Previous pageNext page

INDICATIONS FOR ELECTRON MICROSCOPY IN CENTRAL NERVOUS SYSTEM BIOPSIES

Lauren A. Langford
M.D. Anderson Cancer Center, Houston, TX

Classification of various cellular conformations of the CNS is difficult when trying to distinguish glial and neuronal elements using only light microscopy and immunohistochemistry. Although immunohistochemical techniques have challenged cell analysis by electron microscopy, electron microscopy provides structural information far beyond the presence or absence of a requested epitope. Electron microscopy often plays a pivotal role in diagnosis because it can provide accurate diagnosis when immunohistochemical studies are equivocal or negative. Diagnostic dilemmas posed by small biopsy specimens of CNS tumors are often the result of multiple factors, including fixation artifact, biopsy size, lack of immunohistochemical techniques to distinguish cell types, and on the part of the investigators, unawareness of rare entities.

In an ideal setting, CNS biopsies should be examined with full knowledge of the patient's clinical history, including age, sex, onset and type of symptoms, and previous biopsy and/or treatment. The pathologist should be informed of the precise location of the lesion or lesions and the imaging characteristics. The CNS is a complex collection of neuroanatomical structures and systems and not simply a single anatomic organ; therefore, careful inspection of biopsy tissues and clinicopathological correlation is necessary for accurate evaluation of biopsy specimens. Difficulties in diagnosis frequently arise in three categories: (1) biopsies containing clear cells, (2) in glial fibrillary acidic (GFA) protein-positive spindle cell lesions, and (3) in unsuspected metastases. The following notes consider differential diagnoses in these areas.


GROUP 1: NEOPLASMS CONTAINING CLEAR CELLS
central neurocytoma, oligodendroglioma, ependymoma, meningioma, metastatic clear cell carcinoma

Central neurocytoma is a neuronal neoplasm that was first described by Hassoun et al in 1982 (Acta Neuropathol. 1982;56:151-156). It typically occurs in an intraventricular location in young adults. Light microscopic examination reveals that central neurocytomas comprise cells that have uniform ovoid nuclei and frequent perinuclear halos. With immunohistochemical analysis, delicate cytoplasmic processes may be immunoreactive with antibodies to synaptophysin, but when sufficient autolysis artifact is present, distinguishing intracytoplasmic staining from entrapped neuropil may be difficult. A definitive diagnosis can be made with electron microscopy. Homer-Wright rosette formation may be the first clue to the neuronal nature of the neoplasm. Electron microscopic examination shows that central neurocytomas contain regular nuclei with dispersed heterochromatin and small quantities of cytoplasm containing ribosomal rosettes, short strands of rough endoplasmic reticulum, scattered mitochondria, dense core membrane-bound neurosecretory vesicles, and parallel arrays of tubules that form an axon. Synapses may be present but are rare. Features that help to distinguish central neurocytoma from cerebral neuroblastomas and medulloblastomas with neuronal differentiation are its characteristic location; round, uniform nuclei in contrast to the hyperchromatic, angular nuclei seen in neuroblastoma and medulloblastoma; low mitotic activity; and lack of necrosis. Central neurocytomas are presumed to be much less biologically aggressive than neuroblastomas or medulloblastomas.

Oligodendrogliomas constitute about 5 to 10% of the primary CNS glial neoplasms. Most are identified in adults in their fourth and fifth decades of life, but oligodendrogliomas can occur at any age. Their appearance on light microscopy is one of fairly compact collections of uniform cells with regular ovoid nuclei surrounded by halos. Cortical infiltration, perineuronai satellitosis, cystic changes, and mineralizations suggest the likelihood of an oligodendroglioma. Immunoreactivity with antibodies to G FA protein is positive in about 50% of the pure oligodendrogliomas. The perinuclear halos surrounding closely arranged round to oval nuclei combined with a characteristic honeycomb architectural tissue pattern provide the basis for making the diagnosis; therefore, in the absence of halos or in fragments of insufficient size to identify an architectural pattern, electron microscopy plays an important role in diagnosis. Ultrastructurally, oligodendrogliomas show a range of morphologies, but typically the neoplastic cells have moderate amounts of electron-lucent cytoplasm containing a few microtubules, large numbers of mitochondria, free ribosomes, glycogen particles, and small bundles of filaments. Some cells extend stubby cytoplasmic processes. Although the nuclei may be slightly irregular, nucleoli are small and nuclear borders are smooth with very few indentations. Clustering of atypical cells about neurons is frequent in contrast to the formation of concentric laminations of cytoplasmic processes that are rarely observed. Oligodendrogliomas often contain cells that have some degree of astrocytic differentiation as well as entrapped reactive astrocytes.

Ependymomas are relatively uncommon central nervous system neoplasms. Their estimated incidence is between 2 and 6% of all gliomas. They can occur at any age but occur mainly in children. Ependymomas are usually intimately associated with the ventricles with the fourth ventricle being the most common site and also occur in the spinal cord. Ependymomas express their dual glial and epithelial heritage by the production of GFA protein positive intracytoplasmic filaments and the formation of true rosettes. By light microscopy, the diagnostic features are the presence of ependymal rosettes or elongated clefts bordered by an ependymal epithelium. These features, however, are present only in a minority of cases. Perivascular pseudorosettes are suggestive, but not diagnostic of ependymoma, because they can also be present in astrocytomas. Clear cell ependymomas, often lack any ependymal features that can be distinguished by light microscopy and may not be immunoreactive with antibodies to GFA protein. These features lead to a differential diagnosis among central neurocytoma, oligodendroglioma, or meningioma, because of the uniform cellularity, prominent perinuclear hatos, and the association with the ventricles. The fine structure of clear cell ependymomas is like that of an ordinary ependymoma mimicking normal ependymal cells. The nuclei tend to be ovoid with dispersed chromatin. The cells contain intracytoplasmic intermediate filaments, usually in a perinuclear distribution, conspicuously long junctional complexes, numerous slender surface microvilli, and abnormal cilia. Intracytoplasmic lumina, containing variable numbers of microvilli and cilia are frequently present and occasionally must be distinguished from metastatic neoplasms.

Meningiomas are composed of atypical arachnoid cells, and vascular and fibrous tissue normally found in the meninges. Although the tumors may range in appearance from epithelial to mesenchymal they are characterized by a uniform distribution of cells with shapes ranging from polygonal epithelial-like to spindled and fusiform. Most of the time the cells are arranged in elongated sheets or islands separated by connective tissue trabeculae. A faint whorling pattern can almost always be seen. The nuclei are characteristically abnormal with pseudonuclear inclusions, or simpiy intranuclear intrusions of cytopiasm. Mineralizations, or psammoma bodies, are the birthright of all meningiomas and may be present or absent in any tumor. With standard immunohistochemical procedures, more than 95% of all meningiomas are immunopositive for antibodies to epithelial membrane antigen (EMA). About one-third are immunopositive with antibodies to cytokeratins and another third are immunoreactive with antibodies to S-100 protein. Occasionally, some meningiomas will show immunopositivity to carcinoembryonic antigen (CEA). Despite the variation in light microscopic appearance, electron microscopic examination shows similar features including interdigitations of cell membranes, intracytoplasmic filaments, well-formed desmosomal attachments, and extracellular collagen fibers.

Metastasis of a clear cell carcinoma may be the presenting manifestation of unsuspected primary carcinoma. The documented frequency of metastatic tumors to the nervous system varies depending upon the age and sex of the patient population of the institution as well as whether the study is based on autopsy or surgical material. As a rule, the incidence of metastatic tumors increases with age usually beginning in the fifth decade. Metastases from the kidney account for about 10% of all the metastases to the brain. Other primary neoplasms with a clear cell appearance are numerous and include gastrointestinal tract neoplasms which constitute about 5% of the metastases to the CNS. Ultrastructural examination of clear cell carcinoma tumor cells will show cell to cell junctions, a feature that is absent in gliomas and central neurocytoma. The presence of an associated basal lamina in carcinomas would aid in the exclusion of meningioma.

Metastatic renal cell carcinoma with cytoplasmic
accumulations of partially extracted glycogen.
Metastatic malignant melanoma, with heavily pigmented melanosomes, must also be considered in large or clear cell CNS neoplasms.

GROUP 2: SPINDLE CELLS
tanycytic ependymoma, astrocytoma

Tanycytic ependymoma, an ependymoma variant, is described as a low to moderately cellular fibrillar neoplasm with modest nuclear pleomorphism and an absence of mitotic figures. The perivascular rosettes are easily overlooked and true rosettes are conspicuously absent. This spindle-cell variant of ependymoma is believed to arise from tanycytes. These are elongate unipolar and bipolar cells that extend between the ventricular lumen and the surface of the nervous system.

For the purpose of description, tanycytes can be described as consisting of a soma, a neck, and a tail portion. The somatic portion is in the ependymal layer. The neck portion originates from the soma and extends into the subependymal region. It, like the somatic portion, has fine processes radiating from it. The tail portion terminates as small bulbous swellings either on a vessel or at the pial surface. Tanycytes in the spinal cord radiate from the ependyma of the central canal into the gray matter.

Important morphologic differences distinguish simple ependymal cells from tanycytes. These differences take into account shape of end-feet, numbers of apical microvilli and cilia, cytoplasmic processes, and intercellular junctions. Simple ependymal cells have numerous cilia and few microvilli on their ventricular surface compared with the ventricular surface of tanycytes, which posses numerous microvilli and few cilia. Other differences between the two include nuclear shape and cytoplasmic contents. Tanycytes have more irregular shaped nuclei than do ordinary ependymal cells. Further, tanycytes do not possess the perinuclear swirling bundle of filaments but rather contain more intracytoplasmic microtubules; however, both are immunoreactive to antibodies for GFA protein.

Tanycytes also have intricate relationships with blood vessels. The tail process forms multiple contacts along its course as it crosses adjacent blood vessels prior to its termination as a foot-like expansion in apposition to a deeper placed blood vessel. A number of foot processes may converge onto a single capillary or, as in the spinal cord, larger blood vessels may be involved. The peripheral processes of tanycytes that form end-feet at vascular surfaces are separated from the blood vessel by a basal lamina; thus, both tanycyte end-feet and astrocyte end-feet are separated from the endothelial cell in the same manner.

Accurate recognition of this ependymoma variant requires a high index of suspicion and knowledge of the existence of the entity. Furthermore, tanycytic ependymoma may not be as uncommon as is currently believed. Most likely, many tanycytic ependymomas are buried in astrocytoma or schwannoma files.


GROUP 3: UNSUSPECTED SARCOMA METASTASES
alveolar soft part sarcoma, osteosarcoma

Alveolar soft part sarcoma is a clinically and morphologically distinct soft tissue sarcoma. Cerebral metastases are more common with alveolar soft part sarcoma than with other soft tissue sarcomas. Cerebral metastases may be the first manifestation of the disease. With microscopic examination, the neoplasm consists of nests of large tumor cells with loss of central cohesion resulting in a pseudoalveolar pattern. The nests are separated by thin-walled, sinusoidal vascular spaces. Individual cells are large and rounded with distinct cytoplasmic borders. These large cells may resemble gemistocytic astrocytes, atypical neurons, or metastatic renal cell carcinoma. The cells in alveolar soft part sarcoma are immunonegative with antibodies against GFA protein, but may occasional expression S-100 protein. In the electron microscope the large globoid cells characteristically contain rhomboid, rod-shaped, or spicular crystals having a regular lattice pattern and electron dense secretory granules. The crystals and granules are membrane bound and consist of filaments about 6 nm in diameter. Usually the filaments are arranged in a parallel array with a periodicity of 10 nm. The large globoid cells are separated from the adjacent vascular structures by a basal lamina. Cell to cell junctions have been cited; however, they are uncommon.

Metastatic alveolar soft part sarcoma. Inset shows higher magnification of cytoplasmic crystals. Metastatic osteosarcoma with stromal collagen
fibrils as part of osteoid formation.

Osteosarcoma (osteogenic sarcoma), a primary malignant tumor of bone, is composed of a malignant connective tissue stroma that gives rise to tumor osteoid and bone. With microscopic examination, most osteosarcomas exhibit a pleomorphic, highly anaplastic cell population with osteoid, tumor bone and cartilage matrix formation. Brain metastases are considered a rare event, however, with prolonged survival of patients with osteosarcoma, there seems to be an increased incidence of CNS involvement. Microscopically, in some metastases to the brain, the osteosarcoma cells diffusely infiltrate the neural parenchyma in a manner similar to the diffuse fibrillary astrocytomas. Although the osteosarcoma cells are immunohistochemically negative with antibodies to GFA protein, some cells in diffusely infiltrating astrocytomas may also be negative. Ultrastructural examination of osteosarcoma shows cells that vary in size and shape. The nuclei are pleomorphic with clumped chromatin and a large nucleolus. The cytoplasm contains a disorganized and dilated rough endoplasmic reticulum, mitochondria, lipid droplets and few intermediate filaments. The osteoid is composed of collagen fibrils in a fibrillo-granular matrix. Calcified osteoid consists of hydroxyapatite crystals, collagen fibrils.


References

Central Neurocytoma

  1. Barbosa MD, Balsitis M, Jaspan T, Lowe J. Intraventricular neurocytoma. A clinical and pathological study of three cases and review of the literature. Neurosurgery 1990;26:1045-1054.
  2. Figarella-Branger D, Pellissier JF, Daumas-Duport C, Delisle MB, Pasquier B, Parent M, Gambarelli D, Rougon G, Hassoun J. Central Neurocytomas. Critical evaluation of a small-cell neuronal tumor. Am J Surg Pathol. 1992;16:97-109.
  3. Hassoun J, Gambarelli D, Grisoli F, Pellet W, Salamon G, Pellissier JF, Toga M. Central neurocytoma. An electron microscopic study of two cases. Acta Neuropathol. 1982;56:151-156.
  4. Hassoun J, Soylemezoglu F, Gambarelli D, Figarella-Branger D, von Ammon K, Kleihues P. Central neurocytoma. A synopsis of clinical and histological features. Brain Pathol. 1993;3:297-306.
  5. Hessler RB, Lopes MBS, FrankfurterA, Reidy J, VandenBerg SR. Cytoskeletal immunohistochemistry of central neurocytomas. Am J Surg Pathol. 1992.16:1031-1038.
  6. Kubota T, Hayashi M, Kawano H, Kabuto M, et al. Central neurocytoma. Immunohistochemical and ultrastructural study. Acta Neuropathol. 1991 ;81 :418427.
  7. Langford LA. CNS neoplasms: Indications for electron microscopy. Ultrastruct Pathol. 1996;20:35-46.
  8. Langford, LA, Camel MH. Palisading pattern in cerebral neuroblastoma. Mimicking the primitive polar spongioblastoma. An ultrastructural study. Acta Neuropathol 1987;73:153-159.
  9. Raimondi AJ. Ultrastructure and the biology of human brain tumors. In: Krayenbuehl, H, Maspes, PE, Sweet, WH, eds. Progr Neurol Surg. vol. 1. New York, NY: Karger;1966:1-63.
  10. von Deimling A, Janzer R, Kleihues P, Wiestler OD. Patterns of differentiation in central neurocytoma. An immunohistochemical study of eleven biopsies. Acta Neuropathol. 1990;79:473-479
  11. Yuen ST, Fung CF, Ng TH, Leung SY. Central neurocytoma. Its differentiation from intraventricular oligodendroglioma. Childs Nerv Syst. 1992;8:383-8.

Oligodendroglioma

  1. Baloyannis S. The fine structure of the isomorphic oligodendroglioma. Anticancer Res. 1981;1:243-348.
  2. Cervos-Navarro J, Pehlivan N. Ultrastructure of oligodendrogliomas. Acta Neuropathol. 1981;Suppl Vl1:91-93.
  3. Hossman KA, Wechsler W. Ultrastructural cytopathology of human cerebral gliomas. Oncology. 1971;25:455-480.
  4. Kamitani H, Masuzawa H, Sato J, Okada M. Ultrastructure of concentric laminations in primary brain tumors. Acta Neuropathol. 1986;71 :83-87.
  5. Kros JM, de Jong AAW, van der Kwast TH. Ultrastructural characterization of transitional cells in oligodendrogliomas. J Neuropathol Exp Neurol. 1992;51 :186193.
  6. Langford LA. CNS neoplasms: Indications for electron microscopy. Ultrastruct Pathol. 1996;20:35-46.
  7. Min K.-W, Scheithauer BW. Oligodendroglioma. The ultrastructural spectrum. Ultrastruct Pathol. 1994; 18:47-60.
  8. Robertson DM, Vogel FS. Concentric lamination of glial processes in oligodendrogliomas. J Cell Biol. 1962;15:313-334.

Tanycytic ependymoma

  1. Friede RL, Pollak A. The cytogenetic basis for classifying ependymomas. J Neuropathol Exp Neurol. 1978, 37:103-118.
  2. Langford LA, Barre GM. Tanycytic ependymoma. Ultrastruct Pathol. 1997, 21 :127-134.

Meningioma

  1. Erlandson RA. Diagnostic transmission electron microscopy of tumors. New York: Raven Press, 1994.
  2. Kepes JJ. Meningioma. Biology, pathology, and differential diagnosis. New York: Masson, 1982.
  3. Langford LA. Pathology of meningiomas. J Neurooncology. 1996;29:217-221.

Ependymoma

  1. Brightman MW, Palay SL. The fine structure of ependyma in the brain of the rat. J Cell Biol. 1963; 19:415-439.
  2. Fleischhauer K. Ependyma and subependymal layer. In: Bourne GH, ed. The Structure and Function of Nervous Tissue. vol. Vl. Academic Press, New York, NY: 1972; 1 -46.
  3. Guccion JG, Saini N. Ependymoma. Ultrastructural studies of two cases. Ultrastruct Pathol. 1991 ;15:159-166.
  4. Haustein J, Cruz-Sanchez F, Cervos-Navarro J. Ependymoma. Electron microscopic criteria and demonstration of GFA protein. Zentralbl Alig Pathol 1987;133:163-174.
  5. Ho KL. Abnormal cilia in a fourth ventricular ependymoma. Acta Neuropathol. 1986;70:30-37.
  6. Ho KL, Caccamo DV, Garcia JH. Intracytoplasmic lumina in ependymomas. An ultrastructural study. Ultrastruct Pathol. 1994; 18:371 -380.
  7. Kubota T, Ishiise J, Yamashima T, Yamamoto S. Abnormal cilia in a malignant ependymoma. Acta Neuropathol. 1986;71 :100-105.
  8. Langford LA. CNS neoplasms: Indications for electron microscopy. Ultrastruct Pathol. 1996;20:35-46.
  9. Peters A, Palay SL, Webster H deF. The Fine Structure of the Nervous System. Neurons and Their Supporting Cells. New York: Oxford University Press;1991 :273-327.
  10. Sara A, Bruner JM, Mackay B. Ultrastructure of ependymoma. Ultrastruct Pathol. 1994; 18:33-42.
  11. Zuppan CW, Mierau GW, Weeks DA. Ependymoma with signet-ring cells. Ultrastruct Pathol. 1994:18:43-46.

Alveoiar soft part sarcoma

  1. Bindal RK, Sawaya RE, Leavens ME, Taylor SH, Guinee VF. Sarcoma metastatic to the brain: results of surgical treatment. Neurosurgery. 1994;35:185-191.
  2. Lillehei KO, Kleinschmidt-DeMasters B, Mitchell DH, Spector E, Kruse CA. Alveolar soft part sarcoma: an unusually long interval between presentation and brain metastasis. Hum Pathol. 1993;24:1030-1034.
  3. Menesce LP, Eyden BP, Edmondson D, et al. Immunophenotype and ultrastructure of alveolar soft part sarcoma. J Submicrosc Cytol Pathol. 1993;25:377-387.
  4. Mukai M, Torikata C, Iri H. Alveolar soft part sarcoma: an electron microscopic study especially of uncrystalized granules using a tannic acid fixative. Ultrastruct Pathol. 1990; 14:41 -50.
  5. Ordonez NG, Ro JY, Mackay B. Alveolar soft part sarcoma: an ultrastructural and immunocytochemical investigation of its histogenesis. Cancer. 1989;63:1721-173.

Osteosarcoma

  1. Dickersin GR, Rosenberg AE. The ultrastructure of small cell osteosarcoma with a review of the light microscopy and differential diagnosis. Hum Pathol. 1991 ;22:267-275.
  2. Ghadially FN, Mehta PN. Ultrastructure of osteogenic sarcoma. Cancer. 1970;25: 1457-1467.
  3. Marina NM, Pratt CB, Shema SJ, Brooks T, Rao B, Meyer WH. Brain metastases in osteosarcoma. Report of a long-term survivor and review of the St. Jude Children's Research Hospital experience. Cancer. 1993;71 :3656-3660.
  4. Shapiro F. Ultrastructural observations on osteosarcoma tissue: a study of 10 cases. Ultrastruct Pathol. 1983;4: 151 - 161.
  5. Wexler LH, De Laney TF, Saris S, Horowitz ME. Long-term survival after central nervous system relapse in a patient with osteosarcoma. Cancer. 1993;72:12031208.

Top of page

SUP Logo