Spindle cell ductal carcinoma in situ. An unusual variant of ductal intra-epithelial neoplasia that simulates ductal hyperplasia or a myoepithelial proliferation.

Abstract

Seventeen examples of a variant of ductal carcinoma in situ (DCIS) composed exclusively or predominantly of spindle cells arranged in fascicles, whorls, and solid sheets are described. The fascicular arrangement of the spindle cells simulates the "streaming" phenomenon associated with ordinary intraductal epithelial hyperplasia (IDH). This process also resembles the myoid, solid form of intraductal myoepithelial proliferation. The women ranged in age from 38 years to 79 years with a mean age of 59.3 years. Five patients presented with a palpable mass. The remaining tumors were discovered using mammography. The radiological appearances of the lesions raised concern for carcinoma, but there were no distinctive mammographic findings to suggest an unusual variant of DCIS. Cytological preparations were suspicious for malignancy in two patients and were reported as malignant in another case. Sixteen patients were treated with wide local excision, and one woman had a partial mastectomy. The tumors measured from 3 mm to 15 mm (mean 8.65 mm). In three cases, minute foci of stromal invasion were associated with the spindle cell DCIS. In another specimen, a 2.7-cm invasive ductal carcinoma of no special type was identified in an area away from the foci of the spindle cell DCIS. None of the patients has experienced recurrence or metastasis during the relatively short mean follow-up period of 16.2 months (range 4-77 months). Spindle cell DCIS is distinguished from the streaming pattern of ordinary IDH by its solid growth pattern, lack of secondary spaces or peripheral fenestrations, uniformity of appearance and distribution of nuclei, cytological atypia in the range of low to intermediate-grade DCIS, and negative immunoreaction with CK-34betaE12 (HMW-CK903). When fenestrations are present, they are evident in areas of cribriform DCIS that merge with the solid, spindle cell areas in hybrid ducts harboring both patterns. This admixture, with conventional cribriform DCIS, and the association with foci of invasive ductal carcinoma in some cases further help recognition and confirmation of this lesion as in situ carcinoma. When there is no transition from the spindle cells to recognizable cribriform DCIS, distinction from intraductal myoepithelial hyperplasia (myoepitheliosis) requires immunostains for actin and S-100 protein. Recognition of this pattern of DCIS is important in order to avoid its frequent misclassification as a benign lesion.