Comprehensive Review of Molecular Mechanisms and Clinical Features of Invasive Lobular Cancer

Abstract

Invasive lobular carcinoma (ILC) accounts for 10% to 15% of breast cancers in the United States, 80% of which are estrogen receptor (ER)‐positive, with an unusual metastatic pattern of spread to sites such as the serosa, meninges, and ovaries, among others. Lobular cancer presents significant challenges in detection and clinical management given its multifocality and multicentricity at presentation. Despite the unique features of ILC, it is often lumped with hormone receptor‐positive invasive ductal cancers (IDC); consequently, ILC screening, treatment, and follow‐up strategies are largely based on data from IDC. Despite both being treated as ER‐positive breast cancer, querying the Cancer Genome Atlas database shows distinctive molecular aberrations in ILC compared with IDC, such as E‐cadherin loss (66% vs. 3%), FOXA1 mutations (7% vs. 2%), and GATA3 mutations (5% vs. 20%). Moreover, compared with patients with IDC, patients with ILC are less likely to undergo breast‐conserving surgery, with lower rates of complete response following therapy as these tumors are less chemosensitive. Taken together, this suggests that ILC is biologically distinct, which may influence tumorigenesis and therapeutic strategies. Long‐term survival and clinical outcomes in patients with ILC are worse than in stage‐ and grade‐matched patients with IDC; therefore, nuanced criteria are needed to better define treatment goals and protocols tailored to ILC's unique biology. This comprehensive review highlights the histologic and clinicopathologic features that distinguish ILC from IDC, with an in‐depth discussion of ILC's molecular alterations and biomarkers, clinical trials and treatment strategies, and future targets for therapy.

Implications for Practice

The majority of invasive lobular breast cancers (ILCs) are hormone receptor (HR)‐positive and low grade. Clinically, ILC is treated similar to HR‐positive invasive ductal cancer (IDC). However, ILC differs distinctly from IDC in its clinicopathologic characteristics and molecular alterations. ILC also differs in response to systemic therapy, with studies showing ILC as less sensitive to chemotherapy. Patients with ILC have worse clinical outcomes with late recurrences. Despite these differences, clinical trials treat HR‐positive breast cancers as a single disease, and there is an unmet need for studies addressing the unique challenges faced by patients diagnosed with ILC.

Short abstract

Invasive lobular carcinoma is the second most common subtype of invasive breast cancer and presents unique challenges in detection and treatment. This review reports on the distinct morphological, molecular, and clinical features of invasive lobular carcinoma and highlights the challenges in the management of this cancer.

Introduction

Breast cancer is the most commonly diagnosed cancer among women (excluding skin cancer), with 276,480 new diagnoses in 2020 expected in the United States alone [1]. One in eight women will be diagnosed with breast cancer in their lifetime [2]. Invasive lobular carcinoma (ILC) is the second most common subtype, accounting for 10% to 15% of invasive breast cancer [3]. ILC presents unique challenges in detection and treatment, and yet it remains understudied in research. In this review, we will discuss the distinct morphological, molecular, and clinical features of ILC and highlight the challenges in the management of this cancer.

Materials and Methods

A comprehensive literature search of the PubMed database was conducted. The keywords used were “invasive lobular carcinoma,” “E‐cadherin,” “endocrine therapy resistance,” “HER2,” “FOXA1,” “FGFR,” “WNT,” and “p120 catenin.” All articles, including clinical trials, reviews, and original manuscripts, were considered. The search term “invasive lobular carcinoma” yielded 4,496 search results; thus, the search criteria was modified as follows to narrow down results: “invasive lobular carcinoma AND E‐cadherin” produces 335 results; “invasive lobular carcinoma AND p120 catenin” yields 24 results. Selected articles from the database search are included in this article. For clinical trials, further information was obtained through ClinicalTrials.gov.

Incidence, Mortality, and Clinical Features

The incidence of ILC has risen in recent years. Studies based on Surveillance, Epidemiology, and End Result (SEER) [4, 5] database found ILC rates increased from 9.6% in 1987 to 15.6% in 1999, with invasive ductal carcinoma (IDC) rates staying relatively constant. This increase in risk was associated with use of combined hormone replacement therapy (HRT) of estrogen and progestin, or progestin alone. Following the dramatic reduction in use of HRT, there was a nonsignificant downward trend in incidence of ILC for the period of 2002 to 2006 [6].

The incidence of ILC is highest in White women; however, Black women experience higher mortality from ILC. A 2017 study examined 18,295 patients with ILC from the SEER database, which included White women (n = 15,936), Black women (n = 1,451), and women of other races (American Indian/Alaskan Native and Asian/Pacific Islander, n = 908) found that the 5‐year overall survival (OS) and 5‐year breast cancer–specific survival were worst in Black patients with ILC (76.0% and 84.4%, respectively) compared with White women and women of other races (85.5% and 87.7%, p < .01; 91.1%, and 91.6%, p < .01, respectively) [7]. Other studies have reported similar 5‐year survival rates, approximately 86% for patients with ILC when not differentiating by race [8].

ILC is distinct from other breast cancer subtypes in its clinical, radiographic, and histopathology features (Table 1). Patients with ILC are generally older than other histological subtypes [8, 9, 10]. ILC has higher rates of multicentricity as well [11, 12]. The tumors can be large and are frequently bilateral, with lower rates of breast preservation [7]. ILC does not usually form a discrete palpable mass and is more difficult to detect through physical examination or imaging until at advanced stages [10, 11]. Radiographically, ILC may present as a spiculated mass or architectural distortion; however, mammography has lower sensitivity in the detection of ILC compared with IDC, resulting in potential false negatives reported to occur in 29.9% of ILCs [13]; ultrasound and MRI may be more helpful but could result in overestimation of tumor size or false positives [14, 15]. The classic subtype of ILC is composed of discohesive, small, round to ovoid cells with minimal cytoplasm and occasional cytoplasmic inclusions. Typically, these cells infiltrate the stroma as single cells or in single‐file lines, with minimal host reaction. Loss of membranous E‐cadherin expression as detected by immunohistochemistry (IHC) is seen in >90% of cases [3, 8], with concomitant loss of membranous p120 (increased cytoplasmic expression) and beta‐catenin due to disruption of the E‐cadherin complex [16]. The morphology and immunophenotype of ILC is due to inactivation of the cell–cell adhesion molecule E‐cadherin [8]. It results in the inability of tumor cells to adhere to one another [3] and leads to its unique histomorphology and mammographic presentation [10, 11]. Classic ILC is almost exclusively estrogen receptor (ER) and progesterone receptor (PR) positive and HER2 negative [8]; however, the pleomorphic variant shows higher rates of HER2 positivity [17]. In comparison with IDC, ILC is more likely to have metastases in the luminal gastrointestinal tract, cerebrospinal fluid and leptomeninges, and ovary [18, 19, 20].

Table 1.

Distinct clinical and pathological features of ILC

Clinical/pathological differences	IDC	ILC
Pathology	Pleomorphic with variability in cell size and shape [22] Prominent nucleus [22] Sheets, nests, cords, or individual cell tumor makeup [22] No myoepithelial cell lining, differentiating from DCIS [22]	Small, rounded, nucleated, and discohesive cells [22] Lack of E‐cadherin in about 80% of ILC cases More low‐grade tumors than IDC [104]
Description of growth	Originating in mammary ducts [105] Clustered growth with calcifications [106]	Originating in milk‐producing lobules [107] Single‐file of cells forming web‐like lines through stroma
Clinical features	Most common histological subtype of breast cancer Accounts for about 80% of all breast cancers [8]	Second most common histological subtype of breast cancer Accounts for 8%–14% of all breast cancers Worse overall prognosis [66] Often diagnosed in postmenopausal women [23]
Detection	Commonly presented as defined, firm nodule [104] Typically diagnosed in women at a younger age [106] Less false‐negative mammograms compared with ILC [106]	Lack of discrete, palpable mass More difficult to detect in later stages Higher false‐negative mammograms than IDC [106]
Comments	Highly observed heterogeneity in histologic and morphologic characteristics [109]	More likely to have metastases in GI tract and ovary [108] Less breast preservation [108] Higher frequency of contralateral breast cancer [108]

Abbreviations: DCIS, ductal carcinoma in situ; GI, gastrointestinal; IDC, invasive ductal cancer; ILC, invasive lobular cancer.

Histological Subtypes

Classic type ILC rarely induces a host response, does not typically destroy the architecture of breast, and grows as single‐file concentric patterns around ducts and lobular units of the breast. Patients diagnosed with classic type ILC often have associated lobular carcinoma in situ preceding diagnosis and have better overall prognosis than other ILC variant diagnoses [21, 22, 23]. However, long‐term survival is lower compared with stage‐ and grade‐matched IDC [24]. Pleomorphic lobular cancer (PLC) follows the typical ILC growth pattern but with marked nuclear pleomorphism (larger cell size with more abundant cytoplasm, more prominent nucleolus [Fig. 1], and higher mitotic rate compared with classic type ILC) [25]. Genetically, PLC shows similar genomic alterations to classic type ILC (gains in 1q and 16p, losses of 11q and 16q) along with additional genomic alterations characteristic of high‐grade IDC. PLC shows lower expression of hormone receptors (HR) and high rates of HER2 expression (9.7%–37%). Studies show PLC as a more aggressive variant with higher stage and lymph node involvement and shorter survival compared with classic type ILC [25, 26, 27, 28]. Solid variant ILC consists of discohesive lobular cells that grow in sheets throughout the breast rather than single‐file patterns, with higher grade nuclear atypia and mitotic rate compared with classic ILC [3, 21, 23]. Alveolar ILC is characterized by cells arranged in round aggregates composed of at least 20 classic type ILC cells [21, 23]. Signet ring cell variant of ILC is architecturally similar to classic ILC but contains more than 20% with signet cell morphology [22, 29]. Cells are differentiated from other variants by significant intracellular mucin displacing the nucleus within the cell. Other variants include tubulolobular carcinoma and histiocytoid variant of ILC, the latter associated with shorter survival [22, 30].

Figure 1.

Analysis using TCGA data set. (A): CDH1 alteration frequency in different breast cancer subtypes (TCGA dataset). (B): CDH1 mutations in invasive lobular cancer primarily lead to shallow deletions. Plot shows CDH1 mRNA expression versus copy number alteration due to different mutation resulting in deep deletion, shallow deletion, or gain of function. Each dot represents a sample, color coded for nature of mutation.Abbreviations: GISTIC, genomic identification of significant targets in cancer; NOS, not otherwise specified; RSEM, RNA‐seq by expectation–maximization; TCGA, The Cancer Genome Atlas; VUS, variant of uncertain significance.

Key Molecular Alterations in ILC

Loss of E‐cadherin is the Hallmark Mutation in ILC

One of the characteristic features of ILC is the loss of E‐cadherin. The E‐cadherin gene, CDH1, is located on human chromosome 16q22.1 and codes for a 120kDa single transmembrane glycoprotein [31]. Binding of calcium to its extracellular domain regulates activity of its transmembrane domain, whereas binding of alpha, beta, gamma, and p120 catenins to its cytoplasmic tail links it to the actin cytoskeleton and sequesters catenin from participating in downstream signaling [23, 32]. The key downstream signaling pathways include Hippo, WNT, TGFβ, NFκB, and growth signaling pathways [33, 34, 35]. E‐cadherin supports cell–cell adhesion and regulates motility, and its loss leads to epithelial‐to‐mesenchymal transition. Homophilic binding of cadherins leads to contact inhibition of cell proliferation by activating growth inhibitory signals through the Hippo pathway [36]. Loss of E‐cadherin in ILC is predominantly attributed to CDH1 mutation, which in 89% of cases occurred concurrently with heterozygous deletion in chromosome 16q [37]. Analysis of invasive breast carcinoma cases in the Cancer Genome Atlas (TCGA) dataset (http://www.cbioportal.org) [38, 39, 40] revealed that 66% (107/162) of ILCs harbor mutation in CDH1 compared with only 3% (22/741) of IDCs (Fig. 2A). Over 80 different mutations have been identified throughout the coding region of CDH1, 83% of which were predicted to result in truncation of the protein [37] resulting from shallow deletion (Fig. 2B). CDH1 mutations have also been detected in matching lobular carcinoma in situ, demonstrating that this is an early event [3, 37, 41]. Although some studies have reported downregulation of CDH1 transcription due to promoter methylation [3], based on TCGA data set and later studies by Ciriello et al., methylation‐mediated silencing of Cdh1 in ILC remains controversial [37, 40].

Figure 2.

Histological subtypes of ILC. Hematoxylin and eosin stained sections of classical (A), pleomorphic (B), qlveolar (C), histiocytoid (D), lobular invasive carcinoma‐classical (E), and lobular invasive carcinoma‐pleomorphic (F).

Loss of E‐cadherin markedly increases cytoplasmic accumulation of p120 catenin, which plays a key role in tumor invasion [42]. In normal breast and in situ carcinomas, p120 catenin is primarily bound to E‐cadherin at a juxtamembrane site, where it recruits microtubules and stabilizes E‐cadherin by suppressing endocytosis. p120 also regulates the activity of Rho family GTPases, the molecular switch that is key to cell migration [43]. Using a mouse model of ILC, Schackmann et al. showed that in cytosol, p120 catenin promotes tumor growth by inducing anoikis resistance [44]. Cytosolic p120 interacts with Rho/Rock antagonist myosin phosphatase Rho‐interacting protein and relieves suppression of Rho‐Rock signaling pathway and activated RhoA and its downstream effector Rock1, leading to anoikis resistance. Consistently, human ILC samples showed hallmarks of active Rock signaling [44]. Increase in nuclear p120 was noted by van de Ven et al. in mouse ILC cells and in metastatic cells present in pleural effusion of patients with ILC [45]. Nuclear p120 accumulation relieved transcriptional repression by Kaiso on its well‐known target Wnt11, a non‐canonical Wnt signaling activator. Wnt11 in turn activates RhoA‐Rock signaling and induces anoikis resistance, linking it to E‐cadherin loss. As E‐cadherin loss and cytoplasmic localization of p120 catenin is unique to ILC, this can be a useful diagnostic tool to differentiate between ILC and IDC [16, 40, 46].

ERBB2 Mutation in ILC

ILC is characterized as a Luminal A subtype because of high ER/PR expression, low HER2 amplification, and low‐grade tumor. However, recent studies revealed that HER2 mutations are more common in ILC (2%–15%), particularly in higher grade tumors, compared with other subtypes of breast cancer [9, 17, 40, 47, 48, 49]. Deniziaut et al. reported that 6 out of 39 grade 3 HER2‐negative ILC tumors harbor at least one HER2‐activating mutation, an approximately 15% incidence rate [50]. In addition, primary ILC tumors carrying an HER2 mutation were shown to have early relapse [47]. Based on multiple studies showing higher rates of HER2 mutations in IHC‐scored HER2‐negative ILC tumors, personalizing therapy by targeting this pathway is an important consideration. Targeting HER2 mutations using tyrosine kinase inhibitors along with trastuzumab is an option [51]. Clinical studies are needed to understand when and how to target the HER2 pathway in ILC to improve clinical outcomes.

Endocrine Resistance and Fox A1 Amplification in ILC

Patients with ILC are routinely treated with endocrine therapy, but one third of patients fail to respond to treatment and develop endocrine resistance in the long term. The underlying mechanism behind acquired endocrine resistance in ILC is multifactorial. Comparative gene expression analysis of tamoxifen‐resistant ILC cell line SUM44/LCCTam revealed reduced expression of ERα and increase expression of estrogen‐related receptor γ (ERRγ) when compared with parental cell lines [52]. Orphan nuclear receptor ERRγ was shown to activate AP1‐dependent transcription in the resistant cells. Activation of AP1‐mediated transcription conferring tamoxifen resistance has been demonstrated previously in HR‐positive MCF7 and BT474 IDC cell line [53, 54]. Stires et al. reported upregulation of mitogen‐activated protein kinase (MAPK)/extracellular signal‐regulated kinase (ERK) pathway and glutamate receptor in Sum44/LCCTam cells and their inhibition‐restored tamoxifen sensitivity [55]. Using long‐term estrogen‐deprived (LTED) cell line, Du et al. demonstrated activation of lipid metabolism in MDA‐MB‐134‐VI‐LTED cells, specifically, which increased expression of sterol regulatory binding element protein‐1 (SREBP1) and its downstream target fatty acid synthase [56]. Subsequent analysis of clinical specimens showed a significant association between high SREBP‐1 expression and lack of response to the aromatase inhibitors letrozole, suggesting a role of SREBP‐1 in acquired endocrine resistance. Additional work in LTED cell lines by Sikora et al. identifies WNT4, a ligand in the WNT signaling pathway, to be a driver for estrogen‐induced growth and a potential mediator of endocrine resitance in ILC [57]. Further research into the interplay betwen E‐cadherin loss and the WNT signaling pathway might uncover novel targets to overcome endocrine resistance in ILC.

FOXA1, an ER transcription modulator, plays a key role in cancer progression and development of endocrine resistance [58]. FOXA1 is considered a pioneer factor for its ability to associate with condensed chromatin and demethylate and demarcate the site for tissue‐specific binding of transcription factors, including ER, PR, and androgen receptor [58, 59]. FOXA1 is co‐expressed with ER at high levels in endocrine‐resistant metastatic breast cancer [60]. Through comprehensive genomic analysis, Ciriello et al. observed spatially clustered mutations of FOXA1 in 7% of ILC samples (n = 127), leading to increased FOXA1 expression and activity [37]. A higher rate of FOXA1 mutations was observed in ILC compared with IDC (7% vs. 2%). In contrast, the rate of GATA3 mutation, another key regulator of ER activity, was consistently lower in ILC versus IDC (5% vs. 20%) [37], suggesting that ILC and IDC may rely on different mechanisms to regulate ER‐mediated transcription. Stires et al. showed amplification of FOXA1 in tamoxifen‐resistant SUM44PE/LCCTam cell lines, which supports the notion that FOXA1 could be critical for endocrine resistance in ILC [55]. Further studies are warranted to fully understand the implication of FOXA1 mutations in endocrine‐resistant ILC.

PI3K/AKT Signaling in ILC

Alterations in phosphoinositide 3‐kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) signaling have been observed in numerous cancer types, including breast cancer [61, 62]. Frequent mutations in phosphatase and tensin homolog (PTEN) and PIK3CA, and low frequency of mutations in PIK3R1, AKT1, AKT2, and mTOR, have been reported in ILC [37, 63, 64]. Interestingly, Ciriello et al. noted mutually exclusive nature of PIK3CA mutation (48%) and PTEN mutation (13%) in ILC [37].

Activation of PI3K signaling was confirmed in both mouse and human ILC models by studying phosphorylation status of different intermediate proteins in this pathway. Whereas phosphorylation of AKT (at T473, T308), S6K1 (at T389), and 4EBP1 (at S65) was more prevalent in human ILCs (n = 66), mouse ILCs (n = 30) showed phosphorylation of AKT (at S473), S6 (at S235/S236), and 4EBP1 (at T37/46) [64]. Treatment of K14‐Cre;Cdh1 ^F/F ;Trp53 ^F/F mice, a model of metastatic ILC in either neoadjuvant or prolonged adjuvant setting with a mTOR inhibitor (AZD8055), showed response in primary tumors and delayed metastases. Interestingly, the therapeutic response was suggested to be in part mediated by the adaptive immune system [64]. Ongoing clinical trials with the mTOR inhibitor everolimus in early‐stage HR‐positive breast cancer may shed some light on the outcomes in patients with ILC (NCT01674140).

Fibroblast Growth Factor Signaling in ILC

The fibroblast growth factor (FGF) signaling pathway plays an essential role in cancer cell progression, angiogenesis, and survival. In approximately 7.1% of all cancers and approximately 18% of breast cancers, FGF receptor alterations have been reported [10]. Approximately 14% of patients with breast cancer have amplification of chromosome 8p11‐12, harboring FGFR1 [10, 11, 12]. Reis‐Filho et al. identified FGFR1 as the driver amplicon for 8p12‐p11.2 region in patients with ILC and demonstrated that inhibiting FGFR1 in MDA‐MB‐134 cells inhibited cell growth [14]. Overexpression of FGFR1 in breast cancer cell lines conferred tamoxifen resistance, whereas its silencing increased sensitivity to tamoxifen [15]. FGFR1 is amplified in human ILC cell lines MDA‐MB‐134 and SUM44PE, in which combined use of antiestrogen and FGFR inhibitor led to maximum cell death compared with monotherapy [24]. Using RNA‐sequencing and microarray analysis, Levine et al. identified FGFR4 overexpression in both ILC LTED and tamoxifen‐resistant cell lines. FGFR4 expression was also found to be higher in distant metastasis of patients treated with endocrine therapy [65]. In a mouse model of insertional mutagenesis, Kas et al. identified Fgfr2 as an oncogenic driver for ILC in conditional Cdh1 knockout mice, and the tumors were sensitive to FGFR inhibitor AZD4547 [25]. Eventual development of resistance to AZD4547 in these mice was attributed to activation of the canonical MAPK‐ERK pathway [26]. In other studies, FGFR mutation and amplification have been associated with disease recurrence [28, 49]. It is apparent that aberrations in the FGFR family proteins play a role in poor outcomes of patients with ILC, and further investigation into FGFR‐targeted therapy is warranted. Ongoing clinical trials with the FGFR2 inhibitor pemigatinib in advanced solid tumor malignancies harboring activating FGFR mutations and translocations may shed light on the outcomes in patients with ILC (NCT03822117).

Thus, ILC is not only a unique histological subtype with specific clinical features but also exhibits complex molecular alterations that could be potentially targeted for improved clinical outcomes (Fig. 3).

Figure 3.

Schematic representation of different molecular pathways involved in progression of invasive lobular breast cancer. (A): Mechanism by which a functional E‐cadherin maintains cell–cell adhesion. Juxtamembranous binding of p120, β‐catenin, and α‐catenin facilitates binding to actin cytoskeleton. The catenins are sequestered by E‐cadherin, facilitating the suppression of WNT11 via Kaiso. (B): Loss of E‐cadherin in ILC results in the translocation of p120 into the cytoplasm and nucleus. In the cytoplasm it sequesters MRIP and relieves Rho/Rock repression. In the nucleus, it binds Kaiso and initiates transcriptional expression of WNT11, which then activates Rho/Rock signaling, both leading to anoikis resistance. (C): Amplifications of FGFR in ILC results in increased receptor dimerization and transphosphorylation at several tyrosine residues, leading to the activation of Ras‐dependent MAPK signaling, PI3K/AKT signaling, and PLC‐γ signaling, contributing to endocrine resistance. (D): ILC can harbor activating HER2 mutations, initiating PI3K/AKT and Ras/MAPK signaling cascades, which contributes to cell survival, proliferation, and uncontrolled cell growth. (E): FOXA1 mutations in ILC result in increased expression, contributing to endocrine therapy resistance by facilitating the binding of ERα/E2 complexes to non‐ERE sites and expression of target genes.Abbreviations: AKT, protein kinase B; ERα/E2, estrogen receptor alpha/estradiol; ERE, estrogen response element; GPCR, G‐Protein coupled receptor; ILC, invasive lobular cancer; MAPK, mitogen‐activated protein kinase; MRIP, myosin phosphatase Rho‐interacting protein; PKC, Protein Kinase C; PI3K, phosphoinositide 3‐kinase; PLC, pleomorphic lobular cancer.

Dormant Disseminated Cancer Cells in Late Recurrence of ILC

The overall survival for patients with ILC is favorable compared with IDC within the first 5 years of diagnosis, but the outcome is worse after 5 years [66]. One reason for this is the presence of disseminated cancer cells (DCC) that remain dormant for prolonged periods and can result in delayed relapse at distant metastatic sites. Cancer cells disseminated from the primary site prior to surgery exist as minimal residual disease for a long period of time before switching to overt metastasis and aggressive growth [67]. ER‐negative breast cancers have higher rates of relapse within 5 years of primary tumor diagnosis, whereas ER‐positive breast cancers relapse at greater rates between 5 and 20 years [68]. In a comparative genomic study between ILC and IDC, Du et al. described lower rates of protein translation and metabolism in ILC compared with IDC, which are known features of tumor dormancy [69]. Additionally, Narbe et al. compared the number of circulation tumor cells (CTCs) in patients with ILC and IDC and showed that although patients with ILC had higher number of CTCs, the prognostic significance for CTCs ≥5 was weaker, suggesting a more dormant nature of these disseminated tumor cells [70]. Although in‐depth research on the specific role of the DCCs and the metastatic niche in ILC is lacking, they are believed to play a major role in the late recurrence of ILC, leading to worse patient outcomes. It is argued that targeting the dormant niche could be an effective way of delaying or preventing metastasis [71]. Understanding the dormant niche and specific molecular switch in ILC that awakens the dormant cells to proliferate into macrometastatic disease is crucial to developing novel therapeutics to treat ILC.

Genetically Engineered Mouse Models for ILC

Animal models of ILC are limited and a critical area of need. Conditional deletion of Cdh1 using either MMTV‐, K14‐, or WAP‐Cre failed to form mammary tumors [72, 73, 74]. However, dual knockout of Cdh1 and p53 in mouse mammary epithelium using either K14‐ or WAP‐Cre leads to the formation of mammary tumors resembling human pleomorphic ILC [72, 75]. The WAP‐Cre;Cdh1 ^F/F ;Trp53 ^F/F mouse is a better model, as K14‐Cre;Cdh1 ^F/F ;Trp53 ^F/F mice develop skin cancers as well. The WAP‐Cre;Cdh1 ^F/F ;Trp53 ^F/F mice develop invasive and metastatic ILC similar to human ILC, with metastatic spread to the bone, gastrointestinal tract, lymph nodes, and lung [75]. WAP‐Cre;Cdh1 ^F/F ;Pten ^F/F mice generated by Boelens et al. develop tumors that closely resemble the histological and molecular features, ER status, growth kinetics, metastatic behavior, and tumor microenvironment of classical human ILC [76]. To match the immune‐related (IR) subtype of ILC, An et al. developed the Wap‐Cre;CDH1 ^loxP/loxP;R26‐LSL‐Pik3ca ^H107R, a model that displays morhological features, immune cell infiltration, and T‐regulatory cell signaling observed in human IR‐ILC [77]. In mice with mammary gland–specific loss of E‐cadherin and expression of Cas9, the lentiviral delivery of PTEN targeted single‐guide RNA resulted in efficient induction of ILC [78]. The WAP‐Cre;Cdh1 ^F/F ;Cas9 mice along with guide RNA targeting specific tumor suppressors provides an opportunity to test combinations of tumor suppressor genes implicated in ILC. Although many ILC genetically engineered mouse (GEM) models have been developed over the past decade, an established model with a high rate of ER postivity that closely mirrors the unique histology and metastatic potential of human ILC is critical to study hormone therapy resistance and to validate novel drug targets.

Clinical Management of Invasive Lobular Cancer

Clinical management of invasive lobular breast cancer depends on the biology of the tumor and the stage at diagnosis.

Role of Surgery

A significant number of ILC patients undergo mastectomy. Breast‐conserving surgery (BCS) in combination with postoperative radiation is considered to be an alternative to mastectomy in women with early‐stage breast cancer. Foder et al. showed that BCS with radiation or mastectomy results in similar distant metastasis–free survival, breast cancer–specific survival, and locoregional recurrence–free survival [79]. Rates of BCS in early‐stage ILC have increased by 11.6% from 1993 to 2003 [80], with accumulating evidence suggesting no significant long‐term survival benefit for patients undergoing mastectomy [81, 82]. Approximately 17% to 65% of patients with ILC require a second intervention, and re‐excision lumpectomy is successful in achieving clear margins, particularly in those with node‐negative disease [83, 84]. The unique pathology of ILC leads to under‐recognition of the extent of the disease in the breast on routine preoperative imaging and results in a more extensive surgery in order to obtain clear margins. This is one of the reasons for the higher rates of mastectomies seen with this type of breast cancer.

Role of Neoadjuvant Therapy

Patients with clinical stage II or stage III disease are eligible to have neoadjuvant treatment, which is given to improve rates of BCS [85, 86]. However, ILC is considered to be less chemosensitive compared with IDC, with lower rates of pathologic complete response (pCR) and BCS compared to IDC. In a retrospective study by Tubiana‐Hulin et al., 860 patients received neoadjuvant chemotherapy, of whom 14% (118) had ILC [87]. About 30% of patients with ILC compared with 48% of patients with IDC underwent BCS; pCR was observed in 1% of patients with ILC versus 9% of patients with IDC (p = .002). The poor response to chemotherapy is thought to reflect the biology of ILC, being strongly ER/PR positive with a low mitotic index. Therefore, neoadjuvant chemotherapy should only be offered to patients with ILC after due consideration and a multidisciplinary discussion.

As ILC is often strongly ER/PR positive, treatment with endocrine therapy in the neoadjuvant setting could be considered. A retrospective study by Dixon et al. consisting of 61 patients with ER‐positive ILC found that neoadjuvant letrozole reduced tumor size by a mean of 66% at 3 months, with an 81% rate of successful breast conservation [88]. This further supports the concept that patients with ILC could receive neoadjuvant endocrine therapy instead of neoadjuvant chemotherapy if they desired breast conservation. The ALTernate approaches for clinical stage II or III Estrogen Receptor positive breast cancer NeoAdjuvant TrEatment (ALTERNATE trial, NCT01953588) [89] is an ongoing randomized phase III trial to determine whether neoadjuvant fulvestrant or the combination of anastrozole plus fulvestrant is better than anastrozole monotherapy for tumor shrinkage in postmenopausal patients with stage II to III invasive lobular and ductal breast cancer.

Role of Adjuvant Therapy

Systemic therapy given after definitive surgery impacts disease‐free survival (DFS) and OS in breast cancer [90]. The decision to proceed with adjuvant chemotherapy is often based on the size of the primary tumor, lymph node status, and results of genomic testing (i.e., Oncotype DX; Prosigna‐PAM50) if indicated. Oncotype DX recurrence scores for patients with ILC usually tend to be in the low or intermediate range, making the benefit of adjuvant chemotherapy small [91, 92, 93]. The application of the Oncotype DX prognostic tool in ILC needs further investigation, as very few patients with ILC were included in the trials used in the development of this tool. A retrospective analysis of patients within the Danish Breast Cancer Group database suggests that postmenopausal women with ILC and an intermediate Prosigna‐PAM50 score have 18% 10‐year risk of recurrence compared with 8% in those with low Prosigna‐PAM50 score [94]. Thus, women with intermediate/high genomic scores may have a potentially meaningful clinical benefit from chemotherapy compared with women with low scores. The effect of adjuvant chemotherapy in ILC was retrospectively studied by Truin et al. [95]. Patients were divided into two groups: one group receiving adjuvant hormonal therapy alone, and the other receiving adjuvant hormonal therapy plus adjuvant chemotherapy. The 10‐year OS rate for IDC when treated with hormonal therapy alone and combination therapy was 69% versus 74% (p < .0001); for ILC, 10‐year OS rate was 68% versus 66% (p = .45). This retrospective study suggests that adjuvant chemotherapy offered no additional benefit for patients with ILC receiving endocrine therapy, in contrast to patients with IDC.

In early‐stage HR‐positive breast cancer, 5 years of adjuvant endocrine therapy substantially reduces the risks of locoregional and distant recurrence and death from breast cancer [96, 97]. Tamoxifen is typically indicated for premenopausal women, whereas a third‐generation aromatase inhibitor is given to postmenopausal women. Investigators from the Breast International Group 1‐98 study explored the relative effectiveness of letrozole versus tamoxifen in patients with ILC compared with IDC. After a median follow‐up of 8.1 years, letrozole was associated with a significantly reduced risk of DFS event in patients with both luminal B‐like (HR = 0.34, 95% confidence interval [CI] = 0.21–0.55) and luminal A‐like ILC subtypes (HR = 0.50, 95% CI = 0.32–0.78), concluding that the magnitude of benefit from letrozole was greater in patients with ILC versus IDC [98]. CAN‐NCIC‐MA27 (NCT00066573) was a randomized phase III trial comparing anastrozole versus exemestane. There was no difference in event‐free survival between ILC and IDC; however, in the ILC group, anastrozole showed a marginal benefit (HR = 1.79, 95% CI = 0.98–3.27, p = .55), which may not have any clinical significance [99]. A current presurgery window‐of‐opportunity study (NCT02206984) is investigating which antiestrogen therapy (fulvestrant vs. anastrozole vs. tamoxifen) is most effective in ILC. Hence, current data support the important role of adjuvant/neoadjuvant endocrine therapy in HR‐positive ILC, with some data supporting the use of aromatase inhibitors over tamoxifen.

Conclusion

Given the rather uniform characteristics of ILC tumors with high HR positivity and low proliferation index, the challenge is how to identify tumors that would have a worse prognosis. One molecular signature, LobSig, that captures the unique molecular landscape of ILC along with clinicopathological features appears to provide independent prognostic information for patients with ILC, but it needs validation prospectively [100]. Identifying patients with high risk of recurrence will help in exploring novel targeted therapies and potentially immunotherapy. The ROLO study (NCT03620643), a phase II trial, is investigating the role of ROS1 inhibition with crizotinib in advanced E‐cadherin–negative, ER‐positive lobular breast cancer and diffuse gastric cancer. E‐cadherin defective cells rely on ROS1 for proper function, and ROS1 inhibition demonstrates a synthetic lethality in in vitro and in vivo studies [101].

CDK4/6 inhibitors are currently approved in combination with endocrine therapy as first‐ and second‐line treatment in advanced ER‐positive, HER2‐negative breast cancer. Multiple ongoing clinical trials are investigating the role of CDK4/6 inhibitors in neoadjuvant and adjuvant settings. PALLET (FB‐11; NCT02296801) demonstrated that the addition of palbociclib to letrozole in the neoadjuvant setting significantly improves the suppression of tumor proliferation in HR‐positive early‐stage breast cancer [102]. Palbociclib and Endocrine therapy for LObular breast cancer Preoperative Study (PELOPS, NCT02764541) is a phase II clinical trial assessing clinical utility of neoadjuvant palbociclib in women with HR‐positive, HER2‐invasive ILC. An immune‐related ILC subgroup was identified in the analysis of TCGA and RATHER consortia data sets [37, 103]. A phase II clinical trial is assessing the efficacy of carboplatin and atezolizumab in metastatic lobular breast cancer (the GELATO study, NCT03147040).

Invasive lobular carcinoma of the breast is a distinct entity with unique clinical and molecular features. Although studies show that the outcomes are largely similar to ductal cancers, the delay and difficulty in early diagnosis, inherent resistance to conventional therapies, and risk of late recurrences create significant challenges in management. With so few human ILC cell lines and established GEM models, there is a unique opportunity to develop novel models of ILC to advance ILC research. ILC‐specific diagnostic tools and treatment options are needed to improve overall patient outcomes. The lack of large national studies addressing the needs of the patients diagnosed with this histology is an unmet need.

Author Contributions

Conception/design: Bhuvaneswari Ramaswamy

Provision of study material or patients: Gary Tozbikian

Collection and/or assembly of data: Nikhil Pramod, Akanksha Nigam, Mustafa Basree, Resham Mawalkar, Saba Mehra, Gary Tozbikian, Nicole Williams, Sarmila Majumder, Bhuvaneswari Ramaswamy

Data analysis and interpretation: Nikhil Pramod, Akanksha Nigam, Sarmila Majumder, Bhuvaneswari Ramaswamy

Manuscript writing: Nikhil Pramod, Akanksha Nigam, Mustafa Basree, Resham Mawalkar, Saba Mehra, Neelam Shinde, Gary Tozbikian, Nicole Williams, Sarmila Majumder, Bhuvaneswari Ramaswamy

Final approval of manuscript: Nikhil Pramod, Akanksha Nigam, Mustafa Basree, Resham Mawalkar, Saba Mehra, Neelam Shinde, Gary Tozbikian, Nicole Williams, Sarmila Majumder, Bhuvaneswari Ramaswamy

Disclosures

Bhuvaneswari Ramaswamy: Eisai (C/A); Gary Tozbikian: Roche/Genentech (H), Lilly USA (C/A). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board