Microinvasion— When is it Invasive?

Abstract & Commentary

Synopsis: Ductal carcinoma in situ may contain small areas of invasion termed microinvasion. Unfortunately, varying definitions used in different studies has hindered the long-term prognosis and appropriate management of this entity. This study divided microinvasive lesions into 2 types based on the characteristics of the invading cells. One type has a natural history similar to DCIS. The other type is intermediate to DCIS and invasive breast cancer.

Source: de Mascarel I, et al. Cancer. 2002;94: 2134-2142.

One of the important exercises in evaluating a specimen with ductal carcinoma in situ (DCIS) is the identification of stromal invasion because of the well-known differences between pure DCIS and invasive breast cancer. The identification of small areas of invasion has been termed microinvasion. Unfortunately, the long-term prognosis and appropriate management of this entity have been hindered by varying definitions used in different studies.

In this study from a single institution in France, patients with microinvasive breast cancer (MIBC) were subdivided into 2 groups: those with a few single infiltrating tumor cells (DCIS-MI type 1) and those with a few infiltrating tumor cell clusters (DCIS-MI type 2). For comparison, 2 other groups of patients were studied: 1) patients with DCIS alone; and 2) patients with infiltrating ductal carcinoma but a predominant (at least 80%) intraductal component (IDC-DCIS).

Between 1970 and 1996, 1248 patients were identified with either DCIS, DCIS-MI (types 1 or 2), or IDC-DCIS. The goal was to correlate the histologic appearance with clinical behavior by analyzing the 154 recurrences (12.3%), 84 of which were distant, and the 117 deaths (9.4%). Factors influencing the local recurrence rate were not analyzed.

de Mascarel and associates found that there were no differences between DCIS and DCIS-MI type 1 with respect to metastasis-free or overall survival. On the other hand, patients with DCIS-MI type 2 could be distinguished from DCIS patients by their metastasis-free survival (91% vs 98%, respectively; P < .0001) and their overall survival (88.4% vs 96.5%, respectively; P < .0001). Similarly, the overall survival was worse for IDC-DCIS patients (78.5%) vs. DCIS-MI type 2 patients (88.4%); P = .03. Based on these results, 3 different groups, defined by their likelihood of metastasis-free and overall survival, were identified: 1) those with either DCIS and DCIS-MI type 1; 2) DCIS-MI type 2; and 3) IDC-DCIS.

Comment by Kenneth W. Kotz, MD

Microinvasion is characterized by its association with DCIS; small invasive tumors without an in situ component should not be classified as microinvasive. DCIS is recognized by the proliferation of epithelial cells with malignant cytologic features confined to the glandular component of the ducts. Although exact definitions vary, microinvasion occurs when there is invasion of the stroma surrounding ducts and lobules containing DCIS. The significance of defining MIBC as a separate entity is to determine whether the clinical behavior of this lesion is more closely related to DCIS, invasive breast cancer, or something in-between.

Most pathologists refer to microinvasion when a predominantly in situ lesion contains a focus (or foci) of invasion which is no more than 1 mm in diameter. Using a different scheme, de Mascarel et al divided microinvasive lesions into 2 types based on whether the invasion was by individual cells or tumor cell clusters. In fact, they did not measure the size of each focus of microinvasion. The variety of different diagnostic criteria and definitions of MIBC likely accounts for some of the uncertainty regarding its clinical significance.

Many others report that MIBC behaves similar to, or slightly worse than high-risk DCIS,1,2 with the rate of axillary nodal metastases generally less than 10%,1 but ranging from 0-20%.3 In the study by de Mascarel et al, positive axillary nodes were present in 3 of the 9 patients with DCIS-MI type 2 who ultimately died. In a recent study of DCIS and microinvasion, the sentinel lymph node biopsy technique identified an involved node in 10% of cases.4 On the other hand, a study of 38 patients with DCIS and microinvasion (defined as a single focus < 2 mm or up to 3 foci each < 1 mm) found no histologically positive nodes after an axillary lymph node dissection.5

The likelihood of finding microinvasion in DCIS is greater in lesions of increasing size.2 Depending on the series, the average size of a DCIS lesion, when microinvasion is present, is anywhere from 25-50 mm.6 Furthermore, it is well recognized that microinvasion is most common when the DCIS lesion is the comedo type.2,6 Histologically, MIBC is characterized by malignant cells in the stroma, absence of both basement membrane and myoepithelial cells around the invasive cells, and the frequent presence of a lymphocytic infiltrate.3 Unfortunately, there are well-recognized lesions which can lead to a false-positive or false-negative diagnosis of microinvasion.1,3 However, immunostaining may be useful by 1) differentiating tumor cells from inflammatory cells; and 2) confirming the presence or absence of the smooth muscle, actin-positive layer of myoepithelial cells that should not be present around the microinvasive area.1,7

Defining the clinical significance of MIBC is limited not only by relatively few series with small numbers of patients, but also by varying definitions used. In the large study by de Mascarel et al, the behavior of DCIS-MI type I was similar to DCIS whereas DCIS-MI type 2 was intermediate to DCIS and invasive breast cancer. Whether this approach is superior to definitions based on a 1 mm cut-off is unknown. Although it is difficult to draw definitive conclusion, MIBC lesions do have metastatic potential and decisions on adjuvant therapy should be made based on standard pathologic information.

Dr. Kotz is Clinical Assistant, University of North Carolina; Hanover Medical Specialists, Wilmington, NC.


1. Hoda R, et al. Arch Pathol Lab Med. 2001;125: 1259-1260.

2. Hoda S, et al. Histopath. 1999;35:468-470.

3. Hoda S, et al. Am J Surg. 2000;180:305-308.

4. Klauber-De More N, et al. Ann Surg Oncol. 2000;7: 636-642.

5. Silver A, et al. Cancer. 1998;82:2382-2390.

6. de Mascarel I, et al. Cancer. 2002;94:2134-2142.

7. Prasad M, et al. Anat Pathol. 1998;3:209-232.