Microinvasive Breast Cancer and the Role of Sentinel Node Biopsy: An Institutional Experience and Review of the Literature


Address correspondence and reprint requests to: Amber A. Guth, MD, NYU Clinical Cancer Center, 160 East 34th Street, 4th floor, New York, NY 10016, USA, or e-mail: amber.guth@med.nyu.edu.


Abstract:  Ductal carcinoma in situ with microinvasion (DCISM) is a distinct clinicopathologic entity. Its true metastatic potential has been unclear, due in part to historical differences in the definition of microinvasion. The role of routine axillary staging for DCISM is controversial, given the reportedly low incidence of axillary metastases. We describe our institutional experience with DCISM, and define the role of axillary staging. A retrospective analysis was made of patients with DCISM. Forty-four patients underwent axillary staging (24 axillary lymph node dissection [ALND], 22 sentinel node biopsy [SNB]). Macrometastatic disease was present in three patients (7%), and two patients had isolated tumor cells (itc) in the sentinel node. Patients with axillary metastases tended to be younger. Comedonecrosis, nuclear grade, multifocal microinvasion or presentation as a clinical mass was not associated with a higher rate of axillary metastases. In this series, 7% of patients had macrometastatic disease, and two patients (5%) had itc only. Axillary staging is indicated, and SNB is appropriate for the identification of axillary metastatic disease.

Ductal carcinoma in situ with microinvasion (DCISM) is a distinct clinicopathologic entity, different from ductal carcinoma in situ (DCIS). While it is an invasive breast cancer, with the ability to metastasize, its true metastatic potential has been unclear, due in part to differences in the definition of microinvasion that have been used by different investigators. Since 1997, however, the criteria as defined by the American Joint Committee on Cancer (AJCC) has been widely accepted (Fig. 1) (1), allowing comparison of different studies. This definition has also been accepted by the World Health Organization.

Figure 1.

 1997 American Joint Commission on Cancer Definition of DCISM “Microinvasive breast carcinoma (T1mic) is the extension of cancer cells beyond the basement membrane into the adjacent tissues, with no single focus larger than 1 mm in greatest dimension.”

The role of routine axillary staging for DCISM is controversial, given the reportedly low incidence of axillary metastases. Sentinel node biopsy (SNB) has virtually replaced axillary lymph node dissection (ALND) in the management of early-stage invasive breast cancer, but its role for patients with DCISM is unclear.

In this report, we present a single institutional experience with DCISM in 59 patients, discuss the available literature, and define the role of SNB in the management of DCISM.

Materials and methods

The Department of Pathology data base was queried for all patients with DCISM, identified from among the 3601 surgical breast cases treated from 1991 to 2006, utilizing the AJCC criteria (1). Patients whose histology was reported as “suspicious for” or “suggestive of” microinvasion were reviewed, and included only if they met the AJCC definition of microinvasion.

Clinical and pathologic variables including diagnosis, patient demographics, size of DCIS, grade, multifocality, estrogen and progesterone receptor status, the performance of axillary staging utilizing either ALND or SNB and its results.

Local treatment consisted of either total mastectomy, or segmental excision with subsequent radiation therapy, as decided by the treating physicians and the patient.

From 1991 to 1997, ALND was performed to surgically stage the axilla. Since 1998, SNB has been routinely performed for all cases. SNB was uniformly performed using both filtered 99TM sulfur colloid and isosulfan blue dye, utilizing intraoperative frozen section analysis with subsequent serial sectioning and immunohistochemical staining.

For breast conservation cases, the entire segmental excision specimen was submitted and examined for histologic evaluation. For mastectomy cases, sections of all grossly identified lesions, biopsy cavity and additional random sections were examined. The size of DCIS was determined by taking the total number of slides that contain DCIS divided by the total number of slides in the specimen multiplied by the largest size focus of DCIS on a single slide to give an estimate of the volume of DCIS within the surgical specimen. The largest size focus is determined by the largest extent of DCIS on a single slide measured microscopically.

For axillary dissection all lymph nodes were submitted and one level each was examined. Sentinel lymph nodes were entirely submitted, three times leveled and submitted for immunohistochemical analysis with cytokeratin. Size of metastatic foci was characterized as macrometastasis, micrometastasis (0.2–2.0 mm) or isolated tumor cells (less than 0.2 mm). For isolated tumor cells, indication as to whether these cells were identified by standard histology or by immunohistochemistry only was noted. Characterization of DCIS included histologic subtyping, nuclear grading, and presence or absence of necrosis. Measurement of DCIS included largest microscopic area and estimation of volume. The presence of multicentricity was also noted. Microinvasion was defined as extension of cancer cells beyond the basement membrane into adjacent tissues with no focus measuring greater than 0.1 mm in greatest dimension. Cases with multiple foci of microinvasion were also included given that the size of the largest focus not greater than 1 mm in greatest dimension.

Long-term follow-up was not available at the time of initial data collection. The data were evaluated for statistical significance utilizing chi-squared analysis. Approval for this study was obtained from the Institutional Review Board.


A total of 59 consecutive patients with DCISM were identified (Table 1). Nineteen patients had a preoperative procedure that diagnosed malignancy: nine had the diagnosis of adenocarcinoma made by fine-needle aspiration biopsy, and 10 patients had DCIS diagnosed on core biopsy. Four additional patients were noted to have microinvasion noted on core biopsy; the remaining 55 patients had the diagnosis of DCIS with microinvasion made from the surgical specimen.

Table 1.   Patient Characteristics
CharacteristicStudy patients (= 59)Patients with axillary metastases (= 5)
  1. p-values not significant for all categories.

Mean age(years)57.651.6
Mean size DCIS (mm)34.8 (range 4–120)59.8 (range 19–120)
Mean nuclear grade2.52.8
Comedonecrosis42 (71%)4 (80%)
Microcalcifications43 (73%)5 (100%)
Palpable mass16 (27%)1 (20%)
Multifocal DCISM25 (42%)2 (40%)

Forty-four patients underwent axillary staging: 24 utilizing ALND and 20 by SNB. Nodal metastases were present in five of the 44 patients (11.4%), two following ALND and three following SNB. Three of the five patients had macrometastatic disease, and in the other two patients immunohistochemistry (IHC)-only isolated tumor cells were detected [now staged by the 6th edition of the AJCC staging manual as N0(i+)] (2).

Of the two patients who had positive lymph nodes on axillary dissection, both had one positive lymph node with macrometastases. Among the patients who underwent SNB, one patient had macrometastatic disease in two sentinel nodes and micrometastatic disease in the third, with two additional positive lymph nodes on ALND; the second had isolated tumor cells (itc) in one of three sentinel nodes and did not undergo complete axillary dissection, and the third patient had itc in two sentinel nodes and underwent a negative axillary dissection. In both cases of itc, the tumor cells were visualized only on IHC staining, and not on H&E stains.

The macrometastatic nodal deposits measured 8, 9, and 15 mm. Hormone receptor status was available only for 18 patients, as many of the earlier patients did not undergo bioreceptor evaluation.

In the patients treated prior to 1998, 13 of 17 (77%) underwent axillary staging, while 37 of 48 (77%) underwent axillary staging after 1997.

Patients with axillary metastases tended to be younger. While there was a trend towards association with larger volumes of DCIS, this was not an absolute association; four of the six patients with DCIS measuring >100 mm in diameter did not manifest axillary nodal disease; of the two with lymph node metastasis both were macrometastatic disease The remaining 53 patients had DCIS measuring less than 100 mm; three had lymph node metastasis, one with macrometastatic disease, and two with isolated tumor cells. DCIS presenting as a clinical mass was not associated with a higher risk of metastatic disease (Table 1).

On chi-squared analysis, no significant differences were present for any of the variables in Table 1.


With the dissemination of widespread breast cancer screening in the Western World, breast cancers are being diagnosed at increasingly earlier stages. Currently ductal carcinoma in situ accounts for 23% of newly diagnosed breast cancers (3), and ductal carcinoma in situ with microinvasion is found in 5–10% of cases of DCIS (4). As the detection of early stage breast cancer continues to increase, it can be expected that DCISM will also be diagnosed more frequently. This raises significant questions regarding its clinical behavior and appropriate management. Is DCISM simply a variant of DCIS, as has been suggested by some authors (5,6), or does the presence of penetration through the basement membrane reflect its metastatic potential? On routine axillary dissection for DCIS, the rate of lymph node metastasis is less than 1% (6). Meanwhile, the reported incidence of axillary metastases in DCISM has ranged from 0 to 28% (1,3–24) (Table 2). What is the real incidence of axillary metastases? Should the axilla be routinely staged, and if so, how?

Table 2.   Literature Review: Axillary Lymph Node Involvement in DCISM
AuthorYear publishedNo. patients/ No. with axillary stagingNo. node (+) patients (%)No. micrometastatic nodes
Schuh (17)198630/306 (20%)
Patchefsky (15)198916/162 (12%)
Wong (21)199041/330
Silverstein (19)199717/170
Silver (18)199838/380
Zavotsky (24)199914/142 (14.3%)1 (7%)
Prasad (16)200021/152 (13%)
Padmore (14)200011/110
Klauber-DeMore (13)200031/313 (10%)2 (7%)
Cox (9)200115/153 (20%) 
Wong (22)200224/242 (8%)
Wasserberg (20)200237/283 (11%) 
de Mascarel (10)2003243/19914 (7%) 
Yang (23)200328/260 
Intra (11)200341/414 (9.7%)2 (5%)
Camp (8)200518/185 (28%)1 (6%)
Katz (12)200621/212 (9.5%)1 (5%)
Broekhuizen (7)200612/123 (25%)2 (16%)
Cavaliere (5)200631/210 
Sakr (27)200614/141 (7%)1 (7%)
Current study200759/445 (11.4%)2 (5%)
Total 762/70943 (6%) 

In our series, 44 patients underwent axillary staging, and 11.4% had axillary metastases identified. This incidence of axillary involvement is in line with the majority of reported series (Table 2). The patients with metastases were younger (mean age of 51.6 years versus 57.6 years) than those without metastases. While the presence of comedonecrosis (20), presentation as a palpable mass (25), and multifocal DCIS (10) have all been linked to a higher risk of axillary nodal disease, there were no statistical differences in our patients (Table 1). Likewise, estrogen and progesterone receptor status results, available for the minority of our patients, also did not predict axillary nodal involvement.

Interestingly, the availability of SNB after 1997 did not appear to influence the decision to perform axillary staging in this series. Seventy-seven percent of patients treated before 1998 underwent ALND, and 77% of patients treated after 1997, when SNB was routinely employed, underwent axillary staging.

In an attempt to identify those patients at risk for axillary metastases, De Mascarel et al. (10) divided DCISM into two prognostic groups: DCIS with microinvasion consisting of infiltration of the periductal stroma by single cells (type 1), or clusters (type 2). In their study, there were no axillary metastases in the DCISM type 1, but a 10.1% incidence in DCISM type 2 (Table 2). Disease-free survival, and the overall survival probability was related to the diagnosis of either pure DCIS, DCISM type 1, and DCISM type 2. They concluded that DCISM type 1 should be considered a variant of DCIS and treated as such, while type 2 is closer in clinical behavior to invasive ductal carcinoma. We do not use this classification in our institution; however, review of all five specimens with nodal disease showed them to be De Mascarel type 2, as were the majority our cases with DCISM.

  • image(2)

[  Photomicrograph (20×) of ductal carcinoma in situ with microinvasion (arrow). ]

Presentation of DCIS as a palpable mass has been linked to microinvasion (4,25). Maffuz et al. reported their experience from Mexico, where DCIS most commonly presents as a palpable mass. The incidence of microinvasive disease and axillary metastases were directly related to tumor size. All positive sentinel nodes contained microinvasive disease only, and no additional metastases were found in nonsentinel nodes.

While Wasserberg et al. (20) reported that comedo-type necrosis, the number of ductals involved by DCIS, and nuclear grade three were all independent predictors of axillary metastases, Intra and associates (11) were unable to link risk of axillary metastases to grade, histologic pattern, or receptor status.

The influence of the presence of microinvasive disease on outcome has been hotly debated. Cavaliere et al. (5) suggest that the natural history of DCISM resembles DCIS and is different from T1a invasive ductal cancers. There were no axillary relapses or distant metastases noted in the 31 patients on long-term follow-up (mean 75 months). Of 21 patients described by Prasad et al. (16) one patient (treated with breast-conserving surgery, and radiation therapy) had recurrent DCIS at 18 months. A second patient had a chest wall recurrence 19 months after mastectomy. All patients were alive without evidence of disease at the time of publication. Silver and Tavassoli (18), detailing the Armed Forces Institute of Pathology experience, reported 38 patients. All had negative ALND, and after a mean follow-up of 7.5 years, there were no cases of local or distant recurrence. Very different outcomes were reported by Schuh (17), who reported 30 patients with DCISM in 1986. Twenty percent had axillary nodal involvement, seven developed metastases, and four died of the disease. Comparison of these results is hampered by historical variations in the definition of DCISM, as well as improvements in early breast cancer diagnosis in the intervening 25 years.

There has also been disagreement in the literature on the impact the presence of micrometastatic disease has on survival. Broekhurzen et al. (7) studied patients with DCIS, DCISM, and T1a invasive ductal carcinomas. They concluded that the presence of micrometastases detected only by immunohistochemical staining did not influence survival. The study by de Mascarel et al. (10) demonstrated a significant difference in overall survival probability among DCIS, DCIS-MI type 1, and DCIS-MI type 2. The 10-year overall survival probability was 96.5% for DCIS, 96.3% for DCIS-MI type 1, and 88.4% for DCIS-MI type 2. This led the authors to conclude that DCIS-MI type 1 behaves like DCIS and should be treated the same way, while DCIS-MI type 2 approaches the clinical behavior of invasive ductal carcinoma with extensive in situ component.

The reliability of these conclusions, including those from this study, is limited by small sample size. This has been addressed by a study of Surveillance, Epidemiology, and End Results (SEER) Program data looking at the influence of nodal involvement on outcome. Maibenco et al. (26) identified 1,229 cases of female microinvasive breast cancer treated between 1997 and 2003, using 1997 as their start date so that all cases were diagnosed in accord with the 1997 AJCC staging manual definition of microinvasive DCIS. Among the patients who underwent axillary staging, the frequency of nodal metastases was 10.5%. Five-year survival was 99% for node-negative, and 95% for node-positive women.

In conclusion, our institutional data demonstrate a 7% rate of axillary macrometastases in patients with DCISM. Axillary staging is indicated, and SNB is appropriate for the identification of axillary metastatic disease.