An immunohistochemical analysis of endometrial stromal and smooth muscle tumors of the uterus: a study of 54 cases emphasizing the importance of using a panel because of overlap in immunoreactivity for individual antibodies.

The expression of desmin, h-caldesmon, calponin, CD10, CD34, CD99, inhibin, and keratin (AE1/3-Cam 5.2) was studied in 10 conventional leiomyomas, 9 highly cellular leiomyomas, 9 epithelioid smooth muscle tumors, 9 leiomyosarcomas, 10 endometrial stromal tumors (4 with smooth muscle metaplasia), and 7 uterine tumors resembling ovarian sex cord tumors (UTROSCTs). c-kit expression was tested in 10 endometrial stromal tumors, 7 UTROSCTs, and 9 leiomyosarcomas. Desmin was positive in almost all smooth muscle tumors except those of epithelioid type, which were positive in only about half of the cases. It also stained areas of smooth muscle differentiation in endometrial stromal tumors and five of seven UTROSCTs. h-caldesmon was positive in almost all nonepithelioid smooth muscle tumors and in areas of smooth muscle differentiation in endometrial stromal tumors; it was positive in only about half of the epithelioid smooth muscle tumors and negative in all UTROSCTs. Calponin was positive in most tumor types. CD10 was positive in nine of 10 endometrial stromal tumors and five of seven UTROSCTs, although very focally in the latter group. It was also expressed, however, in almost all leiomyosarcomas, almost 50% of highly cellular leiomyomas, and rarely in the other smooth muscle tumors. CD34 was negative in the tested tumors with rare exceptions. CD99 and inhibin were positive in four of seven and one of seven UTROSCTs. Keratin positivity was found in most (five of seven) UTROSCTs and occasionally in smooth muscle tumors (seven of 37). c-kit was negative in all endometrial stromal tumors, UTROSCTs, and leiomyosarcomas. The major conclusions of this study are as follows: 1) Pure endometrial stromal tumors are usually desmin negative. 2) In contrast to some previous studies, CD10 expression was often seen in smooth muscle tumors, including most leiomyosarcomas and almost half of highly cellular leiomyomas. As a result, a panel of CD10, h-caldesmon, and desmin should be used and will distinguish endometrial stromal tumors from highly cellular leiomyomas in most cases. 3) In contrast to a previous study, no significant differences in immunoreactivity were seen between h-caldesmon and desmin in tumors with smooth muscle differentiation. 4) The absence of h-caldesmon in UTROSCTs helps separate them from epithelioid smooth muscle tumors. 5) UTROSCTs may express epithelial, stromal, and smooth muscle markers, suggesting divergent differentiation. 6) Our study shows less frequent inhibin expression in the sex cord-like elements of the UTROSCTs than in other studies. 7) c-kit may help distinguish metastatic endometrial stromal tumors of the uterus (c-kit negative) from gastrointestinal stromal tumors (c-kit positive). 8) CD34, CD99, and keratin have no or minimal role in this area, but keratin positivity in smooth muscle tumors should not lead to their confusion with epithelial tumors.