V. Jay et al., POLYPHENOTYPIC TUMORS IN THE CENTRAL-NERVOUS-SYSTEM - PROBLEMS IN NOSOLOGY AND CLASSIFICATION, PEDIATRIC PATHOLOGY & LABORATORY MEDICINE, 17(3), 1997, pp. 369-389
In recent years, there is increasing recognition of polyphenotypic hig
h-grade malignancies in the non-central nervous system (CNS) tumor lit
erature. Some of these tumors have been regarded as variants of primit
ive neuroectodermal tumor (PNET) or as extrarenal malignant rhabdoid t
umors (MRTs). This report concerns two posterior fossa neoplasms, both
of which displayed a ''polyphenotypic'' expression of neural, epithel
ial, myogenic, and glial markers, including synaptophysin, neurofilame
nt, vimentin, glial fibrillary acidic protein, S-100, neuron-specific
enolase, desmin, S antigen, MIC2, cytokeratin, epithelial membrane ant
igen, and carcinoembryonic antigen. One tumor showed complex intercell
ular junctions, cytoplasmic intermediate filaments, well-developed rou
gh and smooth endoplasmic reticulum and Golgi apparatus, cilia, and ne
urosecretory granules. The other neoplasm showed pools of glycogen, de
smosomes, and tonofilaments. The histological and ultrastructural appe
arances were inconsistent with glioma, PNET, meningioma, ependymoma, c
horoid plexus carcinoma, sarcoma, germ cell tumor, and other tumors in
the World Health Organization classification. Although the polyphenot
ype raises the issue that these may represent variants of MRT or the a
typical teratoid-rhaboid tumor, the morphologic findings in the two ca
ses were very dissimilar. Our two cases underscore the problems in nos
ology and classification of polyphenotypic tumors of the CNS. This is
particularly significant, as therapeutic protocols for PNET, MRT, and
non-CNS polyphenotypic tumors are different. We review the literature
on polyphenotypic tumors and reiterate the difficulties in precise cla
ssification of these complex tumors.