Breast Pathology: Differential Diagnostic Dilemmas
Dr. Christopher Elston
Case 9 -
Infiltrative carcinoma, atypical medullary type
Dr. Jane Armes
A 38 year-old woman presented with a self-detected left breast lump. A slight increase in density of
the upper inner quadrant of the left breast was noted by mammography. There were no microcalcifications,
architectural disturbances or enlarged lymph nodes. Ultrasound showed a solid lesion in the 11 o'clock
region, measuring 15 mm. A fine needle aspirate was performed, with the diagnosis of 'adenocarcinoma'
and the patient went on to have a wide local excision of the breast lesion, with sentinel lymph node
biopsy and left axillary node sampling.
There was no family history of breast or ovarian cancer. The patient had had a previous total
abdominal hysterectomy for menorrhagia and was on hormone replacement therapy. The patient was P5G3 and
had breast-fed all three children.
Case 9 - Slide 1
The wide local excision specimen consisted of a fibro fatty portion of breast tissue, measuring 45 x
45 x 30 mm. On sectioning, a well-circumscribed cream nodule measuring 15 mm in maximum dimension was
The sections showed an infiltrative carcinoma, measuring 25 mm, formed by broad trabeculae of
syncytial cells (Figure 1a). There was no tubule differentiation (score 3), marked nuclear pleomorphism
(score 3) and approximately 52 mitoses per 10 (0.5mm) hpf (score 3). Thus the BRE grade was 3.
Confluent necrosis was associated with the carcinoma, together with abundant apoptosis (Figure 1b).
There was a moderate, predominantly peripheral lymphocytic response to the tumour. The tumour was not
A small amount of high grade DCIS was present beyond the confines of the invasive carcinoma, but this
did not extend to within the invasive lesion. There was no lymphovascular space infiltration by tumour.
A lymphocytic response was noted around non-neoplastic lobules adjacent to the carcinoma. Elsewhere the
breast tissue was unremarkable.
There was no metastatic carcinoma in the nine axillary lymph nodes examined (including three sentinel
Figure 1a - Infiltrative carcinoma composed of broad trabeculae of cells with areas of confluent necrosis (original magnification x 2).
Figure 1b - High grade nuclei, with numerous mitotic figures (original magnification x 40).
Infiltrative carcinoma, atypical medullary type. Possible basal-like carcinoma.
Immunohistochemistry showed complete absence of staining with antibodies to ER (Figure 2) and PR.
ERBB2 immunohistochemistry was negative (1+ by the DAKO HercepTest scoring system).
Figure 2 - Absence of ER staining in carcinoma (inferior) with positive internal control (normal breast epithelium) (original magnification x 4).
The invasive carcinoma showed accumulation of p53 protein (Figure 3), detected as strong nuclear
staining. There was diffuse, moderate staining of the tumour cells with antibodies to vimentin (Figure
4) and strong CK5/6 staining (Figure 5). Moderate, focal membranous EGFR staining was present.
CKIT/CD117 staining was negative.
Figure 3 - Strong nuclear staining of p53 (original magnification x 10).
Figure 4 - Vimentin staining within tumour cells (original magnification x 10).
Figure 5 - Tumour staining with the basal cytokeratin marker CK5/6 (original magnification x 10).
The patient was treated with chemotherapy (six cycles of fluorurocil, epirubicin and
cyclophosphamide), followed by radiotherapy. There is no evidence of recurrent disease 8 months
This is a high-grade carcinoma occurring in a young women. It fulfils the Ridolfi criteria of an
atypical medullary carcinoma.  since >75% of the tumour cells grow as syncytial sheets
and trabeculae, and the tumour exhibits three of the four secondary features of medullary carcinomas, ie,
marked nuclear pleomorphism, a moderate lymphocytic infiltrate into the tumour and absence of
microglandular architecture. The tumour does not show complete circumscription and hence cannot be
categorised as a true medullary carcinoma.
Recently Perou et al  (and since others), have identified a subset of invasive breast
carcinoma by their similar gene expression profile, using microarray gene expression techniques. The
expression profile of this subset of cancers includes basal cytokeratins, eg, CK5 and CK17. These have
been termed 'basal-like' breast carcinomas and represent approximately 15% of all breast cancers.
Further studies of this type of carcinoma indicate that the majority can be identified without
sophisticated genomic based technologies, by using standard morphologic and immunohistochemical
techniques. Thus Livasy et al  performed a morphologic evaluation of 23 basal-like breast
cancers as defined by microarray-based gene expression. When the histological features of these tumours
were compared with ERBB2 positive and 'luminal-type' carcinomas (also defined by their microarray
expression profile), the basal-like carcinomas were high grade compared with luminal tumours and compared
with all tumours they were distinguished by having an exceptionally high mitotic rate (>25 mitoses/10
hpf), pushing type of invasion, stromal lymphocytes and atypical medullary features.
Nielsen et al 
and Livasy et al  performed an overlapping panel of
immunohistochemical markers on the same cohort of basal-like breast cancers. Compared with other breast
cancers, basal-like tumours were characterised by ER, PR and ERBB2 negativity, p53 protein accumulation
(indicating p53 gene mutation) and EGFR, CK5/6, vimentin and/or CD117 (c-kit) protein expression. It is
important to note that the most consistent immunophenotype of these tumours is ER, PR and ERBB2
negativity. However, due to the potential problems of positively identifying cancer types by negative
staining (eg, technical failures), expression of CK5/6, EGFR, vimentin or CD117 should be analysed.
Interestingly, despite the descriptor of 'basal-like' carcinomas, only a minority of these cancers
express the basal cytokeratin CK5/6, although this rate is higher than in other carcinomas.
As described above, medullary-type features are often found in cancers which fall into the basal-like
group of tumours and there is definite overlap between these two groups of cancer. An
immunohistochemical study of typical medullary cancers by Jacquemier et al  determined that
the majority of these cancers expressed a basal-like phenotype. Another definable group of tumours which
overlap with medullary type and basal-like breast cancers are cancers arising in patients with germline
BRCA1 mutations. BRCA1 breast cancers have
been shown to be frequently 'basal-like' by both microarray expression analysis and immunophenotype.
The tumour described herein shows the morphological and immunophenotypic features of a basal-like
breast cancer. Given the young age of the patient and the basal-like features of the tumour, it would be
beneficial to advise of the risk of germline BRCA1 mutation in this
patient. Very few studies have specified the risk of carrying a germline BRCA1 mutation in a patient presenting with this cancer phenotype. It is likely
that this phenotype is a relatively sensitive indicator of BRCA1 germline
mutation, but not specific for this mechanism of cancer development (our own unpublished data). The
germline BRCA1 mutation status of this patient has not been tested to date.
Several studies have shown that cancers classified as basal-like have a poor prognosis. Further,
since they are ER and ERBB2 negative they are not treatable with either hormone therapies or
trastuzumab. It is interesting therefore that there is a relatively high rate of EGFR and CD117 (c-kit)
immunohistochemical staining in these cancers (between one third and one half of basal-like cancers),
molecules which already are the target of commercial therapies for non-breast malignancies. The
underlying molecular aberrations (if any) producing immunohistochemically-detectable expression of these
two molecules in basal-like breast cancers have yet to be determined, although the majority of basal-like
breast cancers do not carry the characteristic mutations seen in CD117-positive Gastrointestinal Stromal
Tumours.  Nevertheless, the potential for specific targeted therapy in this subset of
breast cancer needs to be explored.
- Ridolfi RL et al. Medullary carcinoma of the breast: a clinicopathologic study with 10 year follow-up. Cancer 1977. 40(4): 1365-1385.
- Perou CM et al. Molecular portraits of human breast tumours. Nature 2000. 406 (6797): 747-752.
- Livasy CA et al. Phenotypic evaluation of the basal-like subtype of invasive breast carcinoma. Mod Pathol 2006; 19(2): 264-271.
- Nielsen TO et al. Immunohistochemical and clinical characterisation of the basal-like subtype of invasive breast carcinoma. Clin Cancer Res 2004; 10(16): 5367-5374.
- Jacquemier J et al. Typical medullary breast carcinomas have a basal/myoepithelial phenotype. J Pathol 2005; 207(3): 260-268.
- Lakhani SR et al. Prediction of BRCA1 status in patients with breast cancer using estrogen receptor and basal phenotype. Clin Cancer Res 2005. 11(14): 5175-5180.
- van 't Veer LJ et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature 2002; 415 (6871): 530-536.
- Simon R.et al. KIT (CD117)-positive breast cancers are infrequent and lack KIT gene mutations. Clin Cancer Res 2004; 10 (1 Pt 1): 178-183.