The risk factors for the development of invasive breast cancer and DCIS are similar. A further dilemma in the classification and histological analysis of DCIS is micro invasion. DCIS with micro invasion (DCISM) may also result in axillary lymph node metastases, whereas patients with DCIS should not, by definition, have axillary metastases. A higher suspicion for axillary metastases with DCISM can be obtained from the primary tumor characteristics. Statistically significant independent predictors of lymph node metastases in DCISM are comedo DCIS (P
< 0.03) and the number of DCIS-involved ducts (P
< 0.002) [1
On pure morphological assessment a potential diagnostic trap is invasive ductal carcinoma with central necrosis. As the name indicates the tumor has a comedo DCIS like appearance and is likely to be diagnosed as DCIS comedo while in reality it is infiltrative breast carcinoma with central necrosis. This situation is identical to invasive cribiform carcinoma, a rare form of breast malignancy which very closely mimics cribriform DCIS [2
]. The most important aspect of this concept is the realization that a breast carcinoma may be partly or entirely DCIS like, yet invasive. Recently a solid variant of invasive cribriform carcinoma is also described [3
]. Similar morphologic patterns are also seen in salivary duct tumors, sweat gland carcinomas [4
] and high grade prostatic adenocarcinomas. In case of the entire morphology having this feature, it is possible to report primary tumor as DCIS, following a conservative approach without further work up or axillary lymph node sampling. The other more common scenario is to incorrectly asses the size of the invasive component resulting in incorrect pTNM staging and management as pathological tumor size for classification (pT) is a measurement of only the invasive component [5
In such a dilemma IHC is very useful in assessment of invasion. In the ideal world invasive cancers are characterized by lack of both basement membrane and myoepithelial cells. However in the real world while invasive cancer lacks myoepithelial cells, some produce basement membrane components adding further to the confusion. Therefore for the assessment of DCIS and invasive comedo DCIS assessment of myoepithelial lining is most reliable. A number of myoepithelial markers including S-100, Alpha smooth muscle Actin, SMM – HC, Calponin and HMW-CK are available with different sensitivities and specificities. SMM-HC is thought to be the most specific while other though quite sensitive but are less specific. Some other myoepithelial markers include Maspin, CD10 and P63. Amongst these markers P63 is particularly useful as it stains the myoepithelial nuclei only with high sensitivity and specificity [6
]. Myoepithelial antibody cocktail is another good choice [7
]. With Actin one should be particularly careful not to confuse periductal myofibroblast staining as myoepithelial staining. In routine surgical pathology practice however it is not practical to do IHC for myoepithelium routinely on all such cases. One morphologic feature which we found useful on H & E was the concentric stromal reaction around these invasive foci (Fig ). In addition irregular circumference of these invasive foci compared to true DCIS comedo was also helpful.
Foci of invasive ductal carcinoma breast with central necrosis. Note concentric stromal reaction around these foci (arrow), a helpful morphologic feature. H & E, Mag: 4×.