Foote and Stewart (1) had coined the term “lobular carcinoma in situ” (LCIS) for a group of in situ carcinomas of the breast that occurred in the terminal ducts and lobules and were characterized by loss of cellular cohesion, presence of cytoplasmic vacuoles, pagetoid extension, and multifocality.

Classic LCIS (C-LCIS) refers to an architectural pattern of LCIS in which the affected acini are minimally distended, and the affected lobules are generally scattered amid breast tissue (Figs. 18.1, 18.2, 18.3). In objective terms, distention is defined as “the presence of 8 or more cells in the cross-sectional diameter of an acinus” (2). C-LCIS may be populated by cells of type “A” or type “B.”

LCIS with type “A” cytology is characterized by smaller cells (<1.5X size of lymphocyte) with minimal cytoplasm, low-grade nuclei, inconspicuous nucleoli, and extremely rare mitoses. LCIS with type “B’ cytology is typified by larger cells (2X size of lymphocyte) with relatively more cytoplasm, intermediate-grade nuclei, presence of micronucleoli, and rare mitoses (Fig. 18.4).

The classification of C-LCIS into types “A” and “B” is of no known clinical significance; however, such categorization serves as a reminder that “some cytologic variation can be observed in bona fide cases of C-LCIS, and these features should not be over-interpreted as representing the pleomorphic variant of LCIS” (2). Nevertheless, in recent years, the cytologic classification of LCIS as “type A” and “type B” has been largely abandoned, in favor of grouping into classic, florid, and pleomorphic variants.

LCIS can be architecturally as well as cytologically (if not molecularly) heterogeneous. Minimal distention of acini is the hallmark of C-LCIS, whereas confluent lobular involvement with marked distention of acini is observed in florid LCIS (F-LCIS). The pleomorphic variant of LCIS (P-LCIS) is typically characterized by high-grade nuclei populating LCIS with florid-type features.

F-LCIS may be populated by cells of types “A” or “B” and can show central necrosis (Fig. 18.5). F-LCIS with signet ring cells has also been described (3). F-LCIS shares several features with C-LCIS including some cytologic characteristics, loss of E-cadherin, gain of 1q, and loss of 16q; however, F-LCIS has been shown to demonstrate more genomic alterations than C-LCIS (4).

P-LCIS refers to an architectural as well as cytologic pattern of LCIS in which there is diffusely confluent involvement of several contiguous lobules, and the individual acini are considerably distended (Fig. 18.6). P-LCIS is populated by cells with variable size and shape (hence, the designation: pleomorphic). P-LCIS cells are much larger (4X size of lymphocyte) and bear high-grade nuclei, irregular nuclear membrane, prominent nucleoli, and frequent mitoses. Central necrosis of the so-called “comedo” type usually accompanies P-LCIS. LCIS that is cytologically pleomorphic can be rarely found with classic architectural morphology in the vicinity of typical examples of P-LCIS. Some cases of P-LCIS possess abundant eosinophilic (apocrine-type) cytoplasm.

The issue of how much lobular involvement is necessary for the diagnosis of LCIS is of questionable relevance in the diagnosis of needle core biopsies (NCBs) that provide limited samples. Because the number of affected lobules in a biopsy has not proven to be related to the risk for subsequent carcinoma among patients not treated by mastectomy (5,6), there is presently no reason for drawing a distinction between one and two or more involved lobules as a basis for the diagnosis of LCIS (5). In some instances, the only evidence of a neoplastic lobular proliferation is one lobule in which some, but not all, acini are involved. It has been suggested, rather arbitrarily, that at least 50% (7) or 75% (8) of one lobule should be involved to establish a diagnosis of LCIS. Specimens with lesser lesions should be included in the category of atypical lobular hyperplasia (ALH).

The umbrella diagnostic term lobular neoplasia (LN), encompassing ALH and LCIS, was introduced by Haagensen et al. (9) in an era when breast conservation therapy was uncommon mainly for the purpose of avoiding overtreatment of the disease by removing the word “carcinoma” (Fig. 18.7). This rationale is no longer valid in the era of breast conservation. Furthermore, the spectrum of lesions encompassed by LN, from ALH to P-LCIS, is so broad as to render it a useless term. The use of this misleading term demonstrates a lack of regard for criteria-based diagnosis.

Gomes et al. (10) have reported good interobserver agreement between general pathologists and specialized breast pathologists in the diagnosis of ALH (Kappa = 0.62) and LCIS (Kappa = 0.66); however, poor agreement was observed for P-LCIS (Kappa = 0.22). It is likely that the presence of overt central necrosis in P-LCIS is the most likely reason for it to be mistaken for high-grade DCIS.

The classic form of lobular carcinoma in situ (C-LCIS) is a lesion that is evident only on microscopic examination; it almost never forms a palpable tumor (11). C-LCIS is rarely detectable with various imaging modalities. Typically, C-LCIS is discovered incidentally in breast tissue biopsied for lesions that produce masses or cause abnormalities on imaging studies. Calcifications are infrequently formed in C-LCIS (12), and when present are usually associated with an underlying lesion. No radiologic technique has proven to be an effective method for detecting C-LCIS, and thus imaging studies cannot be depended upon to assess the presence of the disease (5,13).

Exceptional situations exist in the relatively infrequent instances of F-LCIS and P-LCIS, the two variant forms of LCIS. These forms of LCIS show marked expansion of acini within lobules. F-LCIS and P-LCIS may involve the breast extensively, and can sometimes involve foci of sclerosing adenosis. As stated earlier, F-LCIS can be populated by cells with either low-grade (type “A”), intermediate-grade (type “B”), or high-grade (“pleomorphic”) nuclei. Necrosis and calcification can be associated with F-LCIS inhabited by neoplastic cells that possess either intermediate- or high-grade nuclei, and with nearly every case of P-LCIS. Such lesions display a pattern and distribution more commonly encountered with ductal carcinoma in situ (DCIS) rather than LCIS (14). Thus, the resultant appearance on mammography is likely to suggest DCIS. The most compelling evidence supporting classification of such cases as LCIS is the lack of E-cadherin immunoreactivity as well as the finding of associated C-LCIS and/or invasive lobular carcinoma in some cases.

In a retrospective cohort analysis of SEER (Surveillance, Epidemiology and End Results) database, including 14,048 patients diagnosed with LCIS, a 38% increase in the incidence of LCIS (15) was reported, from 2.0/100,000 to 2.75/100,000, from 2000 to 2009. Ten percent of the LCIS patients reportedly underwent biopsy alone, seventy-four percent underwent excision, and sixteen percent had mastectomy.

In retrospective reviews, published in previous decades, each involving several thousand breast specimens, the frequency of LCIS was 0.5% to 1.5% (16,17,18). Analysis of population-based data from 1978 to 1998 in the United States revealed an increase in the incidence rate of LCIS from 0.90 per 100,000 person years to 3.19 per 100,000 person years (19). Incidence rates increased continuously throughout the two decades-long study period among postmenopausal women with the highest incidence rate among women 50 to 59 years of age in 1996 to 1998 (11.47/100,000 person years). The absence of consistent pathology review creates uncertainty about the reliability of the data, but if it is correct, the increasing use of breast imaging leading to more frequent biopsies is probably the most important factor responsible for this change. In particular, columnar cell alterations, which are predisposed to develop calcifications, often coexist with LCIS (20,21).

Parenthetically, The Rosen Triad comprises LCIS, columnar cell changes, and tubular carcinoma (Fig. 18.8) (22). The latter two lesions are often responsible for mammographically detected calcifications in the absence of a palpable lesion. ALH and LCIS are commonly associated with various types of columnar change (Fig. 18.9). An instance of the Rosen Triad with LCIS of the signet ring cell type has been reported (23).

Up to 25% of LCIS patients are postmenopausal at the time of first diagnosis (24). LCIS occurs infrequently in women younger than 35 years. In a consecutive series of more than 1,000 patients treated for breast carcinoma, the mean age of women with LCIS (53 years) was not significantly different from the mean age of patients who had invasive ductal carcinoma (57 years) (25).

Some conclusions based on the earliest studies of LCIS are still valid and provide invaluable insight into its behavior. These conclusions include the relatively high frequency of ipsilateral multicentric carcinoma, including occult invasive carcinoma in 5% to 6% of women who undergo mastectomy (26,27). Urban (28) had reported finding LCIS in the contralateral breast in

40% of the cases when a random biopsy was performed even in the absence of any clinical abnormality. This procedure is rarely done today.

The anatomic distribution of LCIS in lobules and terminal ducts (and occasionally in ducts) as well as morphologic alterations in these structures influence the histopathologic appearance of LCIS in any given case. In the typical lobular form, a population of neoplastic cells replaces the normal epithelium of acini within lobules. Intralobular ductules are also characteristically involved (Fig. 18.10). The LCIS cells must be sufficiently numerous to cause expansion of these structures. There may be enlargement of the entire lobule in comparison with uninvolved lobules in the adjacent breast tissues; however, the lobular enlargement is not an absolute diagnostic criterion
(Fig. 18.11). The drift to lobular atrophy in postmenopausal women makes expansion of lobules an unreliable diagnostic feature in that patient group. If the diagnosis of LCIS is to be meaningful because it identifies a lesion associated with a substantial risk of later carcinoma, then the architectural or quantitative finding of lobular enlargement cannot be regarded as the paramount diagnostic criterion in lesions that have reached an acceptable qualitative level of cytologic abnormality.

Loss of intercellular cohesion is a characteristic of LCIS, although this is not always readily apparent in acini filled and expanded by the process. When loss of cohesion is prominent, the resultant spaces may be mistaken for glandular lumina.
Degenerative changes may also disrupt the cellular composition of LCIS. In these situations, the neoplastic cells are not arranged in the polarized fashion. The latter is a trait of non-neoplastic cells persisting around true glandular lumina. Loss of cohesion in lobular neoplastic lesions is attributable to genetic alterations in the E-cadherin gene that are manifested by greatly reduced, fragmented, or absent membrane E-cadherin immunoreactivity (29,30).

Intracytoplasmic vacuoles that contain mucin are present in some LCIS cells (Fig. 18.12). The presence of mucin can
be an inconspicuous feature that can be highlighted with mucicarmine, Alcian blue, or periodic acid-Schiff (PAS) stains (31,32). An extreme manifestation of this phenomenon is the formation of signet ring cells having a distended cytoplasmic vacuole that causes the nucleus to appear eccentric, crescentic, or indented. Signet ring cells can have low-, intermediate-, or high-grade nuclei. Intracytoplasmic mucin vacuoles are uncommon in ductal carcinoma cells and are also uncommon in hyperplastic lesions of ductal or lobular epithelium. Thus, the presence of intracellular mucin is an important but not a necessary criterion for the diagnosis of LCIS. Intracytoplasmic mucin is also present in LCIS with clear cell change.

Several uncommon cytologic features can be found in LCIS. Cytoplasmic pallor or cytoplasmic clearing occurs rarely in LCIS. These cells may have intracytoplasmic mucin that is not restricted to vacuoles, an occurrence manifested by diffuse cytoplasmic staining with the mucicarmine or Alcian blue stain. Apocrine change has been described in LCIS (33). Mucin in the cytoplasm of cells in apocrine LCIS is usually evident in H&E-stained sections as cytoplasmic amphophilia or basophilia. Apocrine LCIS often has pleomorphic cytology and tends to grow in a pagetoid manner into ducts where it can be difficult to distinguish from apocrine DCIS. However, apocrine LCIS is E-cadherin-negative, and it is often estrogen receptor (ER)-positive, whereas apocrine DCIS is E-cadherin-positive and almost always ER-negative. LCIS in atrophic lobules and terminal ducts of postmenopausal women sometimes features cells with dark, eosinophilic-to-basophilic cytoplasm and deeply basophilic, eccentric nuclei. This appearance is probably the result of cytoplasmic condensation associated with loss of cohesion and shrinkage of cells. These cells frequently contain intracytoplasmic mucin. In another variant, the cells of LCIS have a mosaic appearance that results from the presence of distinct cell borders between cells and prominent, round, centrally placed nuclei surrounded by pale cytoplasm (Fig. 18.13). Intracytoplasmic mucin vacuoles can usually be found in this type of LCIS.

LCIS typically involves intralobular and extralobular or terminal ductules as well as acinar units within the lobule (Fig. 18.14). In postmenopausal patients with atrophic lobules, ductal involvement may be the only manifestation of LCIS (34). The irregular configuration of ductules affected by LCIS has been described as “saw-toothed” or as resembling a cloverleaf. Pagetoid LCIS cells growing beneath the non-neoplastic ductal

epithelium may be distributed continuously or discontinuously along the ductal system, undermining, and ultimately displacing the normal ductal epithelium. The cloverleaf pattern of LCIS sometimes involves the terminal ducts, seemingly without lobular involvement. The myoepithelial layer is preserved to a variable extent, and it may require p63 and actin immunostains to confirm that it is present. LCIS can be found in lesions with intra-acinar myoepithelial cell hyperplasia (35).

Pagetoid spread of LCIS may also be encountered in papillomas or radial scar lesions. F-LCIS may proliferate to form a solid mass of tumor cells that fill and expand the ductal lumen, and develop central necrosis and calcifications that are detectable in mammograms. A negative, weak, or fragmented E-cadherin immunostain distinguishes this pattern of F-LCIS from E-cadherin-positive solid type of high-grade DCIS (36).

An unusual pattern of ductal involvement occurs when LCIS develops in ducts altered by collagenous spherulosis (37,38). This configuration mimics cribriform DCIS (Fig. 18.15). However, myoepithelial cells outline the spherule material. The latter should be distinguished from true microlumina of cribiform carcinoma. The neoplastic cells in such foci display loss of cohesion and intracytoplasmic vacuoles characteristic of lobular carcinoma. When LCIS involves collagenous spherulosis, myoepithelium can be highlighted with immunostains such as p63 and myosin. LCIS cells will be E-cadherin-negative. LCIS in collagenous spherulosis only rarely occurs as an isolated finding even in NCB samplings. In the series reported by Eisenberg and Hoda (38), LCIS was identified beyond the index lesion in 4 of 38 cases (14%). In this series, 22 of the 38 specimens were NCB samplings.

Losses in chromosome 16q22.1, the site of the CDH1 gene, are the most common cytogenetic abnormality found in LCIS (39). E-cadherin is the protein product of the CDH1 gene. Despite the high frequency of somatic E-cadherin mutations
in LCIS, germline E-cadherin mutations are rarely detected in these patients (39,40).