Brain-stem hemangioblastomas: The seemingly innocuous lesion in a perilous location
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.232294
Source of Support: None, Conflict of Interest: None
Keywords: Brain-stem hemangioblastoma, complication avoidance, outcome, radiology, surgery
Hemangioblastomas are highly vascular but benign lesions present almost exclusively in the central nervous system and pose a unique challenge to neurosurgeons. 75% of them are sporadic while 25% are in association with Von Hippel Lindau (VHL) disease, an autosomal dominant neoplastic syndrome. These World Health Organization (WHO) grade I neoplasms constitute 1.5-3.7% of all intracranial tumours and 7-12% of posterior fossa mass lesions. 95% of these tumours are found in the cerebellar hemispheres (65%), vermis (15%), cerebellopontine angle and the brain-stem (8%). Though often cystic in other locations, brain-stem hemangioblastomas have more of a solid component, a high flow status with significant arteriovenous shunts within the tumour, and, the tumour often does not have a well-defined plane of cleavage with the brain-stem. Thus, brain-stem hemangioblastomas present a unique challenge in view of their eloquent location and vascularity. To understand the surgical nuances involved in the safe resection of these benign vascular tumours, we present our experience from a consecutive series of 27 patients. The objective of this study was to illustrate the relevant surgical nuances required in the excision of these tumours, evaluate the safety and efficacy of microsurgical resection, and analyze the outcomes.
Between 1st January 2000 and 28th February 2018, 27 consecutive cases with a brain-stem hemangioblastoma underwent microsurgical excision at the Department of Neurosurgery, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow. The patient were followed up on an outpatient basis at regular intervals.
Only those patients who had radiological involvement of the midbrain, pons or medulla, with or without the involvement of cerebellum or spinal cord, and in whom the tumour was histopathologically proven to be a hemangioblastoma were included in the study.
All patients were assessed clinically for the presence of neurological deficits, both before and after the surgery and at each follow up visit. The presence of raised intracranial pressure, long tract signs, cranial nerve palsies and Von Hippel Lindau (VHL) disease [Figure 1] was also documented.
All patients underwent a preoperative gadolinium contrast enhanced magnetic resonance imaging (MRI). Radiological involvement of the midbrain, pons or medulla was defined. On the basis of the location of the lesion, it was determined if it was anteriorly or dorsally placed, subpial or intra-medullary. The nature of the lesion, whether solid or solid-cystic, and the presence of haemorrhage were documented, along with the additional involvement of the cerebellum and spinal cord. We divided the tumours into four categories based on size:
Category A: Size less than or equal to 2 cm, Category B: Size between 2 to 3 cm, Category C: Size between 3 to 4 cm, and Category D: Size more than 4 cm.
Based on their contrast enhancement, the tumours were divided into those with heterogeneous or homogenous enhancement, and with a single mural nodule or multiple ones. All patients were screened for the presence of stigmata of Von Hippel Lindau (VHL) syndrome. In patients with a hemangioblastoma, a detailed family history was undertaken. An ultrasound of the abdomen was carried out in every patient to detect any mass/cystic lesion and an indirect fundoscopy for retinal angiomas was also done. Preoperative blood pressure (BP) monitoring was done in the ward at regular intervals. In case any raised BP readings were seen, the 24-hour urinary metanephrine and serum metanephrine levels were estimated. Patients with a positive abdominal ultrasound scan were subjected to a triple phase (portal, arterial and venous phase) abdominal computed tomographic (CT) scan.
Patients in whom a solid, brilliantly enhancing lesion in the brain-stem with multiple flow voids was seen on MRI, a digital subtraction angiography (DSA) was done; and, 2 patients with a high flow lesion underwent a preoperative embolization.
After a thorough pre-operative evaluation, all our patients underwent surgical excision of the tumour. The midline suboccipital craniectomy in the prone Concorde position, with or without a C1 laminectomy or undercutting of the C2 lamina (based upon the inferior extent of the lesion), was performed. Under an operating microscope, a Y-shaped durotomy was performed, arachnoidal membrane opened and the tumour exposed. The tumour was shrunk by coagulation under copious irrigation. In case a well-defined gliotic plane of cleavage from the brain-stem was visible, the tumour was removed in that plane. In case, the tumour was densely adherent to the brain-stem, a thin part of the tumour capsule was left in situ. The extent of resection, the presence of haemorrhage, the feasibility of complete excision of the cyst wall and the extent of involvement of the midbrain, pons or medulla were documented during surgery. One patient with a ventrolateral lesion extending to the cerebellopontine angle underwent a retromastoid suboccipital craniectomy. Contrast enhanced CT scan/MRI was done postoperatively to assess for the extent of resection and to exclude a surgical cavity haematoma. The extent of resection was defined as total, near-total or sub-total. Total resection was defined as absence of any contrast enhancement in the post-operative scan with the intraoperative assessment of total excision by the surgeon. Near-total excision was defined as less than or equal to 10 per cent contrast enhancing residual tumour seen on the postoperative scan. Anything more than 10 percent residual tumour visible on the postoperative scan was taken as subtotal excision. All of these patients, following the surgical excision of their lesions, were electively ventilated for 12 hours, prior to their extubation in the intensive care unit after ensuring that an adequate consciousness level as well as an adequate motor and respiratory response had been established.
The outcome of the patients was described clinically based on the Karnofsky Performance scale. The patients were followed up at 6 weekly intervals for 6 months, then at 3 monthly intervals for one year, and then at six monthly thereafter, on an outpatient basis. The postoperative complications were documented.
Statistical analyses of data were done using the Statistical Package for the Social Sciences (SPSS), version 20 (IBM, Chicago, IL). Mean with standard deviation was calculated for parametric values. For non-parametric values, median with interquartile range was calculated.
27 patients who underwent surgical excision of a brain-stem hemangioblastoma were included in the study. 16 (59.2%) of them were males. The mean age of the study group was 29 years (range: 15-60 years). 17 (63%) of the patients were below 30 years of age.
19 (70.4%) patients had hydrocephalus at presentation. All of them presented with headache with occasional visual blurring (n = 8, 29.6%) and vomiting (n = 16, 59.3%). 14 of them had a moderate hydrocephalus and 5 had a mild hydrocephalus. 9 patients underwent a cerebrospinal fluid (CSF) diversion procedure preoperatively while 2 underwent a postoperative ventriculoperitoneal shunt. A patient also underwent an endoscopic third ventriculostomy in the postoperative period. Gait ataxia was also very commonly associated, with cerebellar signs being present in 21 (77.8%) patients. Features of lower cranial palsy were present in 10 (37%) patients. Long tract signs in the form of motor weakness was present in 13 (48.1%) patients and sensory involvement was seen in 6 (22.2%) patients. The average duration of symptoms was 14 months and only one patient presented with acute onset of symptoms within 3 days. The longest duration of symptoms was 7 years in the case of a recurrent pontomesencephalic tumour [Table 1].
According to their size, the number of tumours in each of the categories was as follows: Category A: 5 (18.5%); Category B: 10 (37%); Category C: 6 (22.2%); and, Category D: 6 (22.2%). Their mean size was 3.34 ± 1.06 cm (range: 1.4-5.5 cm). Thus, 81.5% of the tumours were more than 2 cm in size and 22.2% tumours were giant, i.e., more than 4 cm in their greatest dimension. 11 of the 12 patients belonging to the category C and D had hydrocephalus. 8 of the 15 patients belonging to category A and B also had hydrocephalus.
The mean duration of symptoms before presentation did not vary significantly between the four groups. Thus, the median duration of symptoms in various categories based upon size was: Category A: 4 ± interquartile range 2-15 months, Category B: 5 ± interquartile range 1-34 months, Category C: 7 ± interquartile range 4-30 months, and Category D: 4 ± interquartile range 2.75-25.5 months.
Within the brain-stem, maximum number of tumours were located at the posterior cervicomedullary junction (n = 12). In 7 patients, the tumour was located at the pontomedullary region. This was followed by tumours located at the pons (n = 3), medulla (n = 3) and the pontomesencephalic region (n = 2) [Table 2]. 11 of the tumour were solid and the rest were solid-cystic. Of the 6 tumours of size >4 cm, only one was solid. Two of our patients with solid-cystic tumours near the cervicomedullary junction had a preoperative sinus tachycardia, which resolved gradually in the postoperative period within a week.
18 (66.7%) tumours were subpial and the rest were intramedullary [Figure 2]. Of the 18 subpial tumours, 10 were solid-cystic and 8 were solid. All the subpial tumours were dorsally placed. Of the 9 intramedullary tumours, 3 were solid and 6 were solid-cystic [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8] and [Videos 1, 2]. Among the 9 intramedullary tumours, 3 were ventrolateral in location within the substance of the brain-stem. One patient had a large (5.5 × 4 cm) solid-cystic tumour extending from medulla to the left cerebellomedullary cistern where encasement of the lower cranial nerves was seen [Table 1].
All tumours were T1 hypointense and T2 hyperintense on MRI. Multiple flow voids were seen on T2 weighted images both within the lesion and both proximally and distally to it in more than 50% of patients with a tumour size greater than 2cm [Table 1]. 5 patients had an associated syringomyelia. In 8 patients, the upper cervical cord was diffusely expanded. Regarding the contrast enhancement pattern, 7 had a heterogeneous enhancement, 9 had a homogeneous enhancement, 2 had multiple mural nodules and 9 had a single mural nodule. In 6 patients, the mural nodule was intrinsically adherent to the brain-stem surface without a definite plane of cleavage with it.
Four patients underwent a digital subtraction angiography (DSA). In 2 patients, in whom the tumour was large with a diffuse nidus and a rapid flow from the arterial to the venous phase, and no cleavage from the surrounding brain-stem was visible, a preoperative embolization of the tumour utilizing Guglielmi detachable coils and polyvinyl alcohol was done to reduce vascularity and ensure a safe resection. One of these two patients, who underwent a preoperative embolization, had a 3 × 2 × 3 cm solid pontomedullary lesion fed by the left posterior inferior cerebellar artery (PICA) and bilateral anterior inferior cerebellar artery (AICA) branches. The other 16-year old patient had the VHL syndrome with a 4 × 4 × 3 cm solid pontomedullary lesion fed by bilateral PICA and left AICA branches. The right PICA was embolized twice over two years resulting in a mild decrease in tumour blush. Another patient showed multiple large flow voids in close proximity to the floor of the fourth ventricle on T2 weighted (W) MRI scan. Suspecting an arteriovenous malformation with or without an associated aneurysm, a DSA was carried out. There was no aneurysm seen on DSA and the tumour was fed by bilateral PICA branches and meningeal branches of the right vertebral artery. The fourth patient underwent a DSA as he presented with posterior fossa cisternal and fourth ventricular bleed, and was found to be having a solid 3 × 3 cm medullary lesion in association an aneurysm in the left tonsillomedullary segment of PICA, which was clipped intra- operatively using a 5 mm straight Yasargil clip [Figure 9] and [Figure 10].
Implications of VHL disease
There were 6 patients with VHL disease in our series. 50% of the tumours in patients with VHL disease were solid, and the rest were solid-cystic. 50% of the tumours that occurred in association with VHL disease were intramedullary.
All of 6 patients suffering from VHL disease had retinal angiomas. One patient was diagnosed to be having a renal cell carcinoma with multiple pancreatic cysts and underwent a laparoscopic radical nephrectomy with excision of mesenteric lymph nodes prior to the hemangioblastoma excision. She also had a 2 × 1.5 × 1 cm hemangioblastoma in the left posterior frontal lobe, which was asymptomatic and is currently being managed with serial observation. One patient underwent bilateral adrenalectomy for bilaterally present pheochromocytomas prior to the tumour excision. He also had renal and pancreatic cysts. One patient had multiple hemangioblastomas at the cervical and dorsal levels in addition to the cervicomedullary tumour. She also had a cystic lesion in the pancreas, which was conservatively managed. Two other patients had adrenal masses (which were asymptomatic, and therefore, not biopsied) that were conservatively managed.
Surgery and outcome characteristics
Twenty-six patients underwent a midline suboccipital craniectomy. In 7 patients, the posterior arch of Atlas More Details was removed. In 2 patients, a C1-C2 laminectomy was done. In one patient with the tumour extending from the medulla oblongata to the left cerebellomedullary cistern, a left retromastoid suboccipital craniectomy was done. In 15 patients, total excision was done; in 10 patients, a near-total, and in 2 patients, a subtotal excision was possible [Table 3] and [Figure 11], [Figure 12], [Figure 13].
Three patients died in the perioperative period. Among them, two of the patients had the VHL disease. Of the three patients who succumbed, one patient with a history of VHL disease with a giant pontomedullary solid-cystic tumour reaching to the cerebellomedullary cistern developed exsanguinating intraoperative hematoma and succumbed in the postoperative period. She did not undergo a preoperative therapeutic embolization [Figure 14].
One of the patients had a giant, high flow brain-stem hemangioblastoma [Figure 15], [Figure 16], [Figure 17], [Figure 18], [Figure 19]. In 2014, he underwent an MRI followed by therapeutic embolization. In view of his minimum neurological deficits at that time, he was kept on a regular follow-up and a definitive surgery was deferred. In 2017, he was readmitted with significant lower cranial nerve paresis, requiring a tracheostomy and nasogastric feeding. He also developed aspiration pneumonia, and underwent prolonged antibiotic treatment and physiotherapy. He also had persistent hypertension in the preoperative period. His level of urinary normetanephrine was 5476 microgram per 24 hours (normal values: Less than 600 microgram per 24 hours) but with a normal serum metanephrine (291 microgram per 24 hours; normal values: Less than 600 microgram per 24 hours). He was administered prazocin (selective alpha-1 adrenergic blocker) in the dose of 2 mg TDS for his hypertension. On ultrasound of the abdomen, a single cyst in the right kidney with a cystic lesion in the head of the pancreas was detected. The contrast enhanced CT scan of the abdomen showed a 3.7 × 3.9 cm well-defined, avidly enhancing mass lesion arising from the right adrenal gland; and, another similar lesion 1.5 × 1.2 cm arising from the left adrenal gland. He underwent an open, transperitoneal, bilateral adrenalectomy. After surgery, he was continued on oral prednisolone 5mg in the morning, 2.5 mg in the evening, and fluodrocortisone 0.1 mg maintenance dose per day. Preoperative embolization of the right PICA was done which caused diminution in the tumour blush. During surgery, there was a significant blood loss (4 liters blood loss so that 7 units of packed cell and 7 units of fresh frozen plasma transfusion were given). He had multiple flow voids with a significant communication between the brain-stem and tumour vessels. There was no significant cleavage between the tumour and the brain-stem. Following surgery, he developed a spreading venous infarction with significant brain-stem edema and a hematoma at the operative site and did not regain consciousness. He eventually developed cerebrospinal fluid (CSF) leakage through the wound, had meningitis and fulminant sepsis. The third patient could not be weaned off from the ventilator and developed ventilator associated pneumonia.
Three patients developed an operative site haematoma in the postoperative period. One of them had an intraventricular haematoma, which was evacuated endoscopically and the patient recovered well; the second patient developed a hematoma at the operative site and was re-explored on the next postoperative day, with the haematoma and the residual tumour being evacuated. He also had a good postoperative recovery. The haematoma in the third patient was relatively smaller and could be managed conservatively.
Five patients had a new-onset cranial nerve palsy. In four of these patients, the cranial nerve palsy was transient. Five of the patients in the series also developed bronchopneumonia. Of the 10 patients with a preoperative lower cranial nerve palsy, 4 patients had improvement in their lower cranial nerve palsy in the postoperative period, and in others, there was persistence of mild nasal regurgitation and hoarseness of voice. Two patients developed transient new onset motor deficit. One of these patients who developed quadriplegia postoperatively had complete recovery at the 6-week follow-up visit. The other patient in whom a surgical cavity haematoma had been evacuated, developed quadriparesis and is recovering gradually. One patient developed new onset ataxia.
The tumour location had no bearing on the postoperative outcome of the patient. However, the patients with giant solid tumours (n = 3, 11%) had a high flow vascular status, with a poor plane of cleavage from the brain-stem and often caused significant pressure on the vagal nerve nuclei at the calamus scriptorius region of the brain-stem. Occasionally, they also surrounded the lower cranial nerves emerging from the medulla oblongata. Thus, these tumours were often associated with neurological deterioration of the motor functions, with associated lower cranial nerve palsy and a significant intraoperative bleeding. At discharge, 12 patients had the same KPS as was seen in the preoperative period, 6 had an improved KPS and 6 had a worsened KPS. At their follow-up visit (median: 25 ± interquartile range 2-56 months), 17 patients had an improved KPS, 4 remained in the same status and 3 had a worsened KPS. The three patients with a worse KPS have been recently operated, have a less than 1.5 months follow-up and are on a tracheostomy. They are awaiting closure of their tracheostomy following recovery of their lower cranial nerve function [Table 3]. [Table 4] comparing the result of this series with the major published studies on brain-stem hemangioblastomas shows that surgical excision of these tumours often results in a lasting improvement in the functional grade of the patients.,,,,,,,,,,
Salient clinicoradiological features
Hemangioblastomas are benign tumours derived from embryonal hemangioblasts which contain endothelial and stromal elements. The neoplastic component is the stromal element. The endothelial component, giving rise to the tumour vasculature, is highly leaky with increased permeability, leading to the formation of a peritumoural cyst comprising of plasma ultra-filtrate. Symptoms are usually due to expansion of the cyst or the solid component and the resulting brain-stem compression.
81.5% of the tumours in our study were of more than 2 cm in size and 22.2% of the tumours were giant, i.e., more than 4 cm in their greatest dimension [Table 1]. This may have been responsible for the high incidence of hydrocephalus, lower cranial nerve palsy and long tract signs prevalent in the series.
The high incidence of hydrocephalus resulting in papilledema in nearly 44% patients and the requirement of a pre- or a postoperative cerebrospinal fluid diversion procedure in 12 of them was a unique feature of our study, refuting the often-quoted adage that brain-stem tumours usually do not present with CSF pathway obstruction. Both the CSF obstruction at the level of the aqueduct and the foramen of Magendie and Lushka may have been responsible, especially in those tumours with a rapidly expanding cystic component; multiple microbleeds emnating from the tumour, causing arachnoidal granulation obstruction at the level of the superior sagittal sinus may also have been implicated in the causation of communicating hydrocephalus in relatively smaller-sized tumours.
Another important symptom was the high incidence of lower cranial nerve paresis, seen preoperatively in 10 (40.7%) patients. The critical location of the tumour in the vicinity of vagal nerve nucleus at the calamus scriptorius of the medulla oblongata, the associated compression of the vagal nuclei due to the solid and cystic components of the tumour, the associated syringomyelia and syringobulbia, or the involvement of the lower cranial nerves at the level of the cervicomedullary cistern may have been implicated as the causative factors for lower cranial nerve palsy.
Small (10 mm or less) hemangioblastomas were usually isointense on T1-weighted images and hyperintense on T2-weighted images and showed a homogeneous well-demarcated enhancement. These smaller hemangioblastomas were detected as they showed symptoms due to their association with a significant syrinx (which was seen in 5 patients in our study) or their close proximity to the vagal nerve nuclei in the posterior aspect of the medulla oblongata.
Larger hemangioblastomas tended to be hypointense or mixed hypo- and isointense on T1-weighted images, heterogeneous on T2-weighted images, with heterogeneous enhancement and were usually associated with vascular flow voids. The latter finding was especially marked in our study where tumours less than or equal to 2 cm had no flow voids but nearly 50% or more tumours of size greater than 2 cm were associated with vascular flow voids.
It may be important to differentiate a peritumoural cyst and a syrinx from a tumoural cyst to ensure completeness of tumour excision. This may be done by the observation of a lack of enhancement of the wall of the former two conditions, while the wall of the tumoural cyst is usually enhancing. However, differentiating between a peritumoural cyst and a syrinx is only possible using the well-defined criteria laid down by Chu et al. A tumour associated syrinx is a cystic cavity larger than one vertebral segment in length and contains fluid. A peritumoural cyst, on the other hand, is defined as a cystic cavity adjacent to the tumour and confined within one vertebral segment. Brain-stem and cord edema in the vicinity of the tumour may also present a similar intensity pattern, being hypointense on T1-weighted images, hyperintense on T2-weighted images, and showing no contrast enhancement. However, it is usually represented as a pencil-shaped intramedullary lesion with blurry margins. A diffuse expansion of the cord without any symptomatic implication has occasionally been encountered with these lesions that usually subsides over a period of time after the tumour has been removed and may be related to both venous congestion or edema, or the effect of growth potentiators produced by the tumour.
The presence of multiple mural nodules (seen in 2 patients as multiple enhancing nodules lining the tumour cyst capsule); and, the adherence of the tumour to the brain-stem (determined radiologically by the lack of demarcation of the enhancing tumour from the adjacent brain-stem tissue, the absence of an arachnoidal or a peritumoural cystic plane all along the circumference of the tumour, and the presence of brain-stem edema in the vicinity of the solid component of the tumour) may be responsible for the high incidence of symptomatic recurrence/regrowth of even smaller hemangioblastomas following their primary surgical excision.,, These radiological findings must be assiduously sought after during the surgical planning for these tumours.
One of our patients presented with an intratumoural bleed within the brain-stem hemangioblastoma. This is an extreme rare phenomenon and there is recorded 0.0024 risk of spontaneous hemorrhage/person/year; and the mean diameter of the tumours that bled was usually >3 cm. During surgery, however, a small aneurysm of the tonsillomedullary branch of PICA in close vicinity to the tumour was also detected and was clipped. We are not sure whether the intatumoural haemorrhage or the aneurysmal rupture was responsible for the intraventricular and subarachnoid bleeding that was encountered in our patient.
Risk stratification based upon location of tumour
18 (66.7%) tumours in our study were subpial and the rest were intramedullary. Based upon the site of tumour, we found a progressive increase in the degree of difficulty encountered in excising the tumour. Thus, the grouping of tumours based upon the progressive degree of difficulty encountered included: Type A: Purely cystic subpial; Type B: Purely cystic intramedullary; Type C: Small dorsal solid component with a large ventral cystic component between the tumour and brain-stem; Type D: Midline solid subpial; Type E: Solid subpial at the cerebellopontine angle; Type F: Midline solid intramedullary; and, Type G: Large solid intramedullary with significant arteriovenous shunts.
In subpial tumours, the tumours were surfacing just below the pial membrane at the floor of the fourth ventricle or at the cisterna magna. Hence, the surgeon was able to directly access the tumour. The plane of cleavage between the tumour and the brain-stem, if it existed in some parts of the tumour circumference, was either in the form of a gliotic one, as a peritumoural cyst, as a layer of CSF between the fourth ventricular floor and the tumour, or, as the outer thin wall of the syringomyelic cavity. 18 (66.7%) tumours were subpial, all of which were dorsally placed; 9 (33.3%) were intramedullary (one becoming exophytic in the cerebellopontine angle and three being anterolaterally placed). An approach to these latter lesions was fraught with the risk of injuring vital structures of the brain-stem. We approach these lesions from the part of maximum bulge on the surface of floor of the fourth ventricle, following 'the two point rule' proposed by Spetzler and his team, that is, one point is placed in the center of the lesion, and a second point is placed either where the lesion comes closest to a pial surface or at the safest entry point into the brain-stem, and the lesion is accessed in the corridor subtended by these two points. As the intramedullary lesion becomes more ventral, a considerable point of concern is to dissect it off the brain-stem tissue without compromising any of the normal brain-stem structures. Some of the important landmarks that help to define important neural structures at the floor of the fourth ventricle include: the median sulcus; the sulcus limitans, with the motor nuclei being medial and the sensory nuclei being lateral to it; the striae medullaris thalami, containing the cochlear fibres of the VIII nerve, bisecting the IVth ventricle in a horizontal plane; the median eminence, representing the facial colliculus and the VIth nerve nucleus; the calamus scriptorius representing the hypoglossal and vagal triangles, and the area postrema. The safe corridors for accessing the brain-stem, in case the intramedullary tumour is not surfacing at all, especially if it is more anterolaterally placed, include the following: A. The retro-olivary sulcus of the medulla. B. The peritrigeminal area in the pons with the medial limit being the pyramidal tract, the base being the pontomedullary sulcus and the area extending from the lateral aspect of the pyramid to the flocculus, with the trigeminal nerve entry zone forming the medial boundary and separating this zone from other vital pontine structures. C: In the midbrain, the lateral mesencephalic sulcus is situated between the crus cerebri and collicular plates. The lateral zone extends from the medial geniculate body superiorly to the pontomesencephalic sulcus inferiorly. This space is limited ventrally by the substantia nigra, dorsally by the medial leminiscus, and medially by the IIIrd cranial nerve. The anterior mesencephalic zone, on the other hand, is located between the oculomotor nerve and tract medially and corticospinal tract laterally. Focusing on these anatomical regions, and accessing the brain-stem through these corridors is recommended to considerably minimize the neurological deficits produced in this region.,,,
The technique of excision of brain-stem hemangioblastomas is similar to that of arteriovenous malformations. Meticulous attention must be paid to determining whether or not the artery is actually supplying the tumour or is an en passant artery to the brain-stem. If one coagulates the dilated tortuous draining veins leading to a large and vascular tumour at the beginning of the surgery, there will be odema and swelling of the tumour, often leading to torrential haemorrhage and loss of plane with the surrounding neuraxis. During surgery, if there is difficulty in identifying feeding arteries from draining veins, temporary aneurysm clips can be employed to assess the effect of the temporary interruption of veins that are conducting the venous drainage of the tumour. Once the arterial feeders are coagulated, there is shrinkage of the tumour. During handling of the tumour, one should take care that no traction is applied to the tumour as undue traction will injure the delicate brain-stem structures with which the tumour is intricately related. The surgical field should be kept as clean as possible without blood in it. This helps to visualize clearly the structures in this critical area. The tumour is often highly vascular so cutting through the capsule of the tumour in order to attempt to remove the tumour piecemeal will make the surgical field bloody and further dissection may become difficult. In small tumours, one may be able to resect the tumour en bloc but this maneuver is difficult in larger tumours. In large tumours, coagulation of the feeding arteries as close to the tumour as possible, and surface coagulation of the tumour, will help to shrink the tumour and define the gliotic plane; and, once the spongy, vascular tumour becomes fibrotic and avascular, it may also be removed piecemeal.
Ma et al., advocated a 'half-cut of the coagulated feeding vessel' to avoid retraction and catastrophic bleeding; and, 'underwater coagulation' to minimize thermal damage to the surrounding structures. Preoperative embolization has been found to be helpful in solid giant tumours to prevent the risk of perfusion breakthrough once the feeding vessels to the tumor have been interrupted. We also carried out preoperative embolization of two large and vascular tumours. Although, this considerably decreased the tumour blood supply, the parasitized blood vessels responsible for the perfusion of the brain-stem that were also supplying a majority of the tumour, the difficulty in reaching the terminal and tortuous branches of the tumour vessels, and the multiple, rapid-flow, vascular nidi (that also prevent the embolic material from being retained within the tumour) often present in a tumour, prevented the complete elimination of the blood supply to the tumour using therapeutic embolization.
If no plane of cleavage is present between the tumour and the surrounding brain-stem parenchyma, then a thin shell of capsule should be left adherent to the brain-stem, especially in the region of dorsal medulla oblongata to avoid damaging the nuclei of bilateral vagal and hypoglossal nerves.
Implications of tumour occurring in association with the VHL disease
Six of our patients had the VHL disease. This is an autosomal dominant, neurocutaneous syndrome linked to the short arm of chromosome 3.VHL disease is classified into two types: Type I that may have the presence of a renal cell carcinoma, retinal angioma, pancreatic tumour, endolymphatic sac tumour, and central nervous system hemangioblastoma but without the presence of a pheochromocytoma; and, Type II that has three subtypes: A: showing the presence of a pheochromocytoma, retinal angioma, pancreatic tumour, endolymphatic sac tumour, and central nervous system hemangioblastomas but without a renal cell carcinoma; B: showing the pheochromocytoma along with the presence of renal cell carcinoma, retinal angioma, pancreatic tumour, endolymphatic sac tumour, and central nervous system hemangioblastoma; and, C: showing only the presence of a pheochromocytoma without any other tumour being encountered. One of our patients with a brain-stem hemangioblastoma and renal cell carcinoma but without a pheochromocytoma, belonged to the type I group; while the other, who had a brain-stem hemangioblastoma in association with a pheochromocytoma but without a renal cell carcinoma belonged to the subgroup Type IIA. Two of our patients were also suffering from unexplained hypertension on serial blood pressure monitoring in whom a pheochromocytoma was suspected but not detectable. The very high incidence of abdominal masses and pheochromocytomas makes it imperative to conduct investigations for confirming the presence of these entities in every patient harbouring a hemangioblastoma.
One major issue in patients who have to undergo surgery for a brain-stem hemangioblastoma is the covert presence of a unilateral or bilateral pheochromocytoma; or, of adrenal medullary tissue at remote sites. Based upon the 2004 World Health Organisation (WHO), catecholamines may either be produced by an catecholamine-producing intra-adrenal tumour arising from the chromaffin cells (intra-adrenal paraganglioma); or, from related extra-adrenal sympathetic and parasympathetic paraganglia that are classified as extra-adrenal paragangliomas. Epinephrine (E), norepinephrine (NE), or dopamine may either be secreted by them in a continuous manner (leading to sustained hypertension) or in an episodic pattern (leading to paroxysms of symptoms). There may be hypertension-related end-organ damage, especially to the heart, kidney, eyes, and central nervous system. Diabetes mellitus may often be associated. Prior to the addressal of the brain-stem hemangioblastoma, the pheochromocytoma needs to be resected.
Within the chromaffin cells, NE and E are metabolized to normetanephrine and metanephrine. A rise in plasma and urine metanephrines is thus indicative of increased tumoural production of catecholamines. Measurement of urinary and plasma vanillylmandelic acid (VMA) and catecholamine levels can also be used for screening these tumours. Contrast enhanced CT scan, and MRI along with functional imaging using radiotracers such as 123 I-metaiodobenzylguanidine (123I-MIBG) and 111 In-DTPA-pentetreotide may be utilized to detect these masses. Preoperative preparation includes the prevention of hypertension induced by the alpha-1 mediated vasoconstriction (using non-selective blockers against both alpha-1 and alpha-2 adrenergic receptors such phenoxybenzamine, or selective alpha-1 receptor antagonist drugs such as prazosin, doxazosin, and terazosin); and, the prevention of beta-1 mediated tachycardia and ionotropy (that may be blocked by beta-blockers). Beta-blockers should never be used before the initiation of alpha-blockade in patients with functional tumours as suppression of beta-1 mediated cardiac sympathetic drive before adequate arteriolar dilatation can lead to acute cardiac insufficiency and pulmonary edema.
Calcium channel blockers (such as amlodipine, nicardipine, nifedipine, verapamil and diltiazem) may also be used as primary therapy for blood pressure control. They are especially useful in normotensive patients and in patients with paroxysmal hypertension with no elevation in the baseline blood pressure.
The alpha-1 receptor stimulation by catecholamines causes significant vasoconstriction and the initiation of alpha-blockade can lead to severe orthostatic hypotension. Thus, patients are advised to increase fluid and salt intake. Once the tumour is resected, the sudden withdrawal of catecholamines often leads to hypotension requiring fluid loading with vasopressor infusion. Sudden catecholamine withdrawal after tumour removal may also cause rebound hyperinsulinemia and severe hypoglycemia in the postoperative period. Thus, a close attention to the systemic effects produced is mandatory in order to successfully extirpate a brain-stem hemangioblastoma.
The major causes of morbidity in our study were exsanguinating intraoperative haemorrhage, spreading venous thrombosis to the brain-stem and the development of postoperative hematoma. The high flow arteriovenous shunt within some large and vascular tumours, the perfusion breakthrough due to a change of flow hemodynamics induced by the interruption of vascular supply to the tumour, spreading venous thrombosis, and the inability to unequivocally differentiate between a tumour vessel and an en passant vessel were the culpable factors. The perioperative manoeuvres discussed previously, including the use of preoperative embolization, had a significant effect in reducing this morbidity.
Perhaps the most important cause of long-term morbidity in these patients was the involvement of lower cranial nerves. A periodic assessment of the patients using the functional outcome swallowing scale (FOSS) described by Salassa is recommended. It has 5 stages, namely, stage 0 = normal function and asymptomatic; stage I = normal function but with episodic or daily symptoms of dysphagia; stage II = compensated abnormal function manifested by significant dietary modifications or prolonged mealtime (without weight loss or aspiration); stage III = decompensated abnormal function with weight loss of 10% or less of body weight over 6 months due to dysphagia, or daily cough, gagging, or aspiration during meals; stage IV = severely decompensated abnormal function with weight loss of more than 10% of body weight over 6 months due to dysphagia, or severe aspiration with bronchopulmonary complications, non-oral feeding recommended for most of nutrition, and stage V = non-oral feeding for all nutrition. A majority of our patients with a preoperative lower cranial nerve deficit were in the FOSS stage III and beyond, and therefore, required additional interventions following surgery such as postoperative overnight elective ventilation, elective tracheostomy and prolonged duration of Ryle's tube feeding., Regular evaluation of saliva swallowing, lower cranial nerve function and gag reflex and the exclusion of basal pulmonary crepitations prevented the occurrence of aspiration pneumonitis in most patients., A comparison of the present series with the reported literature clearly indicates the significant improvement that is achievable in the neurological status in patients following the successful extirpation of the brain-stem hemangioblastoma, despite the turbulent perioperative period that they have to bear up with [Table 4].,,,,,,,,,,
In conclusion, tumour coagulation with adequate irrigation promotes its shrinkage and improves the definition of its plane from the brain-stem. An attempt must be made to preserve the draining veins of the tumour until the majority of the tumour has been addressed. Leaving a small bit of the tumour capsule adherent to the brain-stem helps in preserving brain-stem function and in avoiding complications. Postoperatively, the patients should be assiduously monitored for their respiratory status as well as the gag reflex and lower cranial nerve function, to prevent the occurrence of aspiration pneumonitis.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9], [Figure 10], [Figure 11], [Figure 12], [Figure 13], [Figure 14], [Figure 15], [Figure 16], [Figure 17], [Figure 18], [Figure 19]
[Table 1], [Table 2], [Table 3], [Table 4]