Breast biopsy or mastectomy cases having diagnoses of carcinoma in situ with "microinvasion," "minimal invasion," "focal invasion," or "suggestive of invasion" were reviewed and all histologically identified foci of invasive disease from each case were measured using an ocular micrometer. Cases in which any single focus of invasion was greater than 5 mm or the added size of separate invasive foci exceeded 10 mm were excluded, resulting in a study group of 75 patients. Invasive neoplasm was present in the initial biopsy in 69 of 75 cases (92%); however, residual invasive neoplasm was found in the subsequent lumpectomy/mastectomy from 14 of these (20%). In 59% of cases, two or more histologically separate foci of invasion were identified. Invasive foci consisted of isolated cells or cell clusters, each less than 1 mm (microfocal invasion), in 33% of cases. In 12 cases, the sum of individual invasive foci was 5 to 10 mm. Axillary lymph nodes (LN) from 5 of 69 patients (7%) contained metastatic carcinoma (four cases, one LN positive; one case, two LN positive). The cumulative sizes of all invasive foci in the LN-positive group were microfocal invasion (one case), 0.6 mm (one case), 1.1 mm, 2.5 mm, and 5.8 mm. The difference in frequency of axillary node metastasis between tumors with microfocal and measurable invasion (4.3% v 8.6%) was not statistically significant. Follow-up data were available on 55 cases (mean interval, 66.1 months). One (node-negative) patient had duct carcinoma in situ recurrence in the same breast 4 years after initial treatment. Another (with unknown node status) developed an axillary lymph node metastasis 13 months after initial treatment (96% disease-free survival). We conclude that microscopic stromal invasion in breast carcinoma, at least in the setting of significant in situ component, is often initiated from multiple foci. Patients with microscopically invasive breast carcinoma have a small but significant risk of axillary metastases, although a highly favorable survival.