Lobular Neoplasia and Invasive Lobular Carcinoma

Lobular Neoplasia Versus Atypical Lobular Hyperplasia/Lobular Carcinoma In Situ

In 1978, Haagensen and colleagues retrospectively analyzed 211 examples of in situ lobular proliferations and introduced the term lobular neoplasia , a term that encompasses both ALH and LCIS. In their study, the microscopic qualitative and quantitative variations in the lobular proliferation were not found to have any value in predicting subsequent carcinoma. In the follow-up, 17.1% of the patients developed a “frank” carcinoma. The authors recommended systematic follow-up of patients as opposed to the recommendation of mastectomy suggested by Foote and Stewart. Since then, authorities in the field have argued on the most appropriate terminology to be used.

In the latest World Health Organization (WHO) classification of breast cancer, the term lobular neoplasia has been endorsed. In keeping with Haagensen and colleagues’ proposal, this was largely done to remove the word carcinoma from the diagnosis and alter the treatment of LCIS; a change that may no longer be justified now that breast-conserving surgery is first-line treatment for most breast cancers, unless clinicopathological features indicate mastectomy. Arguments in favor of this terminology also include optimal reproducibility for the differential diagnosis between ALH and LCIS and the fact that the subdivision into ALH and LCIS may not be of prognostic significance. Surprisingly and paradoxically, however, the advocates of the term lobular neoplasia have proposed a further subcategorization of lobular intraepithelial neoplasia (LIN) into three grades based on quantitative and qualitative morphological criteria and associated with clinical outcome. Therefore, there is a consensus that some quantification of the amount of in situ lobular proliferation is desirable and may be of help for clinical decision-making.

The three-tiered LIN system for the subclassification of in situ lobular lesions is based on the extent and degree of proliferation and/or cytological features. LIN1 is characterized by a partial or complete replacement or displacement of the normal epithelial cells of the acini within one or more lobules by a proliferation of lobular cells that may fill, but do not distend, the acinar lumina in comparison with adjacent uninvolved acini. LIN1 corresponds to what most observers would classify as ALH. Lesions in which there is abundant proliferation of similar cells as in LIN1 that fill and actually distend some or all acini, but without fusion of the acini (i.e., the acinar outlines remain distinct and separate from one another with persistence of intervening lobular stroma), are classified as LIN2. Those lesions with markedly distended acini that may appear almost confluent, often with central necrosis, and those composed of either pleomorphic cells or pure signet-ring cells with or without acinar distention are designated LIN3. In the initial publication, the incidence of invasive carcinoma of lobular type was significantly higher in patients with LIN3 (86% of all invasive carcinomas) as compared with LIN2 (47%) and LIN1 (11%). A subsequent study described that LIN3 and, to a lesser extent, LIN2 were associated with an increased risk of subsequent carcinoma, whereas LIN1 was not. Although this classification is currently used in some centers, it has not been endorsed by the WHO. Systematic validation has not been demonstrated and the criteria may differ among users. For instance, some observers may require the presence of necrosis to classify a lesion as LIN3. In fact, it is questionable whether this grading system improves the intraobserver and interobserver reproducibility over the ALH/LCIS classification. It should be emphasized that there is limited information on the clinical value of the proposed subclassification of in situ lobular lesions with the LIN system; hence, we would neither endorse nor encourage the use of this terminology.

Several authors have recommended to use the terms ALH and LCIS in surgical pathology reports, and to avoid whenever possible the vague term lobular neoplasia , which encompasses a wide spectrum of lesions. In addition, LCIS must be subclassified into classic or pleomorphic variants, and the presence of necrosis must also be documented because the pleomorphic variant and those cases with necrosis may warrant different therapeutic approaches. In some instances, however, in particular when dealing with core biopsies and when the lobular proliferation involves a preexisting lesion, the distinction between ALH and LCIS cannot be made with confidence. In these contexts, use of the term lobular neoplasia is advocated.

Clinical Presentation

No specific clinical presentation exists that allows one to infer the diagnosis of LCIS. The age at diagnosis varies widely, ranging from 15 to older than 90 years, but lobular neoplasia has been traditionally considered a disease of premenopausal women, with its incidence peak occurring about a decade earlier than DCIS peak. Haagensen and associates speculated that spontaneous regression of the disease would occur during menopause. Although lobular carcinoma is perhaps the prototype of endocrine-responsive proliferations of the breast, no direct evidence to support the contention of spontaneous regression is currently available. Furthermore, several recent reports have described a mean age at diagnosis as older than 50 years, and a substantial proportion or even the majority of LCIS patients are postmenopausal. In addition, there is evidence to suggest that pleomorphic lobular carcinoma in situ (PLCIS) tends to occur at an older age than classic LCIS, and in particular in those with apocrine histology.

Clinical Imaging

ALH and LCIS are microscopic lesions that do not form a palpable tumor and do not result in a grossly apparent alteration. Calcifications occur in less than 50% of cases. Contrary to other in situ lesions and benign breast proliferations, no imaging features can be used to reliably identify a lobular in situ proliferation. LCIS is typically discovered in breast tissue removed for other proliferative lesions that are prone to calcify, such as columnar cell change and hyperplasia and FEA (see Fig. 21.1 ), or for lesions that cause architectural distortion or a mass, such as fibroadenoma. Therefore, the diagnosis of ALH/LCIS is often considered an incidental finding. Retrospective analyses report that most cases of LCIS were classified as Breast Imaging Reporting & Data System (BI-RADS) 3 or 4. Mammography is not effective in detecting ALH/LCIS and is not recommended for assessing multicentricity or bilaterality of the disease.

It should be noted, however, that the common association of ALH/LCIS within the whole spectrum of columnar cell lesions/FEA is certainly not coincidental, reflecting the presence of low-grade–type molecular alterations (see later) in the affected breast.

Although the majority (≤90%) of LCIS cases are diagnosed in biopsies done for screen-detected calcifications, there is a general belief that lobular cells do not induce calcifications themselves unless they are associated with central necrosis, and that when diagnosed in imaging-guided biopsies for calcifications, the latter are directly associated with other co-occurring lesions. Nevertheless, the presence of calcium directly associated with LCIS has been a matter of debate in the literature. Early studies from the 1960s have suggested that punctate linear calcifications were the most common mammographic finding in patients who had biopsies that resulted in diagnoses of LCIS; however, it was conceded that those findings were not specific and could happen in patients with benign breast disease. Later, Pope and coworkers reviewed the mammographic features of 26 patients with biopsy-proven LCIS with no other abnormalities and concluded that there were no specific radiological findings for LCIS and that the diagnosis of LCIS in a biopsy obtained for calcifications was probably an incidental finding. More recent studies, however, have reported a high incidence of colocalization of calcium within nonnecrotic lobular neoplasia, as high as 42% of cases ( Figs. 21.4 and 21.5 ). Some authors have even highlighted the importance of identifying the subpopulation of LCIS with neoplastic cell calcifications because of its higher propensity to be associated with upstaging on follow-up excision biopsy. Those findings warrant further validation because others may argue that colocalization of calcifications within ALH/LCIS may also be incidental, resulting from colonization of calcium-studded adenosis; however, when performing pathological-radiological correlations, one should remember that LCIS, either classic (small punctate calcifications) or pleomorphic (clustered, large, and pleomorphic calcifications), may justify the calcifications detected on mammograms.

Gross Pathology

The whole spectrum of lobular neoplasia is not associated with any grossly recognizable features. When present, gross changes are mostly caused by other coexisting proliferative lesions, such as sclerosing adenosis or fibroadenoma. In cases of very florid classic LCIS or PLCIS, the cut surface of the breast tissue may display a faintly granular appearance because of the enlargement of the lobules.

Microscopic Pathology

LCIS is composed of acini filled with a monomorphic population of small, round, polygonal, or cuboidal cells, with a thin rim of clear cytoplasm and a high nuclear-to-cytoplasmic ratio ( Figs. 21.6 and 21.7 ; see also Fig. 21.1 ). The nuclei are uniform and the chromatin fine and evenly dispersed. Nucleoli, when present, are inconspicuous. A characteristic cytological feature is the presence of cells containing clear vacuoles, known as intracytoplasmic lumina or magenta bodies ( Fig. 21.8 ). When intracytoplasmic lumina or magenta bodies are found in a fine-needle aspiration (FNA) biopsy from a breast lesion, they are suggestive but not diagnostic of a lobular lesion (including ALH, LCIS, and ILC). The cells are loosely cohesive and regularly spaced and fill and distend the acini ( Figs. 21.9 and 21.10 ); however, overall lobular architecture is maintained. Glandular lumina are usually not seen in fully developed cases. Mitotic figures are present but not common, and necrosis is rarely seen.

Pagetoid spread, in which the neoplastic cells extend along adjacent ducts between intact overlying epithelium and underlying basement membrane, is also frequently found ( Figs. 21.11 to 21.13 ). Extralobular ductal involvement has been reported to occur in 65% to 75% of cases and may be the only manifestation in postmenopausal patients with atrophic breasts. In some cases, ductal involvement by lobular cells leads to the so-called cloverleaf pattern, which is characterized by clusters of neoplastic cells beneath the nonneoplastic epithelium and protruding outward around the periphery of the duct ( Fig. 21.14 ). True Paget’s disease, however, with extension of neoplastic cells to the squamous surface of the nipple, is not a feature of LCIS, with remarkably rare exceptions.

LCIS may colonize preexisting breast lesions, such as sclerosing adenosis ( Figs. 21.15 to 21.17 ), radial sclerosing lesions ( Fig. 21.18 ), papillomas, fibroadenomas, and collagenous spherulosis ( Fig. 21.19 ), leading to potential diagnostic errors. For instance, classic LCIS in sclerosing adenosis may be very subtle and missed ( Fig. 21.20 ). Conversely, LCIS in complex sclerosing lesions may be misinterpreted as ILC ( Fig. 21.21 ), whereas LCIS in collagenous spherulosis may be misdiagnosed as cribriform DCIS. In those cases, careful morphological observation is required, and when needed, the demonstration of loss of EAD and/or aberrant β-catenin or δ-catenin, or cytoplasmic p120 expression is the most reliable means of confirming the presence of an underlying lobular proliferation. In the authors’ experience, p120 catenin cytoplasmic immunostaining is the most sensitive and specific for detection of ALH/LCIS and ILC. In subtle lesions with few neoplastic lobular cells, however, membranous reactivity to those markers in adjacent residual epithelial and myoepithelial cells may result in an incorrect interpretation.

For a diagnosis of LCIS, Page and colleagues stated that more than half the acini in an involved lobular unit must be filled and distended by the characteristic cells, leaving no central lumina, whereas others required involvement of at least two lobules by the neoplastic proliferation. The latter criterion, however, may be inappropriate for the diagnosis of core needle biopsy (CNB) specimens that provide limited tissue samples. For practical purposes, distention translates as eight or more cells present in the cross-sectional diameter of an acinus. This cutoff should be applied with caution because, in atrophic breasts, acini filled with as few as five or six neoplastic lobular cells may actually be distended. Comparison with adjacent noninvolved acini may be preferred; however, in a single case, the size of noninvolved acini can vary considerably.

A lesion is regarded as ALH when it is less well developed and less extensive than LCIS as defined by the previous criteria (e.g., when the characteristic cells only partly fill the acini, with only minimal or no distention of the lobule; Figs. 21.22 and 21.23 ; see also Fig. 21.2 ). Lumina may still be identified, and the number of acini involved is usually less than half. Myoepithelial cells may be seen admixed with the neoplastic population. Clearly, the differentiation between ALH and LCIS on these criteria is somewhat arbitrary and prone to interobserver and intraobserver variability. Therefore, the use of the term lobular neoplasia to encompass the whole range of changes, and remove this variability, may be preferable for diagnostic purposes. However, as discussed previously, a justification for continuing to use the ALH/LCIS terminology is that ALH has been shown to have a lower risk of subsequent invasive carcinoma compared with LCIS.

The cells contained in classic ALH/LCIS, as described previously, can also be referred to as type A cells (see Fig. 21.7 ). A well-recognized subtype of LCIS is an architecturally similar lesion containing cells with mild to moderately large nuclei, some increase in pleomorphism, and more abundant cytoplasm, which is often clear. These cells are known as type B cells.

Both type A and type B cells can be seen in classic LCIS as well as the recently described florid variant of LCIS. The florid variant (FLCIS) differs from classic LCIS (CLCIS) in that the lobular arrangement is massively distended by LCIS cells, which can be either type A or type B cells ( Fig. 21.24 ). Punctate or comedo-type necrosis may be present in the distended lobules. In the study by Bagaria, CLCIS and FLCIS had the same incidences of associated ILC. In this same study of 210 consecutive LCIS cases, the prevalence of FLCIS was 19%. The study by Shin et al was the largest reported series of FLCIS cases analyzed by molecular profiling (array comparative genomic hybridization [aCGH]). Compared with CLCIS, FLCIS showed a significantly higher degree of genome instability with more chromosome losses, gains, and amplifications. In addition to the 1q+ and 16qʺ lobular signature, these lesions demonstrated recurring changes including 11qʺ, 17pʺ, and 8pʺ with a higher incidence than in CLCIS, where only 17pʺ was noted. An increased incidence of amplification at 11q13.3 (the region containing the CNND1 gene) was noted. Some FLCIS harbored amplification of 17q21 (the region spanning the HER2 gene), a finding not seen in CLCIS. These findings suggested that the genome FLCIS is an advanced neoplastic lesion compared with CLCIS. The findings imply that FLCIS is more akin genomically to PLCIS and may represent a point on the evolutionary scale to PLCIS.

In contrast to FLCIS, the neoplastic cells in PLCIS show marked pleomorphism and are distinctly larger, with abundant and often granular cytoplasm, a feature of apocrine differentiation ( Figs. 21.25 to 21.27 ). Nuclei are eccentrically placed and display conspicuous nucleoli. Signet-ring cells may be found in some cases. In contrast with what has been reported for the classic variant, apocrine differentiation at the morphological and IHC levels is very frequent in PLCIS, and a majority of these cases are HER2+ ( Fig. 21.25 ). In fact, some advocate that partial or overt apocrine differentiation is one of the most common features of PLCIS. The cells are often more dyscohesive than in CLCIS, and massive distention, central necrosis (comedo-type), and calcification in lobules are commonly found ( Fig. 21.26 ). PLCIS is often encountered in conjunction with cytologically similar pleomorphic ILC ( Fig. 21.27 ) and, occasionally, areas of transition between the two can be observed. Sneige and associates have described type B cells as containing nuclei that are up to twice the size of a lymphocyte (type A cells are 1 to 1.5 times larger), whereas PLCIS nuclei are typically four times larger and harbor more prominent nucleoli. Not so infrequently, different cell types are observed in the same case, even coexisting in the same TDLU ( Figs. 21.28 and 21.29 ). Recognition of the pleomorphic subtype is important because the combination of cellular features, necrosis, and calcification can lead to difficulty in differentiation from DCIS. PLCIS is an entity that causes diagnostic difficulties. PLCIS with marked pleomorphism is often misdiagnosed as high-grade DCIS, whereas the criteria to differentiate CLCIS with some degree of pleomorphism from a bona fide case of PLCIS are yet to be fully established.

Immunoprofile

The lack of membranous ECAD expression, as well as of other cell adhesion molecules, including β-catenin and catenin p120, characterizes ALH and LCIS and is useful for distinction from ductal proliferations ( Figs. 21.30 to 21.32 ). The p120 catenin is linked to the internal domain of ECAD, in which the presence of ECAD is the negative feedback inhibition of p120 catenin production. Therefore, both ECAD and p120 catenin are normally seen in/on the cell membrane in normal duct epithelium and in carcinomas of ductal type. When ECAD goes absent because of CDH1 mutation in lobular neoplasia, feedback inhibition of p120 catenin is lost, and it increases quantitatively and is seen in the cytoplasm instead of the cell membrane.

Caution should be exercised in the interpretation of these ancillary markers (see later). Over 90% of classic LCIS cases display strong expression of ER ( Fig. 21.33 ) and PgR in the majority of neoplastic cells. Conversely, the expression levels of ER and PgR may be low in PLCIS. In fact, a diagnosis of HR– LCIS should be seriously reconsidered unless it displays pleomorphic features. In this scenario, optimal internal controls ought to be demonstrated.

Whereas up to 5% of ILCs overexpress HER2 with gene amplification, most classic LCISs do not express biomarkers typically associated with an aggressive phenotype, being negative for p53 (as defined by >10% of neoplastic cells) and exhibiting a low proliferative (Ki-67) index. This profile is consistent whether LCIS is associated with invasive carcinoma or not (pure LCIS) and also with ILC. However, FLCIS and PLCIS more frequently display HER2 gene amplification (15%–30%) and protein overexpression, p53 IHC positivity (as a surrogate marker for TP53 mutation), and a higher proliferative index. Given the characteristic apocrine features of PLCIS, it is not surprising that these lesions are often positive for GCDFP-15 (gross cystic disease fluid protein-15).

Treatment and Prognosis

In their original description, Foote and Stewart highlighted the risk of development of subsequent invasive carcinoma in breast tissue harboring LCIS and recommended mastectomy for treating patients with LCIS. After their publication, numerous case reports appeared in the literature. Although some confirmed the development of ipsilateral and contralateral carcinomas after a biopsy displayed LCIS, others reported that a significant number of patients not undergoing mastectomy for LCIS remained well even with follow-up of up to 21 years. Retrospective pathology studies were then undertaken and demonstrated that the relative risk for subsequent development of invasive carcinoma among patients with lobular neoplasia (i.e., ALH and LCIS) ranges from about four to 12 times that expected in women without lobular neoplasia. It is currently widely accepted that LCIS confers an increased risk of development of invasive carcinoma of about 1% per year, a 10-year risk of 7% to 8%, and a lifetime risk of 30% to 40%. Of 1,174 women included in 18 separate retrospective studies, diagnosed as having lobular neoplasia and treated by biopsy alone, 181 (15.4%) eventually developed invasive carcinoma. Of these, 102 (8.7%) developed in the ipsilateral breast and 79 (6.7%) in the contralateral breast, indicating an almost equal risk for either breast. These data have been used to support the notion that lobular neoplasia may be only a risk indicator rather than a true direct precursor of invasive cancer. However, it is unclear whether those patients developing invasive cancer in the contralateral breast did not harbor bilateral in situ lobular neoplasia. Moreover, data from a prospective study of 100 cases of lobular neoplasia with 10 years of follow-up revealed that out of 13 invasive recurrences, 11 were ipsilateral.

Studies from Page and coworkers have demonstrated that the risk of subsequent carcinoma after a diagnosis of lobular neoplasia can be stratified according to the disease extent. By analyzing 39 examples of LCIS with an average overall follow-up of 19 years, the authors determined that the risk of invasive breast cancer among women with LCIS is eight to 10 times higher than among women whose biopsy lacked proliferative disease, whereas the risk is relatively lower but still definable (four to five times higher) after a diagnosis of ALH.

It should be noted that the cancer risk seems to increase with extended follow-up. Studies with longer follow-up tended to report a higher frequency of subsequent carcinoma. One study has suggested that the risk of development of invasive cancer increases to 35% for those women who survive 35 years after their initial diagnosis of lobular neoplasia. In addition, a subgroup analysis of participants in the Canadian National Breast Screening Study, published in 2014, showed that the probability of subsequent invasive breast cancer for women with LCIS and DCIS was 5.7% and 11.4%, respectively, after 5 years. In contrast, after 20 years, there was an equal risk for development of breast cancer after diagnosis of LCIS and DCIS (21.3% and 19%, respectively). Moreover, the relative risk increases substantially from 4.9 to 16.1 if a second biopsy shows lobular neoplasia. Efforts have also been made to identify patients with a significantly greater likelihood of developing invasive carcinoma after LCIS. The risk is higher among women with a family history of breast cancer and among nulliparous patients. However, those are general risk factors for all women, regardless of whether they have LCIS.

With the advent of national breast cancer screening programs around the world, there has been a vast increase in the number of investigations, including CNB, performed for screen-detected abnormalities. In a proportion of cases, this will inevitably lead to detection of lobular neoplasia in patients with calcifications associated with benign breast disease. LCIS and ALH are infrequently seen as the sole diagnostic finding in CNBs, accounting for 0.5% to 2.9% of biopsies taken for histological assessment of mammography-detected lesions. Peer-reviewed data and prospective analyses of lobular neoplasia in CNBs are limited, and therefore, most management recommendations have been based more on pragmatism than on scientific evidence. Until recently, most authors agreed that excision should be performed in cases of lobular neoplasia diagnosed on a CNB when:

• there is the presence of another lesion, which would itself be an indication for surgical excision, on the core biopsy (e.g., ADH or a radial scar[RS]);

• there is discordance among clinical, radiological, and pathological findings;

• there is an associated mass lesion or an area of architectural distortion;

• the lobular neoplasia shows mixed histological features with difficulty in distinguishing the lesion from DCIS, or shows a mixed ECAD staining pattern; or

• the morphology is consistent with that of the pleomorphic variant of lobular neoplasia.

It should be noted, however, that the previous approach has not been universally applied. For instance, some units have recommended and undertaken surgical diagnostic excision biopsies of all lobular neoplasias (including ALH and LCIS) diagnosed in CNBs, whereas other groups were excising only those cases defined as previously and, in particular, those with radiological/surgical/pathological discordance. Since 2009, North American authors have suggested that lobular neoplasia should be perceived as a high-risk lesion and excision should be recommended in all cases owing to the underestimation of cancer in up to 33% of lobular neoplasia diagnosed on CNBs. Interestingly, Esserman and colleagues analyzed a series of 26 cases of lobular neoplasia diagnosed on CNBs that were followed by excision biopsy and observed that invasive carcinoma was only found in cases in which the initial diagnosis was of diffuse lobular neoplasia, suggesting that the extent of lobular proliferation in the core biopsy may also be associated with the presence of invasive carcinoma.

Recently, Laws and colleagues reported the 3-year follow-up on patients who had lobular neoplasia on CNB and were managed without surgical excision. The authors concluded that with appropriate criteria as cited above, surgical excision may not be required.

Although it is important to avoid unnecessary diagnostic surgery for patients when ALH/LCIS is the sole finding in a CNB, the risk of associated malignancy in the adjacent breast at the time of diagnosis should be noted. Some have advocated that a multidisciplinary approach for such cases is essential and that each case must be assessed individually. It should also be noted that the paucity of large prospective studies to define accurately the risk of further aggressive lesions is problematic in clinical management.

Despite the rather limited data on PLCIS, there is circumstantial evidence to suggest that these lesions are more frequently associated with higher-risk lesions and may have a more aggressive clinical behavior than classic LCIS. Therefore, many recommend that such cases be subjected to further excision. In addition, the margin status may be relevant in cases of PLCIS. As opposed to CLCIS, pathologists are encouraged to describe in their surgical pathology reports whether PLCIS is present or not at the margins of resection. When present, current evidence (or lack thereof) favors reexcision. The data for the need to excise FLCIS if seen on CNB, and the need to address margin status for FLCIS, are even less clear. In the face of lack of data, the individualized approach may seem prudent—that is, extent of disease, family history, compliance with follow-up, and desire for breast preservation.

Differential Diagnosis

A few well-known pitfalls can occasionally cause diagnostic difficulty in the diagnosis of LCIS. Poor tissue preservation may lead to an artifactual appearance of dyscohesive cells in a lobular unit, resulting in overdiagnosis of LCIS. Similarly, foci of lactational change containing intracytoplasmic lipid droplets or clear cell metaplasia may superficially resemble ALH/LCIS to the unwary. LCIS cells, however, may display clear cell features ( Figs. 21.34 and 21.35 ); in this situation, distinction from hyperplasia of myoepithelial cells with clear cytoplasm or vacuolated myoepithelial cells often found in the luteal phase of the menstrual cycle may require the use of ancillary IHC markers.

Another difficulty arises when LCIS is growing in some types of benign breast lesions (i.e., sclerosing adenosis and RS), which clinically and radiologically can present as a mass (see Figs. 21.15 to 21.21 ). The histological appearance of LCIS in association with these lesions may be misleading, with distortion of lobular units and a rather sclerotic stroma. The combination of abnormal architecture and proliferative lobular cells can easily be diagnosed as an invasive carcinoma by the unwary (see Fig. 21.21 ). In this situation, low-power examination is recommended to appreciate the lobular architecture. In difficult cases, IHC to demonstrate the myoepithelial cell layer, in particular with a combination of nuclear (e.g., p63) and cytoplasmic (e.g., smooth muscle myosin heavy chain [SMM-HC] or calponin) myoepithelial markers (see Fig. 21.16 ), or the basement membrane, is useful in making the distinction.

Perhaps the most important, and also the most difficult, differential diagnosis of CLCIS is with DCIS of the solid, low nuclear grade type. A diagnosis of DCIS carries wholly different management implications for a patient because it mandates surgical excision with or without radiation therapy as definitive treatment, whereas LCIS may arguably warrant only follow-up or tamoxifen to reduce the risk of subsequent breast cancer development. Correct identification, therefore, is essential. The distinction of LCIS from low-grade solid DCIS is challenging because morphologically they may be remarkably similar, especially when DCIS involves the acini (termed cancerization of lobules ) with minimal or no lobular distortion. Morphological indicators include nuclear size and pleomorphism, which may be greater in DCIS, and the presence of secondary lumen formation and cellular cohesion that also point to DCIS rather than LCIS. IHC analysis of the lesion can prove useful in making the distinction. ECAD and β-catenin are typically absent or aberrant in ALH/LCIS (see Figs. 21.24 and 21.25 ) but present on the membrane of neoplastic cells in DCIS. It should be noted, however, that membranous positivity for ECAD does not preclude the diagnosis of LCIS. Some bona fide LCIS may display aberrant ECAD membranous expression, which is often fragmented and distinct from the expression in residual epithelial or DCIS cells ( Figs. 21.36 to 21.38 ). In some of those cases, β-catenin may be of help because it may be lost, indicating that, although ECAD is present on the membrane, it is dysfunctional, not associating correctly with the cadherin-catenin complex ( Figs. 21.39 to 21.41 ). Another, more useful marker is catenin p120, which is expressed on the cell membranes of DCIS cells but found in the cytoplasm of ALH/LCIS cells. In cases of bona fide ALH/LCIS with ECAD expression, cytoplasmic expression of catenin p120 can be used to corroborate a diagnosis of ALH/LCIS. Occasionally, lesions show an overlapping range of morphological features along with variable expression of IHC markers. This suggests that LCIS and low-grade solid DCIS may truly coexist within the same duct–lobular unit. In these circumstances, differentiation between the two is often not possible and both diagnoses should be given. How a patient should be managed in these unresolved cases remains a challenge, but pragmatically, they will receive treatment as for DCIS.

Likewise, PLCIS must be differentiated from high-grade solid DCIS, given that both lesions display similar features, including high nuclear grade, comedonecrosis, and calcifications. Owing to the massive distention of the TDLUs, PLCIS may not appear so dyscohesive and pose diagnostic problems. In this context, IHC plays an essential role because PLCIS must show downregulation of ECAD, β-catenin, and cytoplasmic upregulation of catenin p120 ( Figs. 21.42 to 21.44 ). Nevertheless, the same caution as described previously must be exercised when analyzing the immunostains. A small proportion of PLCIS may show focal membranous positivity for ECAD.