mainly located within the cell processes but not within perikarya. The findings thus mimicked those of classic, pediatric neuroblastoma~.~ In contrast, dense core or neurosecretory-type granules in SITA resemble those of enterochromaffin (endocrine) cells of the digestive tract. They are aggregated within the perikarya, are often pleomorphic in shape, and are larger in size.'-' Considering its immunophenotype and the ultrastructural findings, we consider Case 26 to have a biphasic tumor composed of epithelial and genuine neuroblastic cells rather than a purely epithelial nature tumor such as a SITA.
In addition to this tumor, we had another biphasic tumor in the study (Case 251, one exhibiting ganglioneuroblastoma-like and epithelial (tubular, papillary, and rosette structures) features.' The epithelial differentiation exhibited by these two ONB could be regarded as differentiation toward the supporting cells of the olfactory epithelium, but not toward the mucosal epithelium of the nasal and paranasal sinus (Schneiderian membrane). Taking into account the wide morphologic variety of ONB, which ranges from neuronal to epithelial features, the finding of epithelial differentiation does not exclude the diagnosis of ONB. We do, however, agree with Dr. Vartanian that SITA should be distinguished from ONB in that the two lesions may have different etiologies and biologic behavior.