Double immunolabeling with cytokeratin and smooth-muscle actin in confirming early invasive carcinoma of breast.

Histopathological identification of invasive breast carcinoma in its earliest phases is fraught with pitfalls. Preinvasive malignant lesions complicated by radial scar, sclerosing adenosis, and lobular cancerization, among other lesions, may simulate invasive carcinoma. Fibrosis, inflammatory reaction, and other stromal changes around in situ carcinoma may mask microinvasive foci on routine stains. Conventional immunohistochemistry to demonstrate basement membrane or myoepithelial cell layer may not, by itself, be unequivocally diagnostic of invasion. We performed a novel double immunoenzyme labeling technique using an avidin-biotin complex peroxidase-diaminobenzidine system for smooth-muscle actin followed by an alkaline phosphatase anti-alkaline phosphatase-new fuchsin system for cytokeratin antigen on formalin-fixed, paraffin-embedded histology sections to evaluate 32 such problematic cases. The initial histologic impression with hematoxylin and eosin staining alone was as follows-first group: microinvasive carcinoma-10; second group: carcinoma in situ--"stromal invasion cannot be ruled out"--15; third group: frankly infiltrating carcinoma of various grades and morphologic types-6. The last group served as positive control for invasion. One fibroadenoma with fine-needle-aspiration-induced artifact simulating stromal invasion was also included. The double immunoenzyme labeling technique imparted a dark brown color to the myoepithelial cells and a vivid red color to the epithelial cells, making individual or loosely cohesive groups of malignant epithelial cells infiltrating the stroma easily detectable, whereas their in situ counterparts were contained within dark brown myoepithelial boundaries. The TNM 1997 definition of pT1mic, i.e., extension of malignant cells in the stroma with no focus measuring >0.1 cm, was followed to classify microinvasion. In the first group, microinvasion was confirmed in six cases but was not demonstrable in four. In the second group, definite invasion was identified in five cases, ruled out in nine, and in one case the suspicion of early invasion could not be entirely ruled out even after double immunoenzyme labeling. Thus, it was possible to render a definite opinion regarding presence or absence of invasion in 24 of 25 (96%) cases diagnosed as or suspected to be microinvasive. The precise and simultaneous elucidation of topography between malignant cells and myoepithelial cells on a single permanent section makes this technique a useful diagnostic tool in the evaluation of those cases of breast carcinoma that exhibit equivocal invasion.