Invasive Carcinoma with Medullary Features
The WHO Classification of Tumors of the Breast recommends the use of the collective term “Carcinoma with Medullary Features” to encompass tumors previously diagnosed as classical medullary carcinoma (classical MC), atypical medullary carcinoma (atypical MC), and a subgroup of invasive breast carcinoma of no special type (NST) with medullary characteristics (Lakhani et al. 2012).
An expansile or pushing border
A syncytial growth pattern
High cytonuclear grade
A prominent lymphoplasmacytic infiltrate
These diagnostic criteria are less strict than the original criteria proposed by Ridolfi et al. for a diagnosis of classical MC which requires complete circumscription, a syncytial growth pattern in at least 75% of the tumor, a moderate to marked mononuclear stromal infiltrate, high nuclear grade, no intraduct component, and no microglandular features (Ridolfi et al. 1977).
Atypical MC displays some but not all of the features of classical MC and permits lack of complete circumscription, a less prominent lymphoid infiltrate, accompanying intraduct carcinoma and focal glandular differentiation. A well-circumscribed tumor with classical MC morphology in 75% of the tumor and the remainder composed of invasive carcinoma NST also qualifies as atypical MC.
A subset of invasive carcinoma NST that displays medullary characteristics, including circumscription, high nuclear grade, and a prominent lymphoplasmacytic stromal infiltrate, has been observed, particularly in women with inherited BRCA1 gene mutations.
In view of the need to recognize the latter tumor subgroup and the considerable lack of reproducibility in the diagnosis of classical and atypical MC, the WHO group has taken the pragmatic approach to combine these three groups emphasizing their common characteristics and propose the collective term, “Carcinoma with Medullary Features,” abbreviated to CMF for the remainder of this chapter (Lakhani et al. 2012).
Classical MC, as defined by Ridolfi et al. in 1977, is generally considered to be a rare breast tumor accounting for <1% of all breast cancers (Ridolfi et al. 1977). Rates of up to 7% have been reported and this is likely attributable to interobserver and inter-institutional variation in the application of the histological criteria used to make the diagnosis.
In reported series, MC is less common than atypical MC but the incidence is largely unknown due to variation in definition and lack of and/or inconsistent coding documentation.
CMF has a broader, more inclusive, definition and includes a subset of tumors with medullary characteristics that would previously have been categorized as invasive carcinoma NST. It is difficult to obtain precise figures on overall incidence, but it is estimated that CMF accounts for between 10% and 20% of invasive breast tumors (Hicks and Lester 2017).
The age at diagnosis of women with CMF ranges from 21 to 95 years, with an average age at presentation of between 45 and 50 years. The latter is approximately 10 years younger than the average age profile of women diagnosed with other types of invasive breast carcinoma. One in four diagnoses of CMF occurs in women under the age of 35 years. This is likely to be due to the association of CMF with inherited Breast Cancer susceptibility gene 1 (BRCA1) germline mutation (Eisinger et al. 1998).
CMF is predominantly reported in females in line with all other forms of breast carcinoma.
The literature is scarce on histological tumor type in men. Most breast tumors in men are luminal A or triple negative. MC has been reported in male patients. Male patients with BRCA1 gene mutations are also at increased risk of developing breast cancer. BRCA2 mutations confer a higher risk of breast cancer in men than in women.
Site and Presentation
Patients may present with a discrete palpable breast mass or may be diagnosed through mammographic screening, population based or high-risk surveillance programs. CMF is typically well circumscribed and may have a soft consistency due to the lack of a stromal desmoplastic reaction. In view of these features and the relatively young age at presentation, these tumors may mimic benign lesions both clinically and on imaging studies. The overlying skin may be red. Bilateral synchronous MCs have been reported, more commonly in patients with a positive family history.
CMF may be associated with a rapid growth rate and may also come to medical attention as an interval screening breast cancer.
Axillary lymph node involvement may occur but is relatively less common in CMF compared with invasive carcinoma NST. Concomitant enlarged axillary lymph nodes are more likely to be reactive than involved by metastatic CMF and presentation as axillary lymph node involvement is rare (Hicks and Lester 2017). Presentation as distant metastases is also rare.
Surgery: The primary goal of treatment is complete surgical excision with clear margins. Conservation surgery is clearly desirable and is feasible in most cases due to the absence of ductal carcinoma in situ. Mastectomy may be necessary to treat large and some centrally located tumors.
A diagnosis of CMF, in some countries, will lead to genetic testing in women who have not been previously tested. The results of this may lead to consideration of additional breast and other prophylactic surgery. This requires careful evaluation by the patient in consultation with her/his clinical team.
Chemotherapy: CMF is typically triple negative although cases of hormone receptor positive MC have been reported. In view of the more favorable prognosis associated with CMF, it has been suggested that patients with lymph node negative disease may possibly avoid chemotherapy. In general, this tumor type tends to occur in younger women and the constellation of morphological features is usually associated with triple negative status. At the present time it is recommended that adjuvant or neoadjuvant chemotherapy treatment recommendations for patients with CMF should take account of the clinical, pathological, and biological characteristics as for all invasive breast cancers. PARP inhibitors and platinum agents are reported to be more effective in the treatment of BRCA1-related tumors.
Immunotherapy: Several clinical trials are in progress to investigate the efficacy of immunotherapy agents in the treatment of triple negative breast carcinoma (TNBC) (Dua and Tan 2017). This relates to the discovery that approximately 25% of TNBCs express the T-cell inhibitory molecule PD-L1 (programmed death ligand-1), particularly in tumors with a high lymphocyte count. PDL1 is expressed on cancer cells, tumor infiltrating lymphocytes (TILs), and immune cells (Tung et al. 2016). The binding of PD-L1 to the PD-1 (programmed death 1) checkpoint receptor on T cells is one mechanism for tumor evasion of the immune response. As detailed in subsequent sections, CMF is characterized by a prominent inflammatory cell infiltrate with a high TIL count. CMF is also associated with a high level of genomic instability leading to increased production of neoantigens and greater immunogenicity (Weigelt et al. 2010). These features suggest that CMF is likely to be a candidate tumor type for targeted immunotherapy.
Adjuvant Radiotherapy: Recommendations for post-surgical radiotherapy should be made in accordance with those for other primary invasive breast carcinomas.
In the original report from Ridolfi et al., classical MC, diagnosed using their strict criteria, was associated with a significantly higher survival rate at 10 years, estimated at 84% for patients with classical MC compared with 63% for patients with non-medullary invasive breast carcinoma (Ridolfi et al. 1977). This also pertained to patients with positive axillary lymph nodes when the two groups were compared. In addition to 57 patients with classical MC, the study population included 79 patients with AMC and 56 with non-medullary carcinoma (invasive carcinoma NST, using modern terminology). Patients with AMC also had an improved prognosis with a 10-year survival rate of 74% overall. Interestingly, AMC with a sparse lymphoid infiltrate was associated with a relatively poor prognosis. Apart from this observation, the authors were unable to draw any firm conclusions regarding the prognostic effect of the other morphological features of AMC.
In a more recent study, Mateo et al. reviewed the outcome of patients diagnosed with MC (n = 3,688) and AMC (n = 288) registered with the National US Cancer Database (NCDB) during the period 2004 to 2013 (Mateo et al. 2016). Cases are reported to the NCDB using codes from the International classification of diseases for Oncology (ICD-O-3). During the study period 918,870 cases of non-medullary breast carcinoma were registered. Over 20% of tumors diagnosed as MC or AMC were recorded as hormone receptor positive. The study did not include a review of histopathological slides and the number of AMC tumors was relatively small. Notwithstanding these limitations the authors found no significant difference between MC and AMC in terms of clinicopathological characteristics and clinical outcome.
Subsequent studies from several institutions have also reported a more favorable outcome in patients with MC compared with grade matched invasive carcinoma NST. In a compilation of data from 13 International Breast Cancer Study Group (IBCSG) trials, Huober et al. found that patients with MC diagnosed on morphology alone had an improved prognosis in terms of disease free and overall survival at 14 years (Huober et al. 2012). Restricting the diagnosis to patients with hormone receptor negative disease was associated with an even more favorable outlook.
In the last decade, the role of TILs in mediating response to chemotherapy and improving clinical outcomes in all subtypes of breast cancer has been recognized. TNBCs are associated with a greater degree of lymphoid infiltration than luminal tumors and survival benefit increases with each 10% increase in TILs (Stanton and Disis 2016). The survival advantage associated with TILs appears to be more powerful in TNBCs than in HER2 positive tumors that may also feature lymphoid infiltration. Considering that a prominent lymphoid infiltrate is one of the key morphological criteria for a diagnosis of CMF, it appears likely that the improved prognosis in CMF is related to the host immune response, particularly in triple negative tumors. Gene expression profiling studies (GEP) have also demonstrated that the expression levels of immune response genes are independent predictors of outcome in highly proliferative breast cancers. A high TIL count also has implications for targeted immunotherapy as discussed above.
In the context of prognosis and treatment, the decision by the WHO to broaden the definition and to create the designation of CMF (to include MC, AMC, and invasive carcinoma NST with specific histological appearances) takes account of the compelling data regarding the improved outcome associated with specific MC morphology and that reported in relation to a high TIL count in TNBC.
CMF, by definition, is a well-circumscribed tumor with an expansile or pushing border and may resemble a fibroadenoma on gross inspection. The cut surface is fleshy and the consistency is soft, attributed to the absence of a stromal desmoplastic reaction. The color varies from tan to grey-white.
Foci of hemorrhage and necrosis may be evident but are not extensive.
Ridolfi et al. observed an average tumor size of 2.9 cm (Ridolfi et al. 1977). Some studies report an average tumor size of up to 4 cm but in most studies the recorded median size is in the range of 2– 3 cm.
The imaging characteristics of CMF, including those observed on mammography and ultrasound examination, may resemble a fibroadenoma. Magnetic resonance imaging (MRI) does not differentiate between CMF and other breast tumor types.
Calcification is rare in CMF due to the lack of stromal desmoplasia and to the absence or low volume of accompanying DCIS.
The original criteria for a diagnosis of classical MC included complete tumor circumscription with a pushing border and no extension of tumor cells into adjacent stroma, a syncytial growth pattern in at least 75% of the tumor, a moderate to marked mononuclear stromal infiltrate, high nuclear grade, no glandular features and no intraduct component (DCIS). The last criterion was shown to be less important with a DCIS component having no impact on prognosis.
A tumor showing the majority of these features can be diagnosed as CMF. This approach should facilitate recognition and identification of CMF. However, due to the inherent flexibility in the application of the diagnostic criteria (“some but not all of the features of MC”), interobserver variability and reproducibility of diagnosis continue to be a challenge.
Despite these limitations, tumors categorized as CMF share the following morphological features:
Syncytial growth pattern: This refers to the arrangement of tumor cells in solid, broad anastomosing sheets with minimal intervening stroma. In classical MC, this pattern is present in at least 75% of the tumor. Although the latter is not a requirement for diagnosis, CMF shows a predominant solid growth pattern with scant intervening stroma. The relative paucity of stroma and lack of a desmoplastic reaction confer a soft consistency on these tumors which, together with the typical pushing border, may give a false impression of a benign lesion on gross examination.
The presence of a prominent lymphoplasmacytic infiltrate is associated with an improved prognosis related to the host lymphocyte response.
Gland formation: The complete absence of gland formation is a strict requirement for a diagnosis of classical MC. Although CMF may show some gland formation, this should be minimal.
DCIS: Accompanying DCIS is usually absent or low volume in CMF.
Lymphovascular invasion: This is uncommon in CMF.
Other features: Hemorrhage and necrosis may be present within the tumor. Breast tissue in the vicinity of CMF may show lymphocytic lobulitis.
Histological grading: Opinion is divided regarding the role of histological tumor grading of CMF. Some authorities advise against grading as the morphological characteristics of CMF will conform to a grade 3 carcinoma and not reflect the improved prognosis of CMF relative to a grade matched invasive carcinoma NST. RCPath UK guidelines recommend grading all special type carcinomas in addition to invasive carcinoma NST (Ellis et al. 2016). The reasons cited include the diagnostic difficulty that may be encountered in separating a special type carcinoma from invasive carcinoma NST where knowledge of histological grade may inform further treatment planning.
At the present time, CMF is defined based on the constellation of morphological characteristics outlined above. While CMF is associated with a classical immunohistochemical profile, these are not the defining properties of this tumor type and some variation in the expression of the following proteins will occur.
CMF is one of the morphological subtypes of invasive breast carcinoma with an immunophenotype that correlates with the specific so-called basal-like tumor gene expression profile (Flucke et al. 2010; Jacquemier et al. 2005; Rakha and Green 2017). There is no internationally accepted definition of a basal-like tumor. Immunohistochemical marker expression profiles that have been proposed to define basal-like breast cancers include (a) triple negative for estrogen (ER) and progesterone (PR) receptors and HER2 expression, (b) expression of one or more high-molecular-weight/basal cytokeratins (CK5/6, CK14, and CK17) and (c) triple negative (ER, PR, HER2 negative) immunophenotype combined with expression of CK5/6 and/or epidermal growth factor receptor (EGFR) (Badve et al. 2010).
CMF is negative for ER, PR, and HER2 in the majority of cases. However, expression of these biomarkers has been reported in small numbers of CMF indicating that tumors with this light microscopic morphology are heterogeneous at cellular level (Matkovic et al. 2008; Mateo et al. 2016). Notwithstanding this if any of these biomarkers is positive one should at least consider alternative tumor types in the differential diagnosis.
In addition to triple negative status, CMF has been shown to express basal type cytokeratins, in particular CK5/6, with greater frequency than non CMF breast tumors. Reported rates of CK5/6 expression vary, attributed to variation in antibodies used and the inclusion of hormone receptor positive tumors with classical CMF/MC morphology in some studies. CK 14 expression has also been reported but appears to be less commonly expressed than CK5/6. Up to 70% of CMF tumors have been shown to express EGFR.
Expression of other immunomarkers associated with basal-like tumors has been documented in CMF including P-cadherin, p63, smooth muscle actin, S100, and caveolin.
In keeping with the morphology, CMF is a highly proliferative tumor with a high Ki67 index.
TILs are now known to play a key role in mediating response to chemotherapy and improving clinical outcomes in invasive breast carcinoma, particularly in TNBC. The dense lymphoid infiltrate that is a key criterion for the diagnosis of CMF typically contains a high proportion of CD8+ cytotoxic T lymphocytes that likely accounts for the improved prognosis associated with this morphological variant of breast cancer.
Although CMF shows morphological similarities to Epstein Barr Virus (EBV) associated lymphoepithelioma-like carcinoma of the breast, expression of EBV protein is rare.
Gene expression profiling (GEP) studies, using unsupervised cluster analysis, has led to the classification of breast cancer based on molecular characteristics (Perou et al. 2000). The vast majority of CMF tumors segregate as basal-type, which are typically characterized by lack of expression of ER, PR, and HER2 and by positive expression of genes that are expressed in the basal-like cells of the normal breast and those associated with high proliferative activity. Further transcriptional analysis of the basal-like TNBC group has led to the recognition of heterogeneity within this group of tumors and the identification of basal-like TNBC molecular subtypes (Lehmann and Pietenpol 2013).
CMF tumors display a unique gene expression profile (Bertucci et al. 2006). CMF tumors are characterized by upregulation of genes involved in the immune response including those that control interleukin, interferon, and cytokine production. Genes involved in activation of the apoptosis pathway are also over-expressed compared with non-CMF basal-like tumors. In contrast, genes that are involved in remodeling of the cytoskeleton and those that control cell invasiveness are under-expressed in CMF (Weigelt et al. 2010). This profile correlates with the immunomodulatory subtype of TNBC tumors, associated with an improved response to chemotherapy, a more favorable prognosis and potential for treatment with immunotherapy.
BRCA1, located at chromosome 17q12–21, is a tumor suppressor gene. Germline mutations of BRCA1 account for 30–40% of hereditary breast carcinomas. Most breast tumors that develop in patients with BRCA1 germline mutations are of CMF type. A diagnosis of CMF may, therefore, signify a BRCA1 mutation and consideration of genetic testing. However, not all patients with CMF have a BRCA1 mutation. It is estimated that approximately 20% of patients with CMF test positive for BRCA1 mutation. CMF is uncommon in patients with BRCA2 mutations.
Somatic BRCA1 mutations and BRCA1 promoter hypermethylation have also been reported in CMF. Promoter hypermethylation may lead to inactivation of the BRCA1 gene and contribute to breast carcinogenesis.
The most common somatic mutation in CMF is TP53, which occurs at a higher rate than in non CMF tumors. TP53 mutations are common in all forms of breast cancer and overall are associated with a worse prognosis. However, the prognostic impact appears to vary with molecular subtype and may be limited to ER positive tumors.
In keeping with the high rate of TP53 mutations in CMF, high-density array comparative genomic hybridization (CGH) studies have demonstrated a high level of genomic instability in these tumors. Gains of 1q and 8q are common to all basal-like carcinomas and are observed in CMF. CMF tumors also show a specific pattern of genetic alteration including recurrent gains at 3p, 9p,10p, and 16q and losses at 4p and amplicons of 1q, 8p, and 10p (Vincent-Salomon et al. 2007).
When confronted with a well-circumscribed invasive breast tumor with no associated DCIS and triple negative for ER, PR, and HER2, it is always wise to maintain a broad differential diagnosis.
Recognition of CMF is important due to its association with the BRCA1 germline mutation. A diagnosis of CMF, particularly in a younger patient, will lead to consideration of genetic testing for the patient and her/his family. CMF also has a more favorable prognosis than a grade 3 invasive carcinoma NST. These characteristics emphasize the importance and value of careful assessment of breast tumor morphology, even in this era of sophisticated genetic methodologies.
The main alternative diagnosis is invasive carcinoma NST. To date there has been considerable interobserver variation in the diagnosis of MC and AMC. Application of the more relaxed histological criteria proposed by the WHO and outlined above should improve observer agreement in recognizing CMF and distinguishing it from NST. In view of the clinical relevance, it is more important to entertain the diagnosis than to assign these tumors to the NST category. The preferential expression of basal type immunohistochemical markers may be helpful in making a diagnosis of CMF.
The possibility of a metastasis from an alternative primary site or an intramammary lymph node replaced by metastatic tumor should be borne in mind. This is less likely when dealing with a single tumor. Tumors that metastasize to the breast include melanoma, which may be pigmented, ovarian carcinoma, lung carcinoma, sarcomatoid carcinoma, and sarcoma and primary sarcoma. Thorough evaluation of tumor morphology and consideration of the clinical history and radiological findings are key to reaching a correct diagnosis. In the case of an intramammary lymph node replaced by tumor, a search for lymph node capsule and any residual normal lymph node architecture should assist the diagnosis. Comparison with previous histology and the use of immunohistochemistry will also assist accurate diagnosis of an intramammary metastasis.
Lymphoma of the breast is a rare condition that should also be considered in the differential diagnosis. Breast lymphoma (mostly diffuse large B-cell lymphoma) may occur as a primary process or represent breast involvement by known systemic lymphoma. It usually forms a solid mass and may mimic carcinoma both radiologically and histologically. On close inspection, it lacks the syncytial growth pattern observed in CMF. Consideration of the diagnosis and the use of immunohistochemistry will clarify the diagnosis.
References and Further Reading
- Dua, I., & Tan, A. R. (2017). Immunotherapy for triple-negative breast cancer: A focus on immune checkpoint inhibitors. American Journal of Haematology and Oncology, 13, 20–27.Google Scholar
- Ellis, I. O., Al-Sam, S., Anderson, N., Carder, P., Deb, R., Girling, A., et al. (2016). Pathology reporting of breast disease in surgical excision specimens incorporating the dataset for histological reporting of breast cancer. (2016). Rcpath.org. Available from: http://www.rcpath.org. Accessed Sept 2017.
- Hicks, D. G., & Lester, S. C. (2017). Carcinoma with medullary features. Available from: https://app.expertpath.com. Accessed Sept 2014.
- Huober, J., Gelber, S., Coates, A. S., Viale, G., Ohlschlegel, C., et al. (2012). Prognosis of medullary breast cancer: Analyses of 13 international breast cancer study group (IBCSG) trials. Journal of Clinical Oncology, 11, 2843–2851.Google Scholar
- Lakhani, S., Ellis, I. O., Schnitt, S. J., Tan, P. H., & van de Vijver, M. J. (2012). WHO classification of tumours of the breast (pp. 46–47). Lyon: International Agency for Research on Cancer.Google Scholar
- Vincent-Salomon, A., Gruel, N., Lucchesi, C., MacGrogan, G., Dendale, R., Sigal-Zafrani, B., et al. (2007). Identification of typical medullary breast carcinoma as a genomic sub-group of basal-like carcinomas, a heterogeneous new molecular entity. Breast Cancer Research, 9, R24.CrossRefGoogle Scholar