinical Neuroscience 21 (2014) 250–255 251

potentials was used in all operations. The posterior midline ap-proach was performed for posterior, posterolateral or intramedul-lary MH; thus, the lower part of the occipital bone and theposterior arch of the atlas were opened and enlarged on the sideof the tumor.

For most anterior and anterolateral tumors, a far-lateral retroc-ondylar approach was chosen. An inverted hockeystick skin inci-sion was used that was carried superiorly to the superior nuchalline, curving below the inion and proceeding down to the spinousprocess of C2.

2.4. Postoperative management

The postoperative care of patients depended on the operativeprocedure. Patients were sent to the neurosurgical intensive careunit (NICU) after surgery. Intubation was continued for at least12 hours. If cough and gag reflexes were present on the secondpostoperative day, the endotracheal tube was removed. For pa-tients who were at risk of aspiration pneumonia or dyspnea, a tra-cheotomy was performed to prevent respiratory complications.Nasogastric feeding provided nutritional support for dysphagicpatients.

2.5. Follow-up

Follow-up data were available with clinical examination andMRI scans at 6 months and 1 year after surgery. Long-term

follow-up in this study was based on our outpatient files and, fora few patients, a telephone call or home visit. Thereafter, patientswere followed every 1 or 2 years.

3. Results

3.1. Clinical data

Fourteen patients were men and 20 were women (ratio 1:1.43).Their ages ranged from 17 to 60 years (mean, 38 years). The aver-age duration of symptoms before presentation was 11 months(range, 0.5–120 months). Paresthesia was present in 16 patients(47.1%), headache in 15 (44.1%), LCN symptoms in 12 (35.3%), vom-iting in nine (26.5%), cerebellar symptoms in eight (23.5%), dizzi-ness in seven (20.6%), pyramidal signs in seven (20.6%) and neckpain in three (8.8%). Four patients (11.8%) in this series hadundergone previous surgery, and one (2.9%) had received previousGamma Knife radiosurgery (Elekta AB, Stockholm, Sweden). Thepre-operative Karnofsky Performance Status (KPS) score was74.7 ± 11.1 (mean ± standard deviation; Table 1).

3.2. Tumor characteristics and surgical aspects

According to the imaging appearance and that during operation,tumors were located in the medulla (n = 18), ponto-medulla(n = 13), and cervico-medulla (n = 3). Tumors were cysticin 12 patients (Type A), and solid in 22 patients. Tumors were small

Table 1Demographics and tumor characteristics of the 34 patients with medullary hemangioblastoma

Patient Sex Age(years)

Duration of symptoms(months)

KPS (beforesurgery)

Tumor bloodsupply

Symptoms

1 M 20 1.5 40 1 PICA Cerebellar sign, pyramidal signs, paresthesia2 F 17 12 60 1 PICA + 1 VA Vomiting, pyramidal signs3 F 47 6 40 2 PICA Cerebellar signs, pyramidal signs, cranial nerve deficit4 F 21 4 80 2 PICA + 2 VA Cranial nerve deficit, paresthesia5 F 49 3 80 Unknown Cranial nerve deficit, paresthesia6 F 23 3 80 Unknown Paresthesia7 F 37 3 90 1 PICA Headache, vomiting8 M 27 15 80 1 PICA Headache, paresthesia, cerebellar signs, dizziness9 M 47 9 80 Unknown Paresthesia10 M 22 15 70 2 PICA Headache, paresthesia, cerebellar signs11 M 40 18 60 Unknown Pyramidal signs, cranial nerve deficit, cerebellar signs12 M 36 2 80 Unknown Cranial nerve deficit, paresthesia13 M 41 2 80 1 PICA Headache, paresthesia, cranial nerve deficit14 M 57 3 70 1 PICA Headache, pyramidal signs15 F 58 15 80 1 PICA Cranial nerve deficit, paresthesia16 M 44 24 70 2 PICA Cerebellar signs, dizziness17 M 41 12 90 Unknown Cranial nerve deficit18 F 48 6 80 Unknown Paresthesia19 M 25 12 70 2 PICA Cerebellar signs, dizziness20 M 35 3 80 Unknown Headache, vomiting, dizziness21 F 35 4 80 Unknown Headache, vomiting, cerebellar signs, neck pain22 F 30 3 80 Unknown Cranial nerve deficit, paresthesia23 F 41 0.5 80 Unknown Dizziness, vomiting24 M 34 3 80 Unknown Headache25 F 40 12 80 Unknown Headache, vomiting, dizziness26 F 60 12 80 Unknown Headache, vomiting, dizziness27 F 40 120 70 2 PICA + 2 VA Headache, vomiting, cranial nerve deficit, paresthesia28 M 37 3 70 1 AICA Headache, pyramidal signs29 F 47 6 80 2 PICA Headache30 F 52 24 70 1 PICA + 1 AICA Paresthesia31 M 44 6 70 2 PICA Headache, vomiting, cranial nerve deficit, paresthesia,

pyramidal signs32 M 37 15 80 2 PICA Neck pain33 F 40 2 80 Unknown Headache, cranial nerve deficit34 F 38 1.5 80 Unknown Neck pain, paresthesia

AICA = anterior internal cerebellar artery, F = female, KPS = Karnofsky Performance Status score, M = male, PICA = posterior internal cerebellar artery, VA = vertebral artery.

252 L. Yin et al. / Journal of Clinical Neuroscience 21 (2014) 250–255

in six patients (<3 cm, Type B), large in 12 (3.1–5 cm, Type C), andgiant in four (>5 cm, Type D). A posterior midline approach wasperformed in 33 patients, and a far-lateral approach in the oneremaining patient. Gross total resection was achieved in 34(100%) patients.

3.3. Blood supply of tumors

An angiogram was performed in 18 patients and magnetic res-onance angiography in 10 patients. When detectable, the bloodsupply mainly came from the posterior internal cerebellar artery(94.4%) and anterior inferior cerebellar artery (11.1%). These tu-mors can also derive their blood supply from branches of the ver-tebral artery. Tumors commonly drained into the straight, lateral,and/or petrosal sinuses (Table 1).

3.4. Surgical morbidity and mortality

The mean stay in the NICU was 2.1 days and ranged from 1 to25 days. Three patients had mechanical ventilation after surgery(two Type C patients and one Type D). Temporary dysphagia oc-curred in 10 patients (one Type B, six Type C, three Type D), andnasogastric feeding was continued for 8 to 94 days with a meanof 12.2 days; this included all eight patients who had pre-operativedysphagia. Postoperative hemianesthesia occurred in three pa-tients. One patient (Type C) had postoperative hematoma evacua-tion. One patient (Type C) developed postoperative hydrocephalusand required a ventriculoperitoneal shunt. Postoperative cerebro-spinal fluid leakage occurred in one patient who required surgical

repair. Five patients developed intracranial infection. Postoperativegastrointestinal bleeding occurred in two patients (both Type C).Two (5.9%) patients (both Type C) died, both due to severe pneu-monia after the operation.

Intubation was required for a mean of 12.7 days (range, 1–70 days) and a tracheotomy was performed in 10 patients. In the10 patients who underwent tracheotomy, none were Type A (0%),one was Type B (1/6), seven were Type C (7/12) and two were TypeD (2/4). Six patients developed intracranial infection after tracheot-omy (6/10), two of whom died.

3.5. Follow-up

A total of 34 patients were followed for a median of 30 months(range, 2–82 months). One patient developed a recurrence, and an-other operation was performed. The mean postoperative KPS was81.2 ± 27.3 (standard deviation). Among the 34 followed patients,29 patients (85.3%) had a better KPS score compared to beforethe operation, 28 (82.4%) had a normal life (KPS 80–100), one(2.9%) had moderate disabilities (KPS 50–80), and five (14.7%)had severe disabilities (KPS < 50), including two patients who diedabout 1 month after surgery. There was no significant differencebetween the postoperative KPS of patients with Type A, B, C or Dtumors (p > 0.05; Table 2).

4. Discussion

Intracranial hemangioblastomas account for 3.7% of all intracra-nial tumors at our hospital. These tumors are most commonlyfound in the cerebellar hemispheres (65%), followed by the vermis

Table 2The postoperative complications and Karnofsky Performance Status score at follow-up of the 34 patients with medullary hemangioblastoma

Patient Pneumonia Tracheotomy Hematoma Gastrointestinal bleeding Follow-up time (months) KPS (at follow-up)

1 N N N N 3 902 N Y N N 4 403 N N N N 3 404 N N N N 2 1005 N N N N 4 806 N N N N 30 1007 N N N N 48 908 N N N N 48 1009 N N N N 41 9010 N Y N N 19 9011 N N N N 38 5012 N N N N 48 9013 N N N N 24 9014 Y Y Y Y 18 4015 N N N N 16 10016 N N N N 25 10017 N N N N 25 10018 N N N N 24 10019 Y Y N N 23 8020 N N N N 25 10021 N Y N N 82 8022 N N N N 42 8023 N N N N 40 8024 N N N N 31 10025 N N N N 34 9026 N Y N N 25 9027 N N N N 79 10028 N N N N 79 10029 N N N N 23 10030 Y Y N N 27 031 Y Y N N 28 032 Y Y N Y 31 8033 Y Y N N 10 9034 N N N N 34 100

KPS = Karnofsky Performance Status score, N = no, Y = yes.

L. Yin et al. / Journal of Clinical Neuroscience 21 (2014) 250–255 253

(15%), and cerebellopontine angle and brain stem (8%).1 Althoughthere has been great improvement in the anatomical knowledgeof this region, MH still represent a surgical challenge due to theirarteriovenous malformation-like vascularisation and their uniquelocation. This study aimed to elucidate the surgical treatmentexperience with MH at a single institution. As genetic factors havean unpredictable impact on patient outcome, only patients whohad one MH and were without a family history of hemangioblasto-mas or VHL disease were selected.

Hemangioblastomas are histologically benign vascular neo-plasms composed of endothelial and stromal cells, with a tendencyfor peritumoral edema and cysts. The stromal cells represent theneoplastic component of the tumour with occasional atypical andhyperchromatic nuclei varying in size.5

The complications of treatment of MH are important and con-troversial. Some MH never produce symptoms and do not requiretreatment. Some MH do not produce symptoms initially but doso in time. Resche et al.6 concluded that only symptomatic andgrowing lesions should be operated on. However, radiographic pro-gression alone is not the optimal predictor of the need for therapy.Ammerman et al.3 reported that the initial tumor size and growthrate of MH were predictors of symptom development and the needfor treatment. At our institution, we recommend conservativetreatment for asymptomatic and stable MH. Careful and regularclinical follow-up with MRI examination should be done for thesepatients.1,2,4,6–8

For symptomatic MH, surgery is the best treatment choice. Inour series, all patients were symptomatic and the most frequentsymptoms were paresthesia in 16 patients (47.1%), headache in15 (44.1%) and LCN symptoms in 12 (35.3%). All patients receivedsurgery, and most achieved excellent outcomes. Gamma Knife

radiosurgery (Elekta AB) can be considered for multiple hema-ngioblastomas that cannot be removed in one operation but isnot recommended as the optimal therapy for MH.9–12 The tumoralways involves important structures that are a risk for complica-tions after radiotherapy.9–12 In addition, radiotherapy can make fu-ture resection of the tumor difficult, as it causes the tumor tobecome hard and increased peritumoral edema increases the de-gree of difficulty.

However, considering the vascularity and critical site, if the MHcan not be totally removed, radiotherapy is recommended forresidual tumor. Since the mortality and complication rates increasesignificantly with secondary surgery, it is preferable to carry out enbloc resection during the primary surgery.13

MH surgery should be performed in a neurosurgery skull basecenter that performs such procedures in high volume and withexperienced neurosurgeons. Because of the hypervascular natureof the tumor, biopsy and piece-meal excision should be avoidedas this may cause uncontrollable hemorrhage, regardless ofwhether the tumor is cystic or solid. A clear operative field isessential to protect the important vascular and neural structuresof brain stem. The surgical principles for hemangioblastoma is sim-ilar to that for arteriovenous malformations: selective division offeeding arteries, preservation of the main draining veins, and pro-tection of vessels across the surface of the tumor during dissection.Feeding arteries must be coagulated initially and as close to the tu-mor as possible to prevent damage to other arterial branches feed-ing the brainstem. With a diminishing blood supply, the lesionshrinks. Each succeeding feeding artery should be coagulated care-fully. Cleavage of the main draining veins, which are extremely di-lated, should be carried out at the last moment. Total removal oftumor is vital.1–4

Fig. 2. Preoperative embolization was a helpful and safe procedure that reduced blood loss and allowed for total tumor resection in this patient. (A) Pre-operative sagittal T1-weighted MRI with gadolinium contrast showing a large solid tumor (Type C), (B) digital subtraction angiography before embolization showing the feeding arteries, (C) digitalsubtraction angiography after embolization showing the reduced tumor stain, and (D) postoperative sagittal T1-weighted MRI with gadolinium contrast showing completeremoval of the tumor.

254 L. Yin et al. / Journal of Clinical Neuroscience 21 (2014) 250–255

Most authors believe that pre-operative embolization is a help-ful and safe procedure that reduces blood loss and allows total tu-mor resection, especially in the spine and brainstem, althoughsome authors generally abstain from it.1,2,4,13,14 In our series, weused pre-operative embolization of the main blood supply vesselsfor large tumors (Types C and D) and found that this was especiallyuseful for the large tumors in terms of producing a less vascular tu-mor (Fig. 2). Following the procedures described, gross total resec-tion was achieved in all 34 (100%) patients in our series whileangiogram and pre-operative embolization were performed in 18patients (Fig. 3).

With the development of skull base microneurosurgery, theremoval of MH has become less difficult than it was decadesago. Postoperative management of MH is as important as thesurgery. The temporary LCN deficit rate after surgery in ourseries was 29.4%. Rachinger et al.15 suggested that pre-operativeimpaired nerve function was a negative predictor for functionaloutcome. They stressed that pre-operative LCN functionalevaluation is helpful for postoperative management. Zhou et al.2,4

theorised that because the respiratory center often has a tempo-rarily reduced responsiveness to carbon dioxide, presenting asrespiratory depression, after removal of MH, careful managementof respiratory and cardiac function was of great importance, notonly during the operation, but during the early postoperativedays as well.

In our study, intubation was required for a mean of 12.7 days(range, 1–70 days), temporary dysphagia occurred in 10 patients(one Type B, six Type C, three Type D), and tracheotomy was per-formed in 10 patients. In the 10 patients who underwent tracheot-omy, none was Type A (0%), one was Type B (1/6), seven were TypeC (7/12) and two were Type D (2/4). Six patients developed intra-cranial infection after tracheotomy (60%), two of whom died. TypeA and B tumors had fewer complications than Type C and D. Theendotracheal tube was removed on the second postoperative dayfor all Type A patients.

We have emphasized that airway management is especiallyimportant after MH surgery. Patients who had pre-operative respi-ration difficulty required postoperative tracheotomy. Trachealintubation should be maintained for at least 12 hours postopera-tively for all patients. If cough and gag reflexes are normal on thesecond postoperative day, the endotracheal tube can be removedafter brain swelling and hemorrhage have been ruled out by a CTscan, and thereafter, the patient’s vital signs should be monitoredcautiously for at least 30 minutes. For patients with weak coughand gag reflexes (in our series, most were Type C and D patientswith LCN deficits), tracheostomies were performed to avoid respi-ratory disturbance. Prediction of later removal of the endotrachealtube requires further study. Saliva should be cleared in a timelymanner to avoid aspiration pneumonia especially for the patientswith tracheostomies (6/10 patients in our series, including two

Fig. 3. Intra-operative photographs showing (A) a large medullary hemangioblastoma, and (B) complete removal of the tumor.

L. Yin et al. / Journal of Clinical Neuroscience 21 (2014) 250–255 255

who died). Nasogastric feeding can provide adequate nutritionalsupport for dysphagic patients. After swallow training, mostpatients had improved swallowing function within a short timepostoperatively.

At the end of follow-up of our 34 patients, 28 (82.4%) had a nor-mal life (KPS of 80–100), and only two patients (Type C) died, dueto pneumonia. Although the resection of large solid tumors (Type Cand D) was often associated with serious complications, small(Type B) and cystic (Type A) tumors could be excised safely. Wedid not find any significant difference between the pre-operativeand postoperative KPS between the four types. Though the TypeC and D patients had more complications than Type A and Bpatients, tumor size was not a clear predictor of outcome inour series, which is in concordance with Rachinger et al.15 andWeil et al.16 Nevertheless, we agree with other authors thatoperative treatment is more difficult with increasing tumorsize.2,4,13,17,18

5. Conclusions

For symptomatic MH, surgery is the best treatment choice.Although transient surgical complications are possible, especiallyfor large solid tumors, total surgical resection can be performedwith favorable long-term outcome due to advanced microsurgicaltechniques and understanding of the vascular pattern of thetumor. Postoperative management of MH is as important as thesurgery.

6. Conflicts of Interest/Disclosures

The authors declare that they have no financial or otherconflicts of interest in relation to this research and its publication.

References

1. Wang C, Zhang J, Liu A, et al. Surgical management of medullaryhemangioblastomas. Report of 47 case. Surg Neurol 2001;56:218–26.

2. Zhou LF, Du GH. Diagnosis and surgical treatment of posterior fossa solidhemangioblastomas. Chin Med J 2000;113:129–32.

3. Ammerman JM, Lonser RR, Dambrosia J, et al. Long-term natural history ofhemangioblastomas in patients with von Hippel–Lindau disease: implicationsfor treatment. J Neurosurg 2006;105:248–55.

4. Zhou LF, Du GH, Mao Y, et al. Diagnosis and surgical treatment of brainstemhemangioblastomas. Surg Neurol 2005;63:307–16.

5. Rachinger J, Buslei R, Prell J, et al. Solid hemangioblastomas of the CNS: a reviewof 17 consecutive cases. Neurosurg Rev 2009;32:37–47.

6. Resche F, Moisan JP, Mantoura J, et al. Hemangioblastoma,hemangioblastomatosis and von Hippel–Lindau disease. Adv Tech StandNeurosurg 1993;20:197–304.

7. Cristante L, Herrmann HD. Surgical management of intramedullaryhemangioblastoma of the spinal cord. Acta Neurochir (Wien) 1999;141:333–9.

8. Lonser RR, Oldfield EH. Microsurgical resection of spinal cordhemangioblastomas. Neurosurgery 2005;57:372–6.

9. Matsunaga S, Shuto T, Inomori S, et al. Gammaknife radiosurgery forintracranial haemangioblastomas. Acta Neurochir (Wien) 2007;149:1007–13.

10. Karabagli H, Genc A, Karabagli P, et al. Outcomes of gamma knife treatment forsolid intracranial hemangioblastomas. J Clin Neurosci 2010;17:706–10.

11. Kano H, Niranjan A, Mongia S, et al. The role of stereotactic radiosurgery forintracranial hemangioblastomas. Neurosurgery 2008;63:443–50.

12. Sayer FT, Nguyen J, Starke RM, et al. Gamma knife radiosurgery for intracranialhemangioblastomas – outcome at 3 years. World Neurosurg 2011;75:99–105.

13. Wan JQ, Cui H, Wang Y. Surgical management of large solidhemangioblastomas of the posteriorfossa. J Clin Neurosci 2011;18:39–42.

14. Tampieri D, Leblanc R, TerBrugge K. Preoperative embolization of brain andspinal hemangioblastomas. Neurosurgery 1993;33:502–5.

15. Rachinger J, Buslei R, Prell J, et al. Solid hemangioblastomas of the CNS: areview of 17 consecutive cases. Neurosurg Rev 2009;32:37–47.

16. Weil RJ, Lonser RR, De Vroom HL, et al. Surgical management of brainstemhemangioblastomas in patients with von Hippel–Lindau disease. J Neurosurg2003;98:95–105.

17. Krishnan KG, Schackert G. Outcomes of surgical resection of large solitaryhemangioblastomas of the craniocervical junction with limitations in pro-operative angiographic intervention: report of three cases. Zentralbl Neurochir2006;67:137–43.

18. Parker F, Aghakhani N, Ducati LG, et al. Results of microsurgical treatment ofmedulla oblongata and spinal cord hemangioblastomas: a comparison of twodistinct clinical patient groups. J Neurooncol 2009;93:133–7.