Introduction

Cancers arise from a diverse accumulation of mutations affecting cellular pathways that control growth, cell death, differentiations, and interactions with the environment. Many of these functions are carried out or modulated by key tumor suppressor genes, mismatch repair genes, and oncogenes. Mutations may occur in tissues, termed somatic, and/or in germline DNA. Germline mutations may be passed on to subsequent generations leading to cancer family syndromes, namely families in which carriers of deleterious mutations exhibit increased susceptibility to cancer development. Approximately 10 % of all cancers, including breast cancer (invasive and ductal carcinoma in situ, DCIS), are attributable to inherited cancer-susceptibility gene mutations [1, 2]. Inherited susceptibility to cancer is associated with early-onset disease (varies by cancer site), multiple effected generations, and rare tumor types and/or multiple primary malignancies. Well-known examples of hereditary cancer syndromes include Lynch (hereditary nonpolyposis colorectal cancer syndrome or HNPCC), PTEN hamartomatous tumor syndrome (PHTS or Cowden), Li-Fraumeni, and hereditary breast and ovarian cancer (HBOC) syndromes due to mutations in mismatch repair genes, PTEN, p53, and BRCA1/2, respectively [3] . Identifying individuals with deleterious germline mutations in these and other cancer susceptibility genes is primarily based on family and personal cancer history and is conducive to the prevention and early detection of cancer in high-risk populations. In addition, specific ethnic backgrounds (Ashkenazi Jewish (AJ), Dutch, and Finnish) are known to be at increased risk to harbor mutations in genes associated with inherited susceptibility to cancer [4, 5].

Approximately 5–10 % of breast cancers are attributable to inherited cancer-susceptibility genes. Although there is significant overlap, familial breast cancer may differ from hereditary breast cancer in that familial breast cancer does not as often display early-age onset or the same inheritance patterns as hereditary cancers due to single germline variants [3]. It is important in all hereditary cancers to determine and validate an algorithm for genetic testing. For example, if HBOC is suspected in an individual, there are various predictive models that aid in determining likelihood that an individual is a BRCA1/2 carrier including BRCAPRO [6], UPENN [7], and BOADICEA [8]. However, only a fraction of individuals with familial breast cancer harbor germline mutations. Importantly, there are populations that are enriched for BRCA1/2 germline variants, including persons of AJ descent, among which approximately 1 in 40 individuals harbors a BRCA mutation. The estimated prevalence of BRCA1 and BRCA2 mutations across all women diagnosed with invasive breast cancer (IBC) is 0.4–2.6 % and 1.4–1.5 %, respectively, with percentages increasing in women diagnosed at earlier ages (< 45 years old) [9].

DCIS Within the Spectrum of Breast Diseases Predisposing to IBC and Family History

DCIS is the most common form of non-IBC. There is no age group in which DCIS is more common than IBC. In studies comparing in situ breast disease versus controls and IBC, it has been shown that a family history of IBC leads to a two- to threefold increase in risk of in situ disease when compared to controls. The increased risk conferred by a family history of breast cancer is similar to in situ disease and invasive disease [10]. A large population study comparing women with DCIS (n = 875) and LCIS (n = 123) to age-matched controls (n = 999) found that women with in-situ disease were significantly more likely than their age-matched controls to report a family history of breast cancer in a first-degree relative. In addition, women with a first-degree relative with breast cancer onset prior to the age of 49 were at even higher risk than those with onset after age 49 (2.1-fold increase vs. 1.5) [11]. However, family history of breast cancer in a first- or second-degree relative did not confer an increased risk of DCIS in a study focused on Han Chinese population but did confer increased risk of IBC. The increased risk of DCIS due to family history may be population specific; however, this study included only 123 patients with DCIS and used benign breast cases as a control for comparison [12].

Early studies have debated whether DCIS should in fact be included as part of HBOC syndrome, as BRCA1/2 mutation carriers showed a paucity of DCIS compared to sporadic controls. For example, a study published in 1997 comparing 447 familial breast cancer cases (including 196 BRCA mutation carriers) to 547 age-matched controls found significantly less DCIS surrounding the invasive tumor in BRCA1 carriers when compared to sporadic controls, 41 and 56 %, respectively; however, BRCA2 showed a similar amount of invasive tumor-associated DCIS as sporadic cases [13]. This and other similar early studies led to the hypothesis that BRCA1-associated IBCs may in fact have a shortened preinvasive phase or skip this phase altogether [14, 15]. One subsequent study investigated the prevalence of premalignant breast lesions in 67 women who underwent prophylactic bilateral mastectomy due to high hereditary risk (66 % of which were known BRCA mutation carriers) and found that more than half of these women had premalignant lesions in at least one resected breast, with 15 % being DCIS. Women over the age of 40 from this hereditary group were at statistically significantly increased risk for developing a premalignant lesion compared to younger women [16].

It was pivotal to define certain groups of women with DCIS who would be appropriate for BRCA testing, as it greatly impacts surveillance and prophylaxis. A population-based case-control study evaluating BRCA mutation prevalence among 369 DCIS cases unselected for age, family history, or ethnicity found that 0.8 and 2.4 % of women carried disease-associated BRCA1 and BRCA2 mutations , respectively. These carriers were significantly more likely to have a first-degree relative with breast cancer [9]. Many tools for carrier-status prediction do not incorporate DCIS. A recent study by Mazzola et al. assessed the absolute risk for DCIS among BRCA mutation carriers and found a sixfold increased lifetime risk of DCIS in deleterious mutation carriers when compared to noncarriers [17]. Another study of 118 women with DCIS who were referred for genetic counseling and underwent BRCA1/2 testing found that 27 % (32/118) of these women had a BRCA mutation, with 10 % BRCA1 mutation carriers and 17 % BRCA2 carriers . In addition, they found that in these high-risk women with DCIS, a family history of ovarian cancer and/or a BRCAPRO score ≥ 10 % conferred higher rates of BRCA mutations [18]. These studies and others led to the current recommendation that DCIS should be included as part of the risk assessment for HBOC syndrome, and that patients with a concerning family history should be screened for mutations regardless of diagnosis of IBC versus DCIS. However, in current BRCAPRO model, DCIS is still not weighted as heavily as invasive disease in risk assessment.

It has been shown that the prevalence of DCIS is roughly equivalent among women who carry deleterious BRCA mutations (37 %) versus women who are BRCA-mutation negative, but who have a high familial risk of breast cancer (34 %). In this study, the women with BRCA mutations overall had an earlier age of onset of both DCIS and IBC. Interestingly, there were differences seen even among BRCA1 versus BRCA2 carriers with BRCA1 carriers being statistically more likely to have high-grade DCIS when compared to controls. However, this study comprised a relatively small number of DCIS-only cases in mutation carriers [19].

A follow-up study assessed three groups, a prevalent series of women with DCIS of AJ descent (retrospective), an incident series of women with DCIS of AJ descent (prospective, pre-op), and a clinic-based series of women with DCIS referred for hereditary cancer risk assessment . These cases of pure DCIS were compared to IBC-matched controls. They found that among the women of AJ descent, the control cases with IBC were significantly more likely to harbor BRCA mutations when compared to DCIS cases; however, similar mutation frequency was found among those women who presented for hereditary cancer risk assessment based on family history and early age of onset (12.7 % carried mutations in DCIS versus 14 % in IBC). The risk of harboring a mutation among the DCIS group was higher if the woman had a family history of ovarian cancer or early-onset breast cancer [14]. This highlights the importance of screening for DCIS in mutation carriers, as this might better identify cancers while still in the in situ phase. Additionally, the fact that a lower portion of mutations carriers had DCIS when compared to IBC might suggest that some BRCA mutation carriers have a shortened preinvasive period when compared to noncarriers (Table 15.1).

Table 15.1 DCIS statistics in BRCA mutation carriers
Full size table

Characteristics of DCIS in the Context of Cancer Family Syndromes

Similar to its invasive counterpart, DCIS is a heterogeneous disease across individuals, with distinct molecular pathology and phenotypic outcomes based on mutation carrier status . For example, DCIS associated with invasive disease in BRCA1 carriers tends to show a more basal-like phenotype with low expression of estrogen receptor (ER) , progesterone receptor (PR), and Her2/neu (Her2) while positive for cytokeratin and epidermal growth factor receptor (EGFR). BRCA2 carriers however tend to display a more luminal-type pattern, frequently staining positive for ER/PR and negative for cytokeratin and EGFR. Not surprisingly, DCIS in BRCA1 carriers is more likely to be highly proliferative. In IBC, these molecular subtypes greatly impact the phenotypic outcomes in patients, and this study suggests that these crucial drivers of cancer are already determined in the preinvasive stage [20].

There also exist some similarities between DCIS in BRCA1 and BRCA2 carriers with regard to expression of certain hypoxia markers. Hypoxia-inducible factor-1α (HIF-1α) expression was detected in 63 % of BRCA1 (n = 32) and 62 % of BRCA2 (n = 16) as compared to 34 % of non-BRCA mutation-related (n = 77) DCIS cases (p = 0.005). Similar overexpression of CAIX and Glut-1 was seen in the BRCA-related cases. The expression of these hypoxia-related proteins in BRCA mutation-associated DCIS was similar to the expression in the matched invasive cancer in 60 % or more of cases [21]. These unique biochemical and molecular characteristics of DCIS in BRCA carriers have important implications for biomarkers of early detection and targeted treatment (Table 15.2).

Table 15.2 Immunohistochemical profile of DCIS in BRCA1/2 mutation carriers (n = 28)
Full size table

DCIS lesions in BRCA1 mutation carriers were:

  • Mostly basal type

  • Low ER/PR/Her2

  • Frequently expressing CK5/6, CK14, and EGFR

  • Grade 3, with high proliferation

DCIS lesions in BRCA2 mutation carriers were:

  • Mostly luminal type

  • Frequent expressions of ER and PR

  • Infrequent CK5/6, CK14, and EGFR expression

  • Grade 3, with low proliferation

There exists a variety of other germline mutations associated with hereditary breast cancer syndrome including ATM, PALB2, NBN, BRIP1, MRE11, RAD51C, RAD50, STK11, CDH1, TP53, CHEK2, and PTEN. While their associations with IBC are relatively well defined, there has not been any large study to date focusing strictly on DCIS in persons harboring variants in these genes (Fig. 15.1). A smaller study of 43 malignant breast specimens from 39 women with known TP53 mutations included 32 invasive ductal carcinomas and 11 DCIS cases. The women in this study had an earlier average age of onset of DCIS (34 years), and their DCIS was noted to be of high nuclear grade, with 55 % being ER/PR + and 73 % Her2(+) [22]. Further studies of DCIS in individuals who carry these rarer germline mutations in non-BRCA genes should be pursued as this hereditary population is likely to benefit most from earlier detection and targeted treatment of their malignant lesions .

Fig. 15.1
figure 1

Inherited germline mutations in breast cancer

Full size image

It is important to note that since it has been only a few months that we have been more commonly testing for germline mutations in these non-BRCA genes, we cannot, at this point in time, speculate as to what percentage of the total burden of DCIS present in cancer families is attributable to any one of these genes. We need to entertain the possibility that, when more individuals are tested, we may find that these mutations are in aggregate responsible for a very significant number of cases of DCIS, perhaps in numbers comparable to the BRCA genes, but possibly at lower risk in each individual person.

Early Detection by Circulating Epithelial Cells and Future Research

Recent studies have highlighted the importance of circulating tumor cells (CTC) as an independent prognostic indicator of progression-free survival and breast cancer-related death. CTCs are found in ~ 70 % of metastatic breast cancer cases, and the presence of CTCs prior to neoadjuvant chemotherapy is associated with a significant risk for breast cancer recurrence. A study of 602 patients undergoing breast surgery showed that having CTCs in peripheral blood prior to curative surgery was associated with an increased risk of breast cancer-related death. Interestingly, they found similar percentage of women with DCIS had CTCs as compared to invasive disease (19 %), though those women with DCIS were excluded from further follow-up analysis or risk outcomes [23]. This study raises an interesting thought that CTCs may set the milieu for metastasis even as early as the preinvasive stage.

A vital clinical question that remains largely unanswered is how to better determine which cases of DCIS will recur and which will go on to develop invasive disease. Molecular subtyping analysis has identified differences at the protein level of DCIS specimens from individuals who subsequently go on to develop invasive cancer versus those who have recurrence of DCIS alone. Interestingly, in this study, a family history of breast cancer was not associated with disease recurrence or progression to invasive disease [24]. Clinical prediction tools have been proposed to better predict DCIS recurrence. For example, Memorial Sloan Kettering Cancer Center’s Breast Cancer Normogram includes a clinical calculator to determine risk of DCIS recurrence. This calculator is based on ten clinicopathologic variables based on age, family history , surgery, radiation therapy, endocrine therapy , and pathology; however, family history was not shown to be a statistically significant predictor of recurrence [25].

Previous studies have shown epigenetic and protein changes in mammary cells obtained by random periareolar fine needle aspiration (RPFNA) from asymptomatic women at high risk for developing breast cancer even prior to the development of DCIS. A study of promoter methylation in key tumor suppressor genes that included 40 unaffected premenopausal women who underwent BRCA1/2 testing based on strong family history of breast cancer showed that women with BRCA1/2 mutations had a low frequency of CpG island promoter methylation events in key tumor suppressor genes (RARB, ESR1, INK4a/ARF, BRCA1, PRA, PRB, RASSF1A, HIN-1, and CRBP-1) whereas women without a mutation but still at high risk based on family history showed a high frequency of promoter methylation events in this same gene panel [26]. In a small pilot study of 26 similarly high-risk asymptomatic women (27 % of which had a strong family history of breast cancer), the majority of RPFNA samples with atypia (based on Masood cytology) showed high expression of key cell survival proteins when compared to non-atypical cells [27]. These studies and others shed light on the molecular underpinnings of cancer initiation and progression from premalignant to preinvasive in situ stages.

Risk Management and Therapy for DCIS in Cancer Families

The decision to undergo prophylactic mastectomy in women with a hereditary predisposition to breast cancer is fraught with emotional, social, moral, and ethical issues. The ability to give a woman a personalized breast cancer risk assessment and to better determine timing of risk-reducing surgeries is crucial. Traditional mammography has been the mainstay of screening for breast cancer in recent history; however, false-negative rates of mammography in hereditary breast cancer populations are not inconsequential. Interval cancer rates in hereditary breast cancer populations are as high as 55 % with screening mammography alone [28]. Magnetic resonance imaging (MRI) has a high sensitivity for detection of breast cancer and does not have the cumulative radiation side effect of mammography. In early studies, MRI showed a higher false-negative rate with limited sensitivity for DCIS. However, MRI imaging interpretation for DCIS has improved, and thus MRI is now recommended yearly for BRCA, TP53, and PTEN mutation carriers and for non-BRCA carriers with ≥ 20 % lifetime risk by a breast cancer risk assessment model that incorporates family history . The EVA trial compared and contrasted various breast cancer detection modalities (mammography, ultrasound , clinical breast exam, and MRI) via a prospective, multicenter observational study of 687 women at high risk for developing breast cancer over a 3-year period. Twenty-seven women were diagnosed with breast cancer during this time (including 11 cases of DCIS) with 14 of these cancers diagnosed only by MRI (52 %). Thus, cancer yield achieved by MRI alone was significantly higher with MRI’s sensitivity of 93 % and positive predictive value at 48 % compared to sensitivity of 33 % and positive predictive value of 39 % for mammography alone [28]. This study supports that MRI screening in hereditary/familial breast cancer populations can actually shift detection towards DCIS and away from invasive disease.

Chemoprevention using selective estrogen receptor modulators (SERMs) or aromatase inhibitors (AIs) has shown significant risk reductions in IBC and pre-IBC in high-risk populations. However, chemoprevention is not a widely used practice, and chemoprevention with tamoxifen in BRCA1 and BRCA2 mutation carriers has been somewhat controversial. King et al. initially reported on tamoxifen’s role as a chemoprevention agent using data from the National Surgical Adjuvant Breast and Bowel Project (NSABP-P1) Breast Cancer Prevention Trial comparing tamoxifen versus placebo with end point of breast cancer risk reduction. They compared the BRCA1/2 status of women who developed breast cancer during their participation in the study, assuming that the equal randomization to tamoxifen versus placebo also held for BRCA1/2 gene mutation status. DNA samples from 288 of the 315 women who were diagnosed with IBC during the prevention trial were included in this study. Eight of the women had BRCA1 mutations (five of whom were in tamoxifen arm, three in placebo), and 11 had mutations in BRCA2 (three of whom were in tamoxifen arm, eight in placebo). Breast cancer risk reduction ratios were calculated to be 1.67 (0.41–8.00) for BRCA1 carriers and 0.38 (0.06–1.56) for BRCA2 carriers . Thus, it was concluded based on this very preliminary and small dataset that tamoxifen did not have a significant impact on reducing breast cancer risk in BRCA mutation carriers [29]. However, the sample size was too small to draw any major conclusions affecting this important issue.

A subsequent case-control study evaluated tamoxifen’s role as a chemoprevention agent in 1036 women with IBC, including 285 with bilateral breast cancer and 751 with unilateral disease, who were also known as BRCA mutation carriers. They found the multivariate odds ratio for BRCA1 mutation carriers was 0.50 (0.30–0.85) and for BRCA2 mutation carriers was 0.42 (0.17–1.02). This translated into an approximate 50 % risk reduction for BRCA1 mutation carriers and an approximate 58 % risk reduction for BRCA2 mutation carriers [30]. A recent follow-up study found that short-term use of tamoxifen for chemoprevention in BRCA1 and BRCA2 mutation carriers is likely as effective as a conventional 5-year course treatment [31]. This stresses again the importance of identifying those women at risk for hereditary or familial breast cancer so that prevention strategies can be tailored accordingly.

Hereditary DCIS , like its invasive counterpart, can behave quite differently depending on the individual patient. Both genetic and environmental factors alter the mammary milieu in these at-risk women, and the ability to detect these early changes at the preinvasive or even premalignant phase is crucial. Once not even included as part of a hereditary breast cancer risk assessment, DCIS has been shown to be more prevalent and have a higher risk of progression in these predisposed populations. It is not clear that all invasive cancers stem from a DCIS precursor in hereditary populations, and breast cancer may even skip a preinvasive stage, altogether in some individuals. Further studies that better define the molecular changes of DCIS in HBOC populations are needed. Not all DCIS behaves similarly, and high-risk women would be the group to most benefit from earlier prevention, detection, and targeted treatments.