Malignant idiopathic intracranial hypertension revealed a hidden primary spinal leptomeningeal medulloblastoma

Abstract

Frequently the cause of raised intracranial pressure remains unresolved and rarely is related to spinal tumours, moreover less to spinal medulloblastoma without primary brain focus. An 18-year-old woman had a 3-month history of headache and impaired vision. Neurological examination revealed bilateral sixth cranial nerve palsies with bilateral papilloedema of grade III. No focal brain or spine lesion was found on imaging. Consecutive lumbar punctures showed high opening pressure and subsequent increasing protein level. Meningeal biopsy was negative. At one point, she developed an increasing headache, vomiting and back pain. Spine MRI showed diffuse nodular leptomeningeal enhancement with the largest nodule at T6–T7. Malignant cells were detected in cerebrospinal fluid. She underwent laminectomy with excisional biopsy, and pathology showed medulloblastoma WHO grade IV. She was treated with chemotherapy and craniospinal irradiation and made a good recovery.

Background

We report an interesting and very scarce case of a young woman who had symptoms and signs of raised intracranial pressure (ICP), and despite extensive workup, no cause was found for 11 months until another spine MRI performed for back pain revealed nodular leptomeningeal enhancement that was proven to be a medulloblastoma.

Case presentation

An 18-year-old woman without prior medical conditions presented to a local hospital with a 3-month history of right leg weakness and left eye blurry vision associated with sudden severe headache along with nausea and vomiting. The headache was not related to the head positioning, but it was increasing over time. She sought medical advice at our hospital for the same persisting symptoms. Central nervous system (CNS) examination at the emergency department of our hospital, King Fahad Medical City in Riyadh, was positive for minimal restricted lateral eye movements bilaterally and bilateral papilloedema grade III, reduced vision in the periphery with poor visual acuity. No other focal signs were elicited with the rest of the physical examination unremarkable.

Investigations

The initial laboratory investigations, including complete blood count, electrolytes, liver function tests, coagulation profile and cardiac enzymes, were normal. Brain MRI with contrast showed bilateral papilloedema, bilateral enlargement of optic nerves sheath and Meckel’s cavum with bilateral symmetric severe stenosis of the distal transverse sinuses, and focal stenosis at the junction between the vein of Galen and the straight sinus (figure 1A–C). There was no leptomeningeal enhancement. Spine MRI with sagittal T2-weighted imaging was unremarkable (figure 1D, E). The ophthalmologist documented bilateral papilloedema grade II–III. She underwent extensive workup to look for secondary causes of the raised ICP. The first lumbar puncture (LP) showed high opening pressure of 27 cmH₂O and high protein, with the rest of the investigations within normal limit (see figure 2 for the cerebrospinal fluid (CSF) results). Visual evoked potentials showed a P100 wave on the left eye 131.7 and the right eye 123.6. CT of the chest and neck with contrast was negative for granulomatous lesions such as sarcoidosis or tuberculosis. Metabolic studies, including vitamin B₁₂, folate, vasculitis screening, serum ACE level, urine toxicology, and screening for the hypercoagulable state, were negative (see figure 3 with a table for details). Acetazolamide 250 mg two times per day per os was administered, and the dose gradually increased to 500 mg two times per day with partial improvement. Her headache subsided, and the vision slightly improved. The second LP done 2 weeks later showed an opening pressure of 21 cmH₂O with high protein and other investigations were normal (see figure 2 for the CSF results). Because of persistently elevated ICP, she was sent twice to a specialised eye hospital because of the absence of a neuro-ophthalmologist at our institution. She was documented to have bilateral papilloedema, worsening in the left side, with the recommendation to continue medical treatment, and no indication for optic nerve sheath fenestration. Thus, acetazolamide was increased to 500 mg three times per day, and furosemide 20 mg daily per os was added. The options to repeat another LP or to call the neurosurgery team for lumbar drain or ventriculoperitoneal (VP) shunt insertion were discussed with the patient, but she preferred another LP under fluoroscopy that showed high opening pressure of 37 cmH₂O, and a high protein level (see figure 2 for the CSF results). The case was discussed in a multidisciplinary team meeting with the recommendation to proceed for VP shunt insertion and consider dural biopsy to assess the possibility of inflammatory and or malignant conditions. The neurosurgery team did VP shunt and right frontal burr hole with dural biopsy. CSF analysis showed a high protein level. The histopathology of the dural biopsy showed unremarkable dura. Postoperative brain MRI and magnetic resonance venography showed similar findings of bilateral transverse and straight sinus/vein of Galen stenosis. Her headache had improved, and she was discharged home in good condition on acetazolamide 750 mg three times per day, furosemide 20 mg daily and analgesics as needed. During her outpatient care, she had eight emergency room visits for headaches, which sometimes were associated with nausea and vomiting, and multiple repeated CT of the brain revealed no new changes and stable, draining VP shunt in place. We treated her with analgesics with partial effect. Eleven months from the first symptoms and 8 months after the first admission to our hospital, she was admitted under the care of the neurosurgery team for evaluation of her persistent headaches, new back pain, lower limb numbness and abnormal movements suspected to be seizures with the impression of shunt malfunction. She underwent another LP showing high opening and high protein level (see figure 2 for the CSF results). CSF cytology was positive for malignant cells. A repeat MRI of the brain showed persistent bilateral optic neuritis/neuropathy associated with diffuse nodular leptomeningeal enhancement along the Sylvian fissures, cerebral sulci and cerebellar fissures (figure 4A, B). The spine MRI showed diffuse spinal enhancing nodules with the largest in the intradural extramedullary thoracic area at the level of (T6–T7) measuring 10×6×10 mm (figure 4C, D). After discussion in a multidisciplinary tumour board meeting, the recommendation was to resect and perform a biopsy of the spinal lesion. She underwent laminectomy with resection of the spinal lesion, and histopathology revealed a high-grade malignant round cell tumour. The immunohistochemistry showed synaptophysin positive, INI-1 retained, CD99 focal positive, and negative for MyoD1, GFAP and S100 (figure 5A, B). The final pathological diagnosis was medulloblastoma, WHO grade IV, classic type.

Figure 1.

Brain MRI. (A) Axial T2 fast spin echo showing prominent subarachnoid space around the optic nerves (empty white arrow), flattening of the posterior sclera (full black arrow) with intraocular protrusion of the optic nerve head, and tortuous optic nerves (full white arrows) with normal brain parenchyma. (B) Brain magnetic resonance venography (MRV) with contrast (posterior view) showing bilateral severe stenosis of distal transverse sinuses (black arrows) and (C) brain MRV with contrast (lateral view) showing bilateral severe stenosis of distal transverse sinuses, focal stenosis is noted at the junction between the vein of Galen and the straight sinus, absent inferior sagittal sinus (black arrows). (D) Whole spine MRI, sagittal T2 fast recovery fast spin echo (FRFSE) (upper) showing unremarkable spinal cord, no intramedullary or extramedullary focal lesion demonstrated and image (E) sagittal T2 FRFSE (lower) showing unremarkable spinal cord, no intramedullary or extramedullary focal lesion demonstrated.

Figure 2.

Laboratory data (CSF). AFB, acid-fast bacillus; CMV, cytomegalovirus; EBV, Epstein-Barr virus; HSV, herpes simplex virus; LP, lumbar puncture; MTB, Mycobacterium tuberculosis; RBC, red blood cell; VDRL, venereal disease research laboratory; WBC, white blood cell.

Figure 3.

Laboratory data. ANA, antinuclear antibody; ANCA, antineutrophil cytoplasmic antibody; APCR, activated protein C resistance; GCMS, gas chromatography/mass spectrometry.

Figure 4.

(A) Brain MRI. Axial LAVA (liver acquisition with volume acceleration) with contrast showing tortuous optic nerves (white arrows) with prominent subarachnoid space, signs of increased intracranial pressure and leptomeningeal enhancement (black arrows). (B) Brain MRI. Axial LAVA with contrast showing diffuse leptomeningeal enhancement along the cerebral sulci (black arrows). (C) Spine MRI. Sagittal T1 with contrast showing diffuse spinal enhancing nodules (black arrows). (D) Spine MRI. Sagittal T1 with contrast showing diffuse spinal enhancing nodules with the largest in intradural extramedullary thoracic area at the level of (T6–T7) measuring 10×6×10 mm (black arrow).

Figure 5.

(A) Pathology slide picture from spinal lesion biopsy. Diffuse infiltrate by small blue cells, H&E stain 20×. (B) Pathology slide picture from spinal lesion biopsy. Diffuse synaptophysin positivity seen, 20× hpf.

Differential diagnosis

The patient had symptoms and signs of raised ICP, and extensive workup revealed no cause in the first 11 months, refractory to medical and surgical treatment; hence, we called it ‘malignant’ idiopathic intracranial hypertension. No brain mass lesion or hydrocephalus was found to suggest as a culprit for the raised ICP. The proper imaging ruled out cerebral venous thrombosis; though venous sinuses were severely stenotic, we felt that this was secondary to raised ICP and not vice versa. There was no history of illicit substance use to suggest intoxication, and also urine toxicology was negative. Proper laboratory investigations and imaging ruled out rheumatological disorders, malignancy and sarcoidosis.

Treatment

Her case was discussed in the neuro-oncology tumour board meeting and it was decided to treat her with craniospinal radiation with concurrent chemoradiation followed by chemotherapy. The radiation oncology team arranged for CT simulation, external beam radiation therapy (EBRT) 54 Gy/30 fractions (craniospinal 36 Gy/20 fractions+18 Gy/10 fractions cranial boost+3.6 Gy/2 fractions spinal boost). The patient was also seen in the oncology clinic, and weekly vincristine chemotherapy was given concurrently with the sessions of craniospinal radiotherapy (RT) and later started on adjuvant chemotherapy with cisplatin, vincristine and lomustine every 6 weeks, and completed a total of eight cycles with good tolerance and minimal side effects in the form of nausea, mucositis, and grade I peripheral neuropathy. Goserelin injection monthly for ovarian prophylaxis during the chemotherapy course was given.

Outcome and follow-up

During her regular follow-ups, she had remarkable improvement in her headaches and vision. However, she had residual sixth cranial nerve palsy. A repeat MRI of the brain and spine at the end of her chemotherapy showed complete resolution of the intracranial leptomeningeal enhancements, with small residual seeding at the lumbar spine and along the cauda equina nerve roots. There were no signs of local recurrence at the surgical bed. She was continued on active surveillance with a repeat MRI of the brain and spine every 3–6 months. Her disease remained under control for 10 months, and after that, she developed minimal dysarthria, dysphonia and gait ataxia. A follow-up MRI of the brain and spine, 6 months after the last chemotherapy cycle, showed stable spinal cord lesion (figure 6A) and thickening of cauda equina roots (figure 6B) with new leptomeningeal seedings in the medulla (figure 6C) and left cerebellar area (figure 6D). Her case was discussed in the tumour board meeting, and it was concluded that the changes might represent the recurrence of her disease versus post-treatment changes. It was recommended that short-term follow-up with brain and spine MRI be considered and second-line chemotherapy if there is further disease progression. She is still being followed up in neurology, neurosurgery, medical oncology and radiation oncology clinics 3 years after the initial presentation.

Figure 6.

(A) Follow-up spine MRI. Sagittal T1 with contrast showing postoperative changes seen at the level of T7 with focal cord myelomalacia (black arrow). (B) Follow-up spine MRI. Sagittal T1 with contrast showing thickening and clumping of the cauda equina nerve roots (black arrows), but without significant enhancement and stable residual nodular enhancement is again seen at the caudal end of the thecal sac (white arrow). (C) Brain MRI. Axial LAVA (liver acquisition with volume acceleration) with contrast showing enhancing lesion in the medulla oblongata (black arrow). (D) Brain MRI. Axial LAVA with contrast showing enhancing nodules over the left cerebellar hemisphere (white arrows).

Discussion

ICP is stabilised by homeostatic mechanisms and normally maintained less than 15 mm Hg (20 cmH₂O). Raised ICP of more than 20 mm Hg is considered pathological and represents a common complication of neurological injury.¹ The major causes of increased ICP include: intracranial mass lesions, cerebral oedema, increased CSF production or decreased absorption, obstructive hydrocephalus, obstruction of venous outflow and idiopathic intracranial hypertension (pseudotumor cerebri).² It is quite intriguing that after nearly 11 months of presentation with signs of raised ICP that a space-occupying lesion (SOL) was detected in the neuroaxis. So, what had caused the ICP to rise in the absence of any discernible SOL? To understand the pathophysiology of raised ICP in such a case, we must revisit our knowledge of the haemodynamic of CSF circulation. There have been numerous hypotheses to the mechanism of raised ICP in spinal seeding. According to the Monro-Kellie doctrine, our brain, vascular structures and the CSF are confined in a rigid skull where the volume is maintained fixed,³ whereby ICP reflects the volume of the three compartments. An increase in the volume of any one component will be remunerated by the removal of an equivalent amount of another; if it fails to occur then, ICP increases. In sound health, there are roughly 500 mL of CSF produced in a day. However, it is continually being reabsorbed, leaving only 125–150 mL at any one time. The reabsorption of CSF occurs from a pressure gradient between the arachnoid mater and the venous sinuses.⁴ It is the outpouchings of the arachnoid mater into the venous sinuses called Pacchioni’s granulation that, through its one-way valve system, allows the majority of CSF reabsorption. A smaller portion of CSF returns to the vascular system via the ependymal linings and sheaths of the cranial and spinal nerves, lymphatic vessels along the olfactory nerve through the cribriform plate also receive some amount of CSF. However, various conditions have been implicated in impeding this process, although there is no compelling evidence for any of the theories. Some believe that spinal cord tumours are capable of producing excess protein. These proteins then act as sludge on the arachnoid granulations impairing CSF absorption through them.⁵ Proteins are capable of increasing the viscosity of the CSF, which can contribute to the hindrance.^{6 7} The excess proteins may have a toxic effect on the meninges leading to leptomeningeal inflammation, which further deters CSF absorption.⁸ On the other hand, some experts postulate that spinal tumours generate chemicals which interfere with CSF reabsorption.⁶ The target chemicals include fibrinogen and inflammatory cytokine transforming growth factor-β (TGFβ) mainly.⁶ Fibrinogen is believed to get converted to fibrin in CSF. When the fibrin gets accumulated in excess in the basal cisterns, it promotes ballooning of the ventricles leading to hydrocephalus.⁹ Whereas if fibrins deposit in arachnoid villi, they lead to raised ICP without ventriculomegaly akin to pseudotumor cerebri syndrome.⁶ In contrast, TGFβ, which originates from primary vascular structures such as choroid plexus, is thought to be present in high platelet concentrations.⁶ They can proliferate leptomeningeal cells and create scarring both at the base of the brain and in the area of the arachnoid villi.^{5 10} Besides, many other theories have been implicated, including compromised spinal elastic reservoir of CSF,¹¹ compression of spinal venous plexus causing unfavourable transarachnoid villous hydrostatic pressure,¹² and lastly, neoplastic arachnoiditis¹³ from spinal seeding of the tumour.¹⁴ Our patient could have seeded the meninges with the tumour cells at the microscopic level very early in the disease course, causing arachnoiditis and pseudotumor cerebri-like syndrome. Microscopic seeding has been described in the past in a case of schwannoma, which was detected only at autopsy. A recent case report describes a case similar to ours, where no primary CNS medulloblastoma was found, yet a cauda equina tumour proved to be a medulloblastoma.¹⁵ Medulloblastoma is a common malignant brain tumour of children, with approximately 70% of the patients diagnosed before their 20s. It occurs primarily in the cerebellum and most commonly presents with symptoms of increased ICP, including morning headache, nausea, vomiting and altered mental status, and in some cases, it is associated with risk of CSF seeding and spinal metastasis.^{16 17} The primary treatment of intracranial medulloblastoma involves multimodality therapy consisting of surgical resection followed by RT and chemotherapy.¹⁸ The addition of chemotherapy is considered in infant patients below the age of 3 years, aiming to delay or avoid the need for RT to allow further development of the nervous system. Chemotherapy is also recommended for patients with high-risk features of recurrence, which include residual tumour more than 1.5 cm², presence of metastatic disease, CSF seeding and large cell/anaplastic histology.¹⁹ The multimodality approach has improved the survival outcomes for patients with average-risk medulloblastoma.²⁰ The recent advances in gene expression profiling using tumour tissue microarrays have identified four distinct molecular subgroups of medulloblastoma with diverse clinical behaviours, prognosis and clinical outcomes. The most common molecular subtype is group 4, which accounts for 35% of medulloblastoma cases and has an intermediate prognosis with 5-year survival of 75%. The Wnt-activated medulloblastoma accounts for 10% of cases and has the best prognosis out of the four groups, with a 5-year survival of over 95%. SHH-activated medulloblastoma represents approximately 30% of cases and has an intermediate prognosis with 5-year survival ranging between 60% and 80%. Group 3 medulloblastoma cases have the worst prognosis of the four subgroups with a high likelihood of metastatic recurrence and 5-year survival of approximately 50%.^{21 22} For patients who are planned for RT, the dose of radiation depends on the risk stratification. The average risk disease is treated with either a reduced dose of 23.4 Gy of craniospinal RT or a conventional dose of 30–36 Gy; both are followed by a primary brain site boost for a total RT dose of 54–55.8 Gy to the tumour bed with or without concurrent weekly vincristine. The high-risk disease is treated with 36 Gy of craniospinal RT and boosting the primary brain site to 54–55.8 Gy, which is usually given with concurrent weekly vincristine.¹⁹ Following the RT course, it is recommended for patients with medulloblastoma to complete eight cycles of adjuvant multiagent chemotherapy. There are two commonly used regimens. The first one consists of cisplatin, vincristine and cyclophosphamide. The other one consists of cisplatin, vincristine and lomustine with similar efficacy.¹⁸ Our patient had a high-risk medulloblastoma with extensive CSF seeding metastases and leptomeningeal disease with a large spinal enhancing nodule at the thoracic T6–T7 level representing seeding metastasis. The spine lesion had caused the symptoms of increased ICP, which have improved after the treatment of the underlying cause (the primary tumour) with multimodality therapy.

Learning points.

• Primary spinal leptomeningeal medulloblastoma is extremely rare, especially without primary brain focus, but may cause increased intracranial pressure, even in the early microscopic phases, and it should be considered in the differential diagnosis if conventional and aggressive treatment of idiopathic intracranial hypertension fails.

• We assume that arachnoiditis from tumour seeding caused increased intracranial pressure.

• Appropriate neurosurgical intervention and surgical biopsy are mandated if a suspicious lesion is detected.

• Consider proper rescreening of the whole neuroaxis in refractory cases of intracranial hypertension.

• We assume that tumour resection, along with chemotherapy and radiotherapy, improved our patient’s symptoms.

Ethics statements

Patient consent for publication

Obtained.