Heffalumps, jagulars, and cheshire cats. A commentary on cytokeratins and soft tissue sarcomas.

Historically, antibodies to cytokeratin intermediate filaments have been models of target specificity. In most diagnostic settings, the utility of these antibodies was unquestioned; reactivity for cytokeratin was dogmatically equated with epithelial differentiation. Recently, however, the diagnostic importance of these antibodies has been challenged, prompted by the demonstration of cytokeratin reactivity in a variety of "nonepithelial" neoplasms. In this review, the evolving literature on this topic is explored, and the practical implications of these findings are discussed.