—  SPECIALTY CONFERENCE  —

Neuropathology

Case 5 - Primary CNS Melanoma

Mark L. Cohen
Case Western Reserve University
Cleveland, OH


Click on each slide thumbnail image for an enlarged view
Clinical History
This 25-year-old male presented to his physician with a chief complaint of headaches and double vision. He had a two-month history of occipital headaches of progressive severity and frequency. These headaches began soon after a minor automobile accident and eventually became generalized. The patient also complained of double vision on right gaze with left eyelid twitching. The patient had a long history of left eye "swinging in and out" for which he wore corrective lenses since age 2. Family history was notable for "multiple moles and skin lesions" in first-degree relatives. Physical exam revealed bilateral papilledema, multiple moles and skin lesions, and bilateral sixth nerve palsies. A CT scan of the head was performed and demonstrated a left frontal lobe mass. Removal of the lesion accompanied by "left frontal lobectomy" was performed.


Case 5 - Figure 1 - LLP. Low power view demonstrating relationship of the tumor to the adjacent brain parenchyma.
Case 5 - Figure 2 - HP. The tumor is composed of epithelioid cells, many with prominent nucleoli and/or nuclear pseudoinclusions.



Case 5 - Figure 3 - Mitosis. Scattered mitotic figures are present.
Case 5 - Figure 4 - Infil. The tumor infiltrates along Virchow-Robin spaces.

Diagnosis: Primary CNS melanoma

Discussion
On the basis of a negative PTAH stain, this patient was originally given a diagnosis of "poorly differentiated sarcoma", underwent a course of radiation therapy, and was lost to follow-up for approximately three decades. PTAH histochemical staining was the method of choice for demonstrating astrocytic differentiation prior to the development of GFAP antibodies in the 1970s, and hence was state-of-the-art when this patient was operated on in 1968. The patient returned to visit his initial surgeon's successor in the spring of 2000 complaining of problems related to his left visual field. Neuroimaging studies and brain biopsy showed features consistent with radiation induced changes, but no evidence of tumor recurrence.

Reevaluation of the initial resection specimen (this time with GFAP antibodies) confirmed the absence of astrocytic differentiation. Both S100 protein and HMB-45 antibodies reacted with the tumor cells, and focal pigment was identified, consistent with melanin (Berlin blue negative, Fontana-Masson positive). The tumor is composed of sheets of predominantly amelanotic epithelioid cells and shows CNS invasion along Virchow-Robin spaces. Many of the tumor cells contain intranuclear pseudoinclusions and/or prominent nucleoli. Scattered mitotic figures (less than 1 per 10 high-power fields) are present, but coagulative tumor necrosis is inconspicuous. These features are consistent with primary CNS melanoma.

Three years later (age 58, 33 years since diagnosis), he returned to the hospital with right hemiparesis, speech problems, and progressive decline in mental status. CT scan showed a large intracerebral hemorrhage with edema suspicious for tumor. The patient expired soon after. At autopsy, cutaneous nevi were inconspicuous. Bilateral necrotizing bronchopneumonia was present. There was no evidence of systemic melanoma. There was a well-healed surgical cavity in the left frontal lobe, without evidence of tumor recurrence. Leptomeningeal melanosis was not identified. The left temporal lobe contained a 6 X 6 X 6 cm. glioblastoma multiforme.

Even in the absence of melanin, the differential diagnosis of an epithelioid neoplasm with abundant granular eosinophilic cytoplasm, intranuclear pseudoinclusions, and prominent nucleoli revolves around malignant melanoma. Within the brain parenchyma, where this tumor appeared to have arisen, the differential diagnosis is virtually confined to the two unlikely possibilities of epithelioid glioblastoma multiforme (which would be GFAP positive) [1] and the pure rhabdoid form of atypical teratoid/rhabdoid tumor (a highly aggressive tumor seen almost exclusively in the first years of life) [2]. Among meningeal tumors, several meningioma subtypes have been described that contain epithelioid cells with abundant cytoplasm, including secretory, rhabdoid, and oncocytic variants [3, 4, 5] . Indeed, at the time that this patient had his initial tumor resection, melanocytic meningeal tumors were considered to be melanotic meningiomas. Recognition of these tumors as being derived from pial melanocytes, rather than meningothelial cells, began with electron microscopic studies in the early 1970's[6], and continued with immunohistochemical and molecular studies over the past two decades[7]. These melanocytic neoplasms react with all the usual melanocyte markers (S100 protein, HMB-45, etc.) but not with antibodies to EMA. Once the melanocytic nature of the tumor is established, the problem then becomes one of trying to predict the biologic behavior of primary CNS melanocytic lesions, and on the malignant end of this spectrum, distinguishing primary from metastatic malignant melanomas.

In the absence of three decades of follow-up followed by a post-mortem examination, it would be impossible to definitively diagnose a melanotic tumor with the histopathologic features seen in this case as primary to the nervous system in the absence of congenital cutaneous stigmata of neurocutaneous melanosis. Though this patient has a personal and family history of "moles and skin lesions", these were never evaluated further, either during life or at autopsy. Giant congenital nevi, seen in ~2/3 of patients with neurocutaneous melanosis were not present, nor was there evidence of leptomeningeal melanosis at autopsy. In addition to these pertinent negatives, the familial nature of the skin lesions is NOT in keeping with the presence of neurocutaneous melanosis, which is classified as a congenital, but non-hereditable, phakomatosis [8].

A recent study of 33 primary melanotic tumors of the central nervous system occurring in the absence of evidence of neurocutaneous stigmata found that such tumors can usually be separated into benign melanocytomas and malignant melanomas, and confirmed previous reports that primary CNS melanoma, in the absence of diffuse leptomeningeal involvement, carries a much more favorable prognosis than metastatic melanoma involving the nervous system [9]. Thirteen of the 33 primary melanotic tumors were considered malignant on the basis of histologic and cytologic features seen within this patient's tumor (epithelioid cells growing in sheets, nuclear pleomorphism, prominent nucleoli, mitotic activity, CNS invasion, and [not seen in this case] necrosis). These patients ranged from 15 to 71 years of age at diagnosis. Three of these were supratentorial, with one occurring in the frontal lobe. The remaining 10 tumors were equally divided between the posterior fossa and spinal cord. After resection, 8 of the 13 tumors recurred, with four being fatal. However, only one of 5 totally resected melanomas recurred during follow-up periods ranging from 14 months to three years. Furthermore, only one of these patients received post-operative radiotherapy.

Seventeen of the tumors demonstrated pathologic features of well-differentiated leptomeningeal melanocytomas (spindle shaped cells with monomorphic nuclei and indistinct nucleoli growing in nesting patterns, with very little or no mitotic activity and no necrosis or CNS invasion). These occurred in the same age range (17 to 73 years at diagnosis), with 2/3 involving the spinal leptomeninges. None recurred following total or subtotal resection (with or without postoperative radiotherapy).

Three tumors could not be definitively classified as either benign melanocytomas or primary malignant melanomas. These showed sheetlike growth, CNS invasion, and occasional mitoses, but had indistinct nucleoli and no necrosis. It was recommended that these tumors be diagnosed as "Melanocytic neoplasm, Intermediate grade".

The first case of transition from an intracranial melanocytoma to primary malignant melanoma was reported in September 2004 [10]. This patient was 25 years old at the time that a "fibrous melanocytic meningioma" (subsequently confirmed to be a meningeal melanocytoma) was removed from her petroclival area. Regular MR examinations showed no evidence of tumor progression until 12 years later, when brainstem compression was noted. Subtotal resection of the recurrent tumor demonstrated features consistent with the aforementioned melanocytic tumor, intermediate grade (rare mitoses, no necrosis). Clinical worsening with MR evidence of tumor progression led to a third operation seven months later, which disclosed a pleomorphic epithelioid mitotically active tumor consistent with primary malignant melanoma. Postoperatively, she developed diffuse leptomeningeal melanomatosis. Despite whole brain radiotherapy and chemotherapy, she died 4 months later.

Leptomeningeal dissemination of primary CNS malignant melanoma is uncommon in adults (who usually do not have neurocutaneous melanosis), but is common in children with neurocutaneous melanosis and CNS lesions [11]. These children may either present with a mass lesion followed by leptomeningeal dissemination or with diffuse leptomeningeal melanocytic proliferation in the absence of a mass lesion [12]. The latter group present clinically with signs and symptoms of increased intracranial pressure (headache, vomiting) accompanied by cranial nerve deficits. Pathologically, they may be divided into cytologically bland (melanocytosis) and cytologically malignant (melanomatosis) proliferations, though both groups appear to have an equally dismal prognosis (usually measured in months).

References

  1. Rosenblum MK, Erlandson RA, Budzilovich GN. The lipid-rich epithelioid glioblastoma. Am J Surg Pathol. 1991 Oct;15(10):925-34.
  2. Hilden JM, Meerbaum S, Burger P, Finlay J, Janss A, Scheithauer BW, Walter AW, Rorke LB, Biegel JA. Central nervous system atypical teratoid/rhabdoid tumor: results of therapy in children enrolled in a registry. J Clin Oncol. 2004 Jul 15;22(14):2877-84
  3. Colakoglu N, Demirtas E, Oktar N, Yuntem N, Islekel S, Ozdamar N. Secretory meningiomas. J Neurooncol. 2003 May;62(3):233-41.
  4. Perry A, Scheithauer BW, Stafford SL, Abell-Aleff PC, Meyer FB. "Rhabdoid" meningioma: an aggressive variant. Am J Surg Pathol. 1998 Dec;22(12):1482-90.
  5. Roncaroli F, Riccioni L, Cerati M, Capella C, Calbucci F, Trevisan C, Eusebi V. Oncocytic meningioma. Am J Surg Pathol. 1997 Apr;21(4):375-82.
  6. Limas C, Tio FO. Meningeal melanocytoma ("melanotic menin gioma"). Its melanocytic origin as revealed by electron microscopy. Cancer. 1972 Nov;30(5):1286-94.
  7. Ibanez J, Weil B, Ayala A, Jimenez A, Acedo C, Rodrigo I. Meningeal melanocytoma: case report and review of the literature. Histopathology. 1997 Jun;30(6):576-81.
  8. Di Rocco F, Sabatino G, Koutzoglou M, Battaglia D, Caldarelli M, Tamburrini G. Neurocutaneous melanosis. Childs Nerv Syst. 2004 Jan;20(1):23-8.
  9. Brat DJ, Giannini C, Scheithauer BW, Burger PC. Primary melanocytic neoplasms of the central nervous system. Am J Surg Pathol. 1999 Jul;23(7):745-54.
  10. Roser F, Nakamura M, Brandis A, Hans V, Vorkapic P, Samii M. Transition from meningeal melanocytoma to primary cerebral melanoma. Case report. J Neurosurg. 2004 Sep;101(3):528-31.
  11. Makin GW, Eden OB, Lashford LS, Moppett J, Gerrard MP, Davies HA, Powell CV, Campbell AN, Frances H. Leptomeningeal melanoma in childhood. Cancer. 1999 Sep 1;86(5):878-86.
  12. Reyes-Mugica M, Chou P, Byrd S, Ray V, Castelli M, Gattuso P, Gonzalez-Crussi F. Nevomelanocytic proliferations in the central nervous system of children. Cancer. 1993 Oct 1;72(7):2277-85.