Improvements in DCIS Pathologic Diagnosis

CHICAGO—The diagnosis of ductal carcinoma in situ (DCIS) usually is straightforward from a pathologic point of view. However, microinvasion and histologically ambiguous lobular carcinoma in situ (LCIS) lesions may be difficult to distinguish from DCIS, Stuart Schnitt, MD, said at the Second Annual Lynn Sage Breast Cancer Symposium, sponsored by Northwestern University Medical School.

Microinvasion was not clearly defined until 1997, said Dr. Schnitt, associate professor of pathology, Harvard Medical School, and director of surgical pathology, Beth Israel Deaconess Medical Center, Boston.

At that time, the AJCC/UICC staging manual created the T1_mic category for breast cancer cells that extend beyond the basement membrane and have no single focus that is greater than 0.1 cm or that exceeds 1.0 mm at its greatest dimension.

Because of the lack of agreement about the pathologic characteristics of the microinvasion disease state, clinical studies applied a wide variety of descriptions. These include evidence of stromal invasion, limited microscopic stromalization invading less than 10% of the surface of the histologic specimen, maximal invasion that is 2.0 mm or less or represents less than 10% of the tumor, and site of invasion that is 1.0 cm in size or less.

Pathologic changes associated with DCIS or tissue sampling also can be mistaken for microinvasion, such as ductal branching, DCIS involving lobules or benign sclerosing lesions, tangential sectioning of the involved ducts, crush artifact, cautery effect, or water vacuole displacement.

The clinical significance of microinvasion is not known largely because the literature contains so many different definitions of the disease state, Dr. Schnitt said.

In about a dozen small clinical studies, the incidence of axillary lymph node involvement in patients who were given the diagnosis of microinvasion ranges from 0% to 20%. Outcomes studies of patients with microinvasion provide few data about disease-free and overall survival, and no clear survival comparisons with patients who have pure DCIS.

Use of the AJCC/UICC T1_mic stage should improve the recognition of microinvasion as well as the understanding of the effect of this form of breast cancer on patients, Dr. Schnitt said.

He believes that in centers that apply the strict AJCC/UICC definition of microinvasion, patients should have a low rate of axillary lymph node metastasis and a cure rate approaching 100% after appropriate localized therapy.

Distinguishing From LCIS

Because of differences in treatment, lobular carcinoma in situ must be differentiated from ductal carcinoma. However, pathologists are more frequently encountering difficulty distinguishing between DCIS and LCIS, perhaps because the increasing use of screening mammography is revealing unusual forms of breast abnormalities, Dr. Schnitt said.

Although most cases of lobular carcinoma can be easily characterized path-ologically, some lesions have indeterminate histologic features that are similar to those of DCIS. LCIS may extend into extralobular ducts, and DCIS may involve obvious lobules.

Even highly experienced pathologists have difficulty establishing a diagnosis when small-cell in situ carcinoma lesions have features that suggest microinvasion or features of both DCIS and LCIS, he pointed out.

Immunostaining for the adhesion molecule E-cadherin may provide a molecular basis for characterizing LCIS in questionable situations, Dr. Schnitt said.

A number of recent clinical studies indicate that DCIS cells consistently stain strongly for E-cadherin regardless of the stage of the lesion, while invasive LCIS cells do not express the adhesion molecule. Several studies have shown that invasive LCIS has a mutation in the gene for E-cadherin expression.

Dr. Schnitt studied a series of DCIS tissue samples, as well as samples of carcinoma in situ that were indeterminate, from his pathologic consulting practice at Beth Israel Deaconess Medical Center.

In all cases, immunostained DCIS samples were positive for E-cadherin, showing intense membrane staining for the molecule. All immunostained samples of LCIS were negative for E-cadherin, with no evidence of adhesion molecule staining.

Some lesions that were indeterminate on hemoxylin and eosin staining had strong E-cadherin expression, suggesting a closer kinship to DCIS than LCIS, he said.

Dr. Schnitt concluded that immuno-staining for E-cadherin may be a useful adjunct in determining whether an in situ lesion is LCIS or DCIS in histologically ambiguous cases. Such testing, therefore, may more definitively guide clinicians to attempt eradication of breast lesions in women who most likely have DCIS or to follow patients with suspected LCIS as a marker of an overall increased risk of breast cancer.