Answer: Sarcomatoid carcinoma (squamous cell carcinoma, spindle cell type)

Histology: The tumor is polypoid and its surface is extensively ulcerated. The subepithelium is expanded by a storiform proliferation of spindled cells. The spindled cell exhibit moderate atypia, focal cytoplasmic clearing, and a relatively high mitotic rate. There is no evidence of squamous differentiation, and the non-ulcerated surface does not demonstrate dysplastic changes. Immunohistochemically, the tumor cells are positive for AE1/AE3 and p63.

Discussion: Sarcomatoid carcinoma of the upper respiratory tract is a distinct variant of squamous cell carcinoma trademarked by a prominent or even exclusive spindle cell component. It is commonly encountered in patients who have undergone radiation therapy of a more conventional squamous cell carcinoma, and radiation exposure has been implicated as a predisposing factor. At the light microscopic level, this variant of squamous cell carcinoma may cause confusion with other processes, particularly when a patient has undergone radiation for a prior squamous cell carcinoma. At one extreme, the spindle cell proliferation may closely resemble a reactive fibroblastic proliferation and be mistaken for a pseudosarcomatous reaction to an associated carcinoma or to radiation. At the other extreme, the spindle cell proliferation may be mistaken for a true sarcoma.

The distinction of sarcomatoid carcinoma from certain non-neoplastic spindle cell proliferations of the upper respiratory tract such as radiation induced stromal atypia can be problematic. Certain phenotypic features of sarcomatoid carcinomas may contribute to this diagnostic confusion. First, the degree of cellular atypia and cellular density in sarcomatoid carcinomas is variable. Sparsely cellular tumors that lack significant pleomorphism may not be easily recognized as a frankly malignant process. Second, the spindle cell proliferation often overshadows the epithelial component. In those instances where the presence of surface dysplasia and/or conventional squamous cell carcinoma is not easily appreciated, confirmation of a carcinoma becomes highly dependent on immunohistochemical documentation of cytokeratin expression in the spindled cells. Third, immunohistochemistry is not always reliable in its ability to confirm epithelial origin. The frequency of sarcomatoid carcinomas demonstrating cytokeratin positivity by immunohistochemistry ranged from 50 to 75%. In effect, absence of staining does not exclude the diagnosis of sarcomatoid carcinoma. Conversely, the presence of positive cytokeratin staining in an atypical spindle cell proliferation does not necessarily confirm a diagnosis of sarcomatoid carcinoma. Rather, focal positive staining should be interpreted with extreme caution given that many sarcomas and pseudosarcomatous stromal proliferations aberrantly express cytokeratin. In addition, some investigators have reported that the majority of sarcomatoid carcinomas retain a strong immunoreactive for p63 – a marker that is expressed in non-neoplastic squamous epithelium of the upper respiratory tract and in most conventional head and neck squamous cell carcinomas. Importantly, p63 expression is not encountered in most sarcomas or benign spindle cell lesions. In the present case, the diffuse staining for cytokeratin and p63 confirmed the diagnosis of sarcomatoid carcinoma.