Thymic carcinomas, but not thymomas and carcinomas of other sites, show CD5 immunoreactivity.

Thus far, there are no immunohistochemical markers that are specific for thymic epithelial neoplasms, although demonstration of immature T cells in an epithelial tumor can indirectly support a diagnosis of thymoma. In this study, the usefulness of a paraffin section-reactive CD5 antibody (clone CD5/54/B4) for supporting the thymic origin of an epithelial neoplasm was evaluated. Antigen retrieval was effected by microwaving in citrate buffer. Sixteen of 24 thymic carcinomas (67%) were immunoreactive for CD5, including nine of nine squamous cell, two of two undifferentiated, two of four lymphoepithelioma-like, and one case each of basolid carcinoma, clear cell carcinoma, and unclassified thymic carcinoma, but none of four thymic small cell carcinomas. None of 17 cases of benign thymoma and 21 cases of invasive thymoma (including six cases classifiable as well-differentiated thymic carcinoma using the Muller-Hermelink criteria) was immunoreactive for CD5, in the presence of CD5-positive lymphocytes as an internal positive control. Two of three thymic neoplasms with features borderline between thymic carcinoma and invasive thymoma were immunoreactive for CD5. In contrast, none of 61 cases of other malignant neoplasms with a tendency to involve the mediastinum was immunoreactive for CD5, including 40 nonthymic carcinomas and 13 malignant germ cell neoplasma. CD5 staining of thymic epithelial tumors correlated with the absence of tumor-associated CD99-positive thymocytes, as demonstrated in our previous studies. We conclude that CD5 is a useful marker of primary thymic carcinomas. Taken together, CD5 and CD99 (or other immature T-cell markers such as TdT and Cd1a) should be particularly useful in evaluating mediastinal and other biopsy samples of possible thymic epithelial neoplasms and in the subtyping of these tumors.