Special techniques such as serial macroscopic sectioning (SMS) or immunohistochemical staining (IH) improve the detection rate of micrometastases but this detection is of value only if it improves the prediction of recurrence and survival. We first studied the prognosis of 120 patients with a single micrometastasis detected by SMS in a series of 1,680 primary operable breast carcinoma with a median follow-up of 7 years. A significant difference in recurrence (P = 0.005) and in survival (P = 0.0369) was found between node-negative patients and those with one single SMS micrometastasis, but SMS micrometastases were not a predicting factor by multivariate analyses according to the Cox model. We then studied the prognostic significance of patients with a micrometastasis detected by IH in node-negative carcinoma: 37 micrometastases from a series of 89 invasive lobular carcinoma (ILC) and 13 single micrometastases from a series of 129 invasive ductal carcinoma (IDC). In the ILC group, IH micrometastases had no prognostic value (median follow-up: 9.3 years). In the IDC group, IH micrometastases were correlated with recurrences (P = 0.01) and were the most significant predicting factor, but were less correlated with survival (median follow-up: 15.6 years). Three main points emerge from this study: (1) SMS micrometastases have a prognostic significance and macroscopic sectioning is recommended as a routine technique not requiring excessive work. (2) IH micrometastases in infiltrating lobular carcinoma have no prognostic significance. (3) The value of IH is debatable in infiltrating ductal carcinoma, since the technique is of principal use in predicting recurrences. It should therefore be carefully assessed vs other prognostic factors currently under study.