Breast Pathology: Differential Diagnostic Dilemmas
Dr. Christopher Elston
Malignant Phyllodes Tumour
Dr. Puay Hoon Tan
A 44 year old Chinese lady underwent screening mammography in November 2002 which detected left breast
microcalcifications and two right breast nodules at 3 o'clock and 9 o'clock. Stereotactic-guided
mammotome biopsies of the left breast microcalcifications revealed benign fibrocystic changes, while
ultrasound guided trucut biopsies of the right breast nodules showed histological features consistent
with derivation from fibroadenomas.
In August 2003, she presented with a palpable enlarging right breast mass measuring up to
9.3 cm on ultrasonography. Correlation with the previous mammograms and ultrasound scans performed in
November 2002 disclosed that this enlarging right breast mass was in fact the nodule discovered in the
right breast at the 3 o'clock position nine months ago, which was approximately 3 cm in size then. The
patient had been advised to undergo excision of this nodule in November 2002, but had defaulted
follow-up. When she first started to experience progressive enlargement of the right breast lump the
following year, the country and its hospitals were facing serious issues with SARS (severe acute
respiratory distress syndrome). The patient therefore delayed seeking treatment, and only returned to
the hospital in August 2003, whereupon the mass was surgically excised. The section is from this excised
The submitted section showed sheets and fascicles of spindle cells with moderate hypercellularity and
obvious stromal cytologic atypia. Mitoses numbered up to 25 per 10 high power fields, with occasional
atypical mitoses. No necrosis was present in this section. Zones composed of only spindle cells devoid
of epithelial elements were present.
A fibroepitheliomatous appearance was noted at the edge of the lesion, resembling the
intracanalicular growth pattern of a fibroadenoma. A few small epithelial islands were embedded within
the hypercellular spindled areas, the latter merging fairly imperceptibly with the fibroepitheliomatous
Elsewhere, not represented in this section, the borders of the lesion were permeative and spindle
cells percolated into surrounding adipose stroma.
No heterologous elements were found. There were no atypical or malignant epithelial areas.
Lesional tissue abutted the surgical margins in a few foci.
The diagnostic conclusion was of malignant phyllodes tumour.
Review of the previous trucut biopsies (three3 cores were taken) showed features characteristic of a
fibroadenoma, and did not reveal changes that would have altered the histologic diagnosis rendered then.
- Malignant phyllodes tumour
- Metaplastic carcinoma
- Primary breast sarcoma.
The patient underwent a simple mastectomy in September 2003 that did not disclose any residual
She remained on regular follow-up, and was disease free at her last hospital visit in March 2006.
The presence of fibroepitheliomatous areas of the lesion with perithelial hypercellularity, compressed
epithelium lining elongated clefts, merging with hypercellular zones of stromal overgrowth, stromal
cytologic atypia and permeative margins, warranted a diagnosis of malignant phyllodes tumour.
Phyllodes tumours, also known as cystosarcoma phyllodes, are fibroepithelial neoplasms that generally
occur in an older age group than fibroadenomas. They tend to be large, seen mammographically as round or
lobulated, sharply defined, opaque masses. Indistinct borders are noted in a minority of cases. On
ultrasound, they are well circumscribed, but the presence of cysts and epithelium-lined clefts may lead
to an inhomogeneous appearance. Histologically, there is a leaf-like architecture with an exaggerated
intracanalicular growth pattern and increased stromal cellularity. Benign, borderline and malignant
phyllodes tumours are classified based on assessment of a constellation of histological features that
include stromal cellularity, stromal atypia, mitotic activity, stromal overgrowth, invasive versus
pushing margins and the presence of necrosis and haemorrhage. All phyllodes tumours, regardless of its
benign or malignant nature, have a propensity for local recurrence, especially if inadequately excised by
enucleation without a surrounding rim of tissue.
The need to separate benign phyllodes tumour from fibroadenoma traditionally has been related to the
recurrent and progressive potential of the former contrasted with the innocuous nature of the latter.
Fibroadenomas can occur synchronously with 4.2% of phyllodes tumours, as reported in one study. The
relationship between these two entities, however, is not fully understood. Early reports documented the
presence of areas of fibroadenoma within otherwise characteristic phyllodes tumours and alluded to
fibroadenoma being the possible precursor or progenitor of phyllodes tumour. More recently, molecular
evidence for the progression of fibroadenoma to phyllodes tumour was described, with Kuijper et al
expressing that fibroadenomas might progress not only in a stromal manner into phyllodes tumour, but also
in an epithelial direction toward carcinoma in situ. Because fibroadenoma-like areas are not encountered
uncommonly in phyllodes tumours, adequate macroscopic sampling is required for excision specimens to
avoid overlooking areas typical of phyllodes tumours, which might be focal within the tumour. In the
current age of pre-operative core biopsy diagnoses, the possibility of undersampling also raises
important management issues.
The trucut biopsies carried out earlier were deceptive in that there were no histologic clues to
suggest the presence of an underlying phyllodes tumour. The subsequent excision also disclosed focal
areas at the periphery of the mass that resembled fibroadenoma. It is presumed that the trucut biopsies
may have sampled this fibroadenomatous part of the lesion, and therefore were not representative.
An accurate preoperative distinction of phyllodes tumour from fibroadenoma cannot always rely purely
on histological appearances, as demonstrated in this case where only benign fibroadenomatous features
were present on the core biopsies. Whether additional cores would have been more elucidative is
uncertain, as in general, recommendation is for about three biopsy cores to be sampled from radiologic
breast densities, which was accomplished in this case. Clinico-radiologic correlation with consideration
of lesion size and patient age are other factors in deciding whether or not to advocate excision biopsy.
This lady was advised to undergo excision in view of the size of the lesion, but had defaulted, with
further subsequent delay occurring due to unforeseen circumstances.
Adjunctive biologic markers have been explored as possible predictors of the behavior of phyllodes
tumours, including p53, c-myc, c-kit, MIB-1 and S-phase fraction, microvessel density, CD34 and factor
XIIIa stromal expression. Some of these have also been investigated as potential adjudicators in core
biopsy findings of fibroepithelial neoplasms, but their utility needs further validation.
The two main differential diagnoses of malignant phyllodes tumours are metaplastic breast carcinoma
and primary breast sarcoma, both of which feature malignant spindle cell or sarcomatoid appearances.
Sometimes, the sarcomatoid component in all these three pathologic entities may be in the form of
malignant chondroid, osteiod or rhabdomyosarcomatous elements. The key to the diagnosis of malignant
phyllodes tumour is the presence of typical epithelium lined fronds, which may be focal within the
tumour, therefore requiring adequate macroscopic sampling. In metaplastic breast carcinoma with
sarcomatoid features, there are malignant epithelial elements in the biphasic subtype, either in the form
of ductal carcinoma in situ (DCIS) or invasive carcinoma. In the monophasic spindle cell variety of
metaplastic breast carcinoma, immunohistochemistry with epithelial markers can confirm keratin positivity
and therefore epithelial differentiation in the malignant spindle cells. CD34 and bcl2 have also been
reported to be useful in distinguishing phyllodes tumour from spindle cell/metaplastic carcinoma, as the
former tends to express CD34 and bcl2 in their stromal cells. Metaplastic carcinomas are treated as for
primary breast carcinomas with axillary dissection, as opposed to malignant phyllodes tumours in which
removal of axillary lymph nodes is not indicated. In primary breast sarcoma, malignant spindle cell
areas are neither associated with the fronding appearance of phyllodes tumour nor malignant epithelial
component of metaplastic carcinoma. Further immunohistochemical workup to determine the cell of origin
may be needed.
Final Diagnosis: Malignant phyllodes tumour
- Kuijper A, Buerger H, Simon R et al. Analysis of the progression of fibroepithelial tumours of the breast by PCR based clonality assay. J Pathol 2002; 197: 575-581.
- Noguchi S, Yokouchi H, Aihara T et al. Progression of fibroadenoma to phyllodes tumour demonstrated by clonal analysis. Cancer 1995; 76: 1779-1785.
- Ellis IO, Humphreys S, Michell M et al. Guidelines for non-operative diagnostic procedures and reporting in breast cancer screening. Sheffield : NHS Cancer Screening Programmes, 2001.
- Tan PH, Jayabaskar T, Chuah KL et al. Phyllodes tumours of the breast: the role of pathologic parameters. Am J Clin Pathol 2005 Apr; 123(4): 1-12.
- Tan PH, Jayabaskar T, Yip G et al. p53 and c-kit (CD117) protein expression as prognostic indicators in breast phyllodes tumours: a tissue microarray study. Mod Pathol 2005; 18(12): 1527-1534.
- Jacobs TW, Chen YY, Guinee DG Jr et al. Fibroepithelial lesions with cellular stroma on breast core needle biopsy: are there predictors of outcome on surgical excision? Am J Clin Pathol 2005; 124(3): 342-354.
- Jacklin RK, Ridgway PF, Ziprin P et al. Optimising preoperative diagnosis in phyllodes tumour of the breast. J Clin Pathol 2006; 59(5): 454-459.
- Dunne B, Lee AH, Pinder SE et al. An immunohistochemical study of metaplastic spindle cell carcinoma, phyllodes tumour and fibromatosis of the breast. Hum Pathol 2003; 34(10): 1009-1015.