Breast cancer can only be life threatening when it becomes invasive, at which point it carries potential for spreading and metastasis. Therefore, it is critical to distinguish invasive carcinomas (IC) from noninvasive lesions, the latter including ductal carcinoma in situ (DCIS) and benign breast lesions. While this distinction is usually made based on histologic evaluation alone, in a small but significant number of cases, accurate diagnosis may be impossible, particularly in the context of core needle biopsies. To this end, a number of immunohistochemical markers have been utilized to help establish the presence (or absence) of stromal invasion. The fact that the loss of the myoepithelial cell (MEC) layer is a hallmark of IC (but not DCIS) suggests the use of antibodies to MEC to distinguish IC from DCIS. However, these markers have a wide range of specificity and sensitivity, with the potential for problems in interpretation. The use of many of these markers is limited by high rates of 'false positive' or 'false negative' immunostaining. In this review, the biology of stromal invasion in breast carcinomas will be discussed, as well as the various myoepithelial markers, with emphasis on pitfalls related to their sensitivity and specificity in the detection of MECs in the breast. Finally, the authors will discuss diagnostically challenging breast lesions, which may require the use of MEC marker studies to reach a definitive diagnosis.