—  SPECIALTY CONFERENCE  —

Neuropathology

Case 5 - Large Cell/Anaplastic Melanotic Medullomyoblastoma

Mark Edgar
Memorial Kettering Cancer Center
New York, NY


Click on each slide thumbnail image for an enlarged view
Clinical Summary
This two-year and two-month-old previously healthy boy presented with a one-month history of intermittent, progressive vomiting followed by gait ataxia. His pediatrician ordered a cranial MRI that revealed a 4 cm diameter, heterogeneously enhancing posterior fossa mass, probably arising from the cerebellar vermis. Subtle enhancement was also present overlying the medulla, raising the possibility of leptomeningeal disease. The patient was taken to surgery and gross total resection was achieved.


Case 5 - Figure 1 - Sagittal T1-weighted post-contrast MR image showing fourth ventricular mass.
Case 5 - Figure 2 - Axial T1-weighted post-contrast (left) and T2-weighted (right) MR images showing fourth ventricular mass.
Case 5 - Figure 3 - Low power view of tumor showing sheets of round cells.



Case 5 - Figure 4 - Higher power view showing round cells with macronucleoli.
Case 5 - Figure 5 - Intratumoral "apoptotic lakes".
Case 5 - Figure 6 - Tubular epithelial structures in one part of the tumor: note the intracytoplasmic finely granular pigment.



Case 5 - Figure 7 - Closer view of pigmented cells.
Case 5 - Figure 8 - High-power view of pigmented cells.
Case 5 - Figure 9 - Collections of cells with abundant eosinophilic cytoplasm consistent with rhabdomyoblasts.



Case 5 - Figure 10 - Desmin immunostain.
Case 5 - Figure 11 - Cytokeratin (AE1/AE3) immunostain.
Case 5 - Figure 12 - Myogenin immunostain.


Diagnosis
Large Cell/Anaplastic Melanotic Medullomyoblastoma

Discussion

Summary of the case
Patient age (2 years) and tumor location (posterior fossa, with involvement of cerebellar vermis) are typical for medulloblastoma, and the histologic appearance of a small blue cell tumor is characteristic. However, the presence of prominent nucleoli and collections of apoptotic figures indicate large cell variant of medulloblastoma. Furthermore, the presence of frankly epithelial foci with well-formed tubules composed of cells containing cytoplasmic melanin pigment is typical of melanotic medulloblastoma. Finally, some sections showed convincing rhabdomyoblasts; these and small blue cell areas where skeletal muscle differentiation was not evident histologically showed immunopositivity for desmin, muscle specific actin and myogenin (i.e. medullomyoblastoma).

Discussion
Although medulloblastoma is generally a straightforward diagnosis when the pathologist is confronted with a small blue cell tumor of the cerebellum, this diagnostic category has become more complicated over the past 15 years as clinicopathologic studies have uncovered histologic features with prognostic significance. The result is that this aggressive, WHO grade IV neoplasm can (and should) be examined for features which portend an especially poor prognosis with unresponsiveness to therapy. Although it might seem superfluous to histologically substratify an aggressive small blue cell tumor for prognostic purposes, an analogous situation exists for other pediatric embryonal tumors such as Wilms tumor and neuroblastoma where anaplasia and Shimada classification, respectively, are important predictors of outcome.

In 1992, Giangaspero et. al. published four cases of large cell medulloblastoma characterized by tumor cells with large, vesicular nuclei, prominent nucleoli, and strong expression of synaptophysin; c-myc amplification was demonstrated in one case [1]. All four cases pursued an aggressive course with early CSF dissemination despite radiation and chemotherapy. Perhaps because of thoughts that this tumor might be a variant of AT/RT, little else was published on the topic until 2000 when a Pediatric Oncology Group study suggested a special category of "large cell/anaplastic" medulloblastoma [2]. This designation placed large cell medulloblastoma (as described in 1992) and anaplastic tumors showing large, irregularly shaped nuclei with coarse chromatin together in a single, prognostically adverse category. Both large cell and anaplastic types frequently showed large collections of apoptotic bodies ("apoptotic lakes"), increased mitotic activity, and a tendency for tumor cells to wrap around one another. This study confirmed the earlier finding of c-myc amplification and found evidence of genetic changes typical of medulloblastoma (isochromosome 17q) but not AT/RT (monosomy 22). The study concluded that LC/A medulloblastoma was separable from AT/RT but acknowledged that the tumors shared some histologic features and an aggressive clinical course with a tendency to early CSF dissemination.

Since the POG study, several series have confirmed the significance of the large cell/anaplastic (LC/A) category for medulloblastoma and presented evidence that LC/A tumors evolve from classic medulloblastoma [3, 4]. One of these studies examined the importance of degrees and extent of histologic anaplasia, finding that moderate or severe anaplasia was prognostically significant and that tumors with diffuse anaplasia (defined as anaplasia present in all low power fields) were associated with a poorer outcome than those showing focal anaplasia [3]. Large cell tumors are considered prognostically equivalent to medulloblastomas with severe anaplasia. Probably because the criteria for anaplasia are subjective, the frequency of the LC/A category among all medulloblastomas has varied from 4 to 24% in childhood medulloblastoma. Although histologic anaplasia has not yet been incorporated into criteria for risk stratification for pediatric medulloblastoma, there is considerable discussion of the topic among pediatric neuro-oncologists. Interestingly, a recent report suggests that anaplasia is a rare finding in adults where it does not influence prognosis [5].

A second striking feature of our case is a focus of well formed tubules; while the presence of glandular elements in a small blue cell tumor of the cerebellum might raise the possibility of AT/RT, these tubules show cytoplasmic brown, finely granular pigment consistent with melanin. This appearance fits with the "malignant melanin-forming tumor of the cerebellum" originally described by Fowler and Simpson [6] in 1962 which has come to be known as melanotic medulloblastoma. The pigment is oculocutaneous melanin (as opposed to neuromelanin), and melanosomes have been demonstrated by electron microscopy. It is worth noting that the WHO 2000 makes a somewhat more inclusive definition of melanotic medulloblastoma: "A rare tumor of childhood with a predominant element of small round cells closely resembling classic medulloblastoma and a minor component of melanin-forming neuroepithelial cells" [7]. This definition therefore accepts as melanotic medulloblastoma a subset of tumors which demonstrate melanin in the cytoplasm of small round cells rather than in tubulopapillary structures [8, 9]; the relationship of these tumors to melanotic medulloblastoma with an epithelial component is not clear. Tubules typically stain for cytokeratin and may show positivity for S-100 protein. Although melanotic medulloblastoma is very rare with less than 20 reported cases, examples with rhabdomyoblasts have been described [10].

Although melanotic medulloblastoma is too rare a tumor to carry out survival analysis on series of cases, virtually all reported cases of the classic tubulopapillary variety have behaved aggressively, most seeding the CSF and causing death [11, 12]. And while it is difficult to assess these cases retrospectively for evidence of LC/A features, some describe or illustrate large cells with prominent nucleoli suggesting large cell elements [6, 12].

Although they are not well represented in the slides we distributed for the conference, our case also showed rhabdomyoblasts which stained for muscle-specific actin, desmin, and myogenin: this third feature of the tumor leads us to the diagnosis of medullomyoblastoma. Pure medullomyoblastoma (first described by Marinesco and Goldstein in 1933) is a cerebellar small blue cell tumor with histologic evidence of striated muscle elements [13]. It should be noted that classic medulloblastoma not uncommonly expresses desmin in the absence of rhabdomyoblasts by histology, and these cases are generally not considered medullomyoblastomas. There is a 3 or 4:1 male predominance, and most patients are children. Unlike melanotic medulloblastoma, medullomyoblastoma is not clearly more aggressive than classic medulloblastoma. For instance, a series of 6 medullomyoblastomas (five of which showed LC/A histology) found survival similar to that for classic medulloblastoma with median follow up of 92 months [14]. This study also found that biphasic nodularity on imaging studies correlated with discrete islands of small blue cells and rhabdomyoblasts by histology.

The histogenesis of medulloblastomas with rhabdomyoblasts and pigmented epithelium has been debated. Currently, there is no reason to suppose that these components represent more than an example of divergent differentiation characteristic of PNET in general. However, a neural crest origin has been suggested, specifically for medullomyoblastoma. Ectomesenchyme is a neural crest (i.e. ectodermal) derivative which in some parts of the body (like the head and neck) gives rise to skeletal muscle. Portions of the leptomeninges are thought to be derived from ectomesenchyme, and this may explain the occasional presence of heterotopic meningeal skeletal muscle and rhabdomyosarcoma.

Given the presence of epithelial and rhabdoid elements, differential diagnosis in this case should include AT/RT. AT/RT frequently includes a small blue cell component indistinguishable from medulloblastoma, but the typical large cells with pale, vesicular nuclei, eosinophilic cytoplasm and prominent cytoplasmic vacuolation/ballooning are absent. Genetic studies are also helpful: in this case, sequencing of the INI1 gene did not demonstrate a mutation, making AT/RT unlikely. Cytogenetic or FISH studies for 22q abnormalities would also be helpful in excluding AT/RT. Immunostaining for the INI1 gene product is a readily available technique which essentially excludes AT/RT if the tumor shows nuclear staining.

Pigmented epithelial elements, synaptophysin, cytokeratin, and S-100 staining can all be seen in melanotic neuroectodermal tumor of infancy (melanotic progonoma, retinal anlage tumor). This tumor (which generally affects the craniofacial region, rarely the brain) lacks the histologic anaplasia evident in our case, and shows a more orderly architecture with pigmented epithelial cells forming a peripheral rim around nests of small blue, neuroblastic cells. Rhabdomyoblasts are not expected in MNTI.

Other Types of Medulloblastoma
The terms nodular and desmoplastic medulloblastoma are often used interchangeably in reference to those tumors with a combination of pale islands (which resemble lymphoid follicles with germinal centers) made up of large cells with advanced neuronal/neurocytic differentiation and internodular desmoplastic zones in which cells typical of classic medulloblastoma grow in linear arrays in a collagenous background. Such tumors have been described in association with Gorlin's syndrome (nevoid basal cell carcinoma syndrome, caused by mutations in PTCH gene) and there is conflicting evidence about whether this histologic pattern is associated with a superior prognosis. However, there are examples of medulloblastoma in which nodularity is extensive and in which nodules show cells with advanced neurocytic differentiation, rare mitoses, and prominent neuropil. Such tumors generally occur in infants, often show a peculiar grapelike appearance on neuroimaging, and are associated with a favorable prognosis [15]. The term medulloblastoma with extensive nodularity is currently favored over the older "cerebellar neuroblastoma" for such tumors.

Cerebellar liponeurocytoma is the name currently given to cerebellar tumors (formerly "lipomatous medulloblastoma" or "medullocytoma") of adulthood which combine advanced neurocytic and adipocytic (and even myogenic) differentiation with low proliferative potential and favorable prognosis [16]. These terms are best separated from classic medulloblastoma (to which their genetic relationship is unclear) in order to avoid unnecessarily aggressive treatment.

References

  1. Giangaspero F, Rigobello L, Badiali et. al. Large cell medulloblastomas. A distinct variant with highly aggressive behavior. Am J Surg Pathol 16(7):687-693, 1992.

  2. Brown HG, Kepner JL, Perlman EJ et. al. "Large cell/anaplastic" medulloblastomas: A Pediatric Oncology Group study. J Neuropathol Exp Neurol 59(10):857-865, 2000.

  3. Eberhart CG, Kepner JL, Goldthwaite PT et. al. Histopathologic grading of medulloblastomas. A Pediatric Oncology Group Study. Cancer 94:552-60, 2002.

  4. Leonard JR, Cai DX, Rivet DJ et. al. Large cell/anaplastic medulloblastomas and medullomyoblastomas: clinicopathological and genetic features. J Neurosurg 95:82-88, 2001.

  5. Giordana MT, D'Agostino C, Pollo B et. al. Anaplasia is rare and does not influence prognosis in adult medulloblastoma. J Neuropathol Exp Neurol 64(10):869-874, 2005.

  6. Fowler M, Simpson DA. A malignant melanin-forming tumor of the cerebellum. J Path Bact 84:307-311, 1962.

  7. Kleihues P, Cavenee WK. Tumors of the Central Nervous System. Pathology and Genetics. Lyon: IARC Press, 2000.

  8. Jimenez CL, Carpenter BF, Robb IA. Melanotic cerebellar tumor. Ultrastruct Pathol 11:751-759, 1987.

  9. Kalimo H, Paljarvi L, Ekfors T, Pelliniemi LJ. Pigmented primitive neuroectodermal tumor with multipotential differentiation in cerebellum (pigmented medullomyoblastoma). Pediat Neurosci 13:188-195, 1987.

  10. Banerjee AK, Kak JK. Teratoid tumor of the cerebellum. J Path 111:285-287, 1973.

  11. Sung JH, Mastri AR, Segal EL. Melanotic medulloblastoma of the cerebellum. J Neuropathol Exp Neurol 32:437-445, 1973.

  12. Dolman CL. Melanotic medulloblastoma. A case report with immunohistochemical and ultrastructural examination. Acta Neuropathol 76:528-531, 1988.

  13. Marinesco G, Goldstein M. Sur une forme anatomique, non encore decrite, de medulloblastoma: meddulo-myoblastome. Ann Anat Pathol 10:513-525, 1933.

  14. Helton KJ, Fouladi M, Boop FA, et. al. Medullomyoblastoma: A radiographic and clinicopathologic analysis of six cases and review of the literature. Cancer 2004 101:1445-1454, 2004.

  15. Giangaspero F, Perilongo P, Fondelli MP et. al. Medulloblastoma with extensive nodularity: a variant with favorable prognosis. J Neurosurg 91:971-977, 1999.

  16. Giangaspero F, Cenacchi G, Roncaroli F et. al. Medullocytoma (lipidized medulloblastoma): a cerebellar neoplasm of adults with favorable prognosis. Am J Surg Pathol 20:656-664, 1996.