|PHYLLODES TUMORS ARE RARE
fibroepithelial neoplasms that account for 0.3 to 1 per cent of all breast cancers in
females. The term cystosarcoma phyllodes was coined by Johannes Muller in 1838 to describe
the gross appearance of this "cystic, fleshy, leafy, tumor of the breast," which
he clearly stated to be a benign tumor. Lee and Pack reported the first case of
metastatic phyllodes tumor in 1931. Since then, metastatic disease has been noted to occur
to the lung, pleura, axial skeleton, central nervous system, pancreas, duodenum, thigh,
sacrum, and mandible.. The term cystosarcoma implies a relationship to cystic
disease and tends to overstate the malignant potential of this lesion. Thus, cystosarcoma
phyllodes is now felt to be a misnomer, and the World Health Organization now classifies
this breast neoplasm as a phyllodes tumor (PT). (Breast sarcomas often appear quite large,
see pic of angiosarcoma
of breast and CT
of the breast.)
The malignant counterpart of fibroadenoma is cystosarcoma
phyllodes or the newer term of phyllodes tumor, in which the epithelial elements are
benign, but the stromal tissue is malignant. It results from malignant degeneration of
fibroadenoma, estimated to occur in 1% of patients, who have fibroadenoma for many years.
At presentation, most women are between 40 and 50 years
with a typical presentation of a rapidly growing mass.
The phyllodes tumor has a lobulated, leaf-like appearance
and varies in size from 1 cm to greater than 15 cm. Microscopically, the branching,
hyperplastic ducts are surrounded by a stroma, which is mostly fibrous and much more
cellular than fibroadenoma. In addition, nuclear atypia and increased mitotic activity
The stromal components may contain liposarcoma,
leiomyosarcoma, rhabdomyosarcoma, malignant fibrous histiocytoma, angiosarcoma,
chondrosarcoma and osteosarcoma. For this reason, it is important to adequately sample the
neoplasm to determine whether ductal elements are present. In the absence of epithelial
cells, the neoplasm is classified as primary stromal sarcoma of the breast, which is
generally more aggressive than phyllodes tumor.
The clinical behavior of phyllodes
tumor is unpredictable. The majority of phyllodes tumors are local problems and
do not metastasize. Less than 20% of phyllodes tumors metastasize by vascular spread, most
commonly to the lung, pleura, and bone. (Hart, 1978; Pietrusz, 1978). Thus, lymph node
dissection is not indicated. Local recurrence is likely, if incompletely excised.
Therefore, a wide local excision is required.
Some recent studies have attempted to separate cystosarcoma
phyllodes into benign and malignant groups. The benign group is characterized by smooth,
non-infiltrative borders and the fibrous elements have minimal nuclear atypia and low
mitotic activity. This is in contrast to infiltrative borders, presence of nuclear atypia
and increased mitotic activity usually greater than five mitotic figures per 10 high power
fields in the malignant group (Hart et al; Pietrusz and Barnes). It should be mentioned
that not all metastatic phyllodes tumors meet the above criteria. Thus, it is more
appropriate to classify the phyllodes tumors into low and high grades and to treat these
tumors with wide clear margins.
Diseases of the Breast
Editor(s): Harris, Jay R., Lippman, Marc E., Morrow, Monica, Osborne, C. Kent
3rd Edition © 2004
Surgical Excision and Margins
The core principle of local therapy for phyllodes tumors, whether benign and malignant, is
wide excision with negative margins, preferably more than 1 cm, to achieve definitive
local control . Negative surgical margins independently predict improved disease-free
survival and decreased local tumor recurrence. Mangi found that recurrence
correlated with excision margin, but not with tumor grade or size. In 40 patients, local
recurrence occurred in 5, each who had margins less than 1 cm on initial excision or
reexcision. These five patients remained disease free after reexcision with a 1-cm
margin.The desired 1-cm margin width is based on retrospective analysis. Because these
lesions are rare, any trial to study optimal margin width is impractical. A pseudocapsule
of dense, compressed, normal tissue containing microscopic projections of the lesion
commonly surrounds phyllodes tumors. As a result, more tissue typically needs to be
removed to achieve the desired histologic 1-cm margin than might be predicted on the basis
of preoperative physical examination or imaging findings. Some authors argue that 2 cm
should be considered the standard desired surgical margin for excision of phyllodes
tumors, and 2 to 3 cm once a phyllodes tumor locally recurs . In practice, margins of 2 to
3 cm can be difficult to achieve, except when the breast is quite large.
Breast-conservation Versus Mastectomy
In the M.D. Anderson experience of 101 patients with phyllodes tumors, surgery included
local excision with breast conservation (47%) or mastectomy (53%). Microscopic surgical
margins were negative in 99% of patients. Six patients received adjuvant radiotherapy.
Local recurrence occurred in four patients, with an actuarial 10-year rate of 8%. The
investigators concluded that local failure in this group of largely margin-negative
patients with phyllodes of the breast was low, showing that breastconserving surgery with
appropriate margins is the preferred primary therapy.
Some investigators recommend mastectomy for histologically aggressive lesions. Kok,
Telesinghe advocates breast-conserving surgery with adequate resection margin for
benign and borderline lesions but recommend simple mastectomy without axillary dissection
for malignant lesions. This distinction is based on clinical rationale rather than
objective data. Kleer found that histologically high-grade malignant phyllodes
tumors have a favorable prognosis if widely excised without mastectomy. Most investigators
have concluded that breast conservation is reasonable as long as margins more than 1 cm
can be achieved, whether the lesion is benign, borderline, or malignant.
Reexcision following Narrow Margin Excision
Approximately 20% of phyllodes tumors recur locally if excised with no margin or with a
margin of only a few millimeters . The proportion may be somewhat higher with borderline
or malignant varieties and lower with benign phyllodes tumors. Nonetheless, authors'
opinions differ over whether immediate reexcision is necessarily required when an
unsuspected phyllodes tumor is diagnosed on permanent section after narrow-margin or
no-margin excision. In a review of 106 patients with benign phyllodes tumors from
Singapore, Chua, concluded that because only 16% of patients presumptively operated on for
fibroadenoma develop a local recurrence, a policy of close follow-up may be acceptable.
Zurrida from Milan made a similar argument based on their series of 216 patients.
With a mean follow-up of 118 months, they found that only 8% of benign lesions recurred
(vs. 20% to 23% for borderline or malignant phyllodes tumors, respectively) and that these
recurrences occurred significantly later following initial diagnosis (32 months vs. 18 to
22 months on average). On this basis, these authors suggest that a
wait-and-watch policy for benign phyllodes tumors may be considered in place
of mandatory surgical reexcision.
Technical Considerations in Lumpectomy
To achieve wide (?1 cm) surgical margins with lumpectomy, the surgeon may need to follow
special oncoplastic approaches, particularly when a phyllodes tumor develops
in a smaller breast. Tunneling through the fibroglandular tissue from a periareolar
incision, though appropriate for a fibroadenoma biopsy, is contraindicated with phyllodes
tumor excisions, because of the potential for tumor seeding in the tunneled bed. Even a
standard curvilinear circumareolar incision directly over the mass without removal of skin
may either be too small to obtain adequate surgical margins or leave excessive redundant
skin behind when a large section of fibroglandular tissue is removed.
Like the quadrantectomy advocated by Italian surgeons for breast conservation in the
1980s, ample full-thickness excisions including skin and resection of glandular tissue
down to chest wall muscle can be very helpful in increasing the surgical margins to the
desired 1-cm width. To remove a generous amount of tissue around the phyllodes tumor, an
incision can be designed to excise an island of skin immediately superficial to the
phyllodes mass. This approach allows en bloc removal of skin, tumor, and surrounding
fibroglandular tissue (Fig. 61.5B). The excision is then carried out full thickness from
the skin island, widely around the mass, and down to and including the pectoral muscle
The remaining tissue defect typically requires a mastopexy flap advancement for closure,
where fibroglandular breast tissue on each side of the residual tumor bed is widely
mobilized and advanced across the tumor bed to close the tissue deficit, both at the level
of the pectoral muscle and at the level just deep to the skin. This will improve the
cosmetic result and prevent adhesion between the skin and muscle surface.
Phyllodes tumors, like soft tissue sarcomas, are unlikely to have nodal metastases, making
axillary dissection unnecessary in patients with phyllodes tumors. Although axillary nodes
are palpable in 20%, fewer than 5% will have nodal metastases. In a series of 45 patients
with phyllodes tumors who underwent axillary staging, none was found to have axillary
metastases. Isolated cases of patients with gross nodal disease have been reported, albeit
in association with locally advanced histologically aggressive disease . Thus, axillary
dissection can be considered in the exceptional case with clinically suspicious axillary
nodes, although minimally invasive nodal sampling is preferable, given the low-level
likelihood of finding axillary disease.
Radiation and Systemic Therapy
Overall, the role of radiation therapy for phyllodes tumors has
remained unclear. Clinical data supporting the use of adjuvant or neoadjuvant
radiation and/or systemic therapy are based on anecdotal case reports as opposed to large
patient series. No series has shown radiation therapy to be of benefit in the primary
treatment of phyllodes tumors. The M.D. Anderson series did not support the use of
adjuvant radiotherapy for patients with adequately resected disease.
Based on anecdotal experience, adjuvant radiation therapy may be appropriate treatment of
locally recurrent phyllodes tumors in selected cases . Phyllodes tumors are so uncommon
and biologically heterogeneous that no large series of locally recurrent phyllodes tumors
is ever likely to be collected. Thus, individualized therapy choices must be made. For
example, radiation therapy may be considered when phyllodes tumors recur following
mastectomy. In these cases, adjuvant radiation therapy may be justified after reexcision,
because any further local recurrence could be debilitating and difficult to treat, despite
the absence of definitive proof of efficacy. Chest wall recurrence following mastectomy
may demand extensive surgery such as multiple partial-rib resections with rigid chest wall
methylmethacrylate reconstruction and myocutaneous flap coverage, which would be better
avoided if possible.Anecdotal cases support the use of combined chemoradiation following
phyllodes tumor recurrence. In a case study of a locally recurrent malignant phyllodes
tumor, neoadjuvant hyperfractionated radiotherapy, superficial hyperthermia, and
ifosfamide were administered after the second local recurrence of this tumor. Toxicity was
reportedly mild. Resection of the tumor bed revealed a pathologically complete response,
with an actual disease-free follow-up of 48 months. Although phyllodes tumors have been
shown to variably express steroid receptors, there is no known value to adjuvant endocrine
therapy with tamoxifen or aromatase inhibitors. There would be little rationale for using
these drugs, because (a) steroid receptor protein expression decreases with increasing
malignancy, (b) steroid receptors are primarily expressed by the epithelial component of
phyllodes tumors, and (c) only the stromal component of phyllodes tumors metastasizes).
Overall, the systemic treatment principles of phyllodes tumors are driven by similar
principles to those governing the management of sarcoma. Chaney observed that
patients with stromal overgrowth, particularly when the tumor was larger than 5 cm, were
found to have a high rate of distant failure. The authors suggest that such patients merit
consideration of a trial that examines the efficacy of systemic therapy (2). Experimental
data in a mouse xenograft model suggest that malignant phyllodes tumors may be sensitive
to doxorubicin, vincristine, and cyclophosphamide. Burton found that two of three
patients with metastases achieved effective palliation when treated with cisplatin and
etoposide combination chemotherapy. Ifosfamide may be an active agent for treatment of
Phyllodes tumors clinically resemble fibroadenomas with the following exceptions: (a) They
tend to present in women about 10 years later than typical palpable fibroadenomas, (b)
they are capable of rapid growth, and (c) in advanced cases, skin ulceration and open
nonhealing wounds can occur. Clinical suspicion of phyllodes tumor based on clinical
criteria (e.g., older age, rapid growth, or large tumor size) is important, because both
imaging and tissue sampling have serious limitations with this tumor. When a phyllodes
tumor is clinically suspected, mammograms and ultrasound evaluation are advised, although
the imaging studies may fail to distinguish the phyllodes tumor from a benign
fibroadenoma. Core needle sampling is preferable to FNA for preoperative tissue sampling,
because fibroadenomas and phyllodes tumors have similar cytologic features. Even core
needle sampling can misdiagnose phyllodes tumors, because their histology can resemble
fibroadenomas, especially with low-grade benign phyllodes tumors. If core needle biopsy
yields a diagnosis of phyllodes tumor and breast tissue is dense, breast MRI may be
considered for assessment of associated pathology and extent of disease, as well as to
serve as a baseline for surveillance.
Clinical management of phyllodes tumors consists of wide surgical excision to achieve
surgical margins more than 1 cm to decrease the likelihood local recurrence in the breast.
Wide resection of phyllodes tumors appears to be the most important aspect of therapy.
When phyllodes tumors are excised with a narrow or no margin, reexcision should be
performed, with the possible exception of benign phyllodes tumors in which reexcision is
excessively difficult or deforming. Distant metastasis is uncommon, even with malignant
phyllodes tumors, but is more common among tumors exhibiting stromal overgrowth.
Histologic differentiation between benign, borderline, and malignant phyllodes tumors only
loosely correlates with distant metastatic risk. Tumor markers (e.g., Ki-67, p53, ER, and
PR) fail to predict long-term outcome and are not useful in clinical management questions.
The role of radiation therapy and systemic therapy is unproven and must be considered on a
case-by-case basis. Radiation therapy has no role in the initial management of primary
phyllodes tumors. However, phyllodes tumors that are initially excised with wide margins
or mastectomy and that then recur locally may warrant adjuvant chest wall radiation after
reexcision, because the high morbidity of second or third local recurrences from these
locally aggressive lesions could be devastating. The role of systemic therapy remains
unclear, because of lack of supportive clinical data other than from anecdotal reports.
When used for treatment of metastatic disease, guidelines for treating sarcoma, rather
than breast carcinoma, should be followed.
there is controversy about therapy, consider the following;
Cystosarcoma Phyllodes or Phyllodes
This tumor is unique to the female
breast. It should be clearly differentiated from a JUNEVILE FIBROADENOMA.
These tumors can be of any size but are usually diagnosed as a large, rapidly
growing, bulky breast tumor (over 5 cm in size). They can occur at any age but is seen
mostly in women in their fifties.
These tumors have the mammographic appearance of a fibroadenoma.
histological diagnosis is made by excisional biopsy. Most of these tumors are usually
benign. However, a few can be malignant. We usually request that our pathologist
classifies these tumors as, benign, malignant or borderline Cystosarcoma Phyllodes.
treatment of these tumors is surgical. As they have a significant rate of local
recurrence, surgical local control is essential. For small benign cystosarcoma phyllodes,
a wide local excision can be performed with meticulous, life long, post-operative
monitoring. For large tumors or malignant / borderline tumors, a total mastectomy is the
procedure of choice. No axillary lymphadenectomy is performed as the rate of axillary
metastasis is less than 0.9%. Some authors have proposed a simultaneous sentinel
lymphadenectomy; although this approach is reasonable, no guidelines have been set.
consider the following study which
discusses the role of radiation:
An analysis of 78 breast sarcoma
patients without distant metastases at presentation
S. McGowan, B.J. Cummings, B. O'Sullivan, C.N. Catton, N. Miller, T. Panzarella International Journal of Radiation Oncology*Biology*Physics,
46:2 : 383-390
Breast sarcoma is a rare tumor comprising less than 1% of
breast malignancies. Recommendations for surgical treatment have been varied, with some
advocating mastectomy , and others allowing that local excision may be acceptable therapy
under certain conditions Most authors have felt that nodal dissection is
A retrospective review of the experience at The Princess Margaret Hospital for the
treatment of breast sarcoma was undertaken to correlate outcome with various patient and
treatment factors.Purpose: A retrospective review of a single cancer center experience was
undertaken to identify clinical or treatment prognostic factors for these unusual tumors,
to allow for a recommendation regarding management.
Methods and Materials: The charts of 76 women and 2 men
with breast sarcoma and without distant metastases at presentation registered from 1958 to
1990 were reviewed. Pathology was centrally reviewed in 54 cases. Histology, tumor size,
grade, nodal status, age, menopausal status, history of benign breast disease, extent of
surgery, resection margins, and radiation dose were each examined as potential prognostic
factors by univariate analysis. To allow an analysis of radiation dose, total dose was
normalized to a daily fraction size of 2 Gy.
Results: The median age at diagnosis was 50.5 years
(1382 years). The pathologic diagnosis was found to be malignant cystosarcoma
phyllodes in 32 patients, with the remainder being stromal sarcoma (14), angiosarcoma (8),
fibrosarcoma (7), carcinosarcoma (5), liposarcoma (4), other (8). Eighteen patients had
grade I or II tumors, 43 had grade III or IV, and 18 were not evaluable. The 5- and 10-year actuarial rates for all 78 patients were 57% and 48%
for cause-specific survival (CSS), and 47% and 42% for the relapse-free rates (RFR),
respectively. The local relapse-free rate (LRFR) was 75% at both 5 and 10 years. The
5-year CSS for grade I or II tumors was 84% versus 55% for grade III or IV tumors (p
= 0.01). Conservative surgery versus mastectomy did not lead to statistically significant
different outcomes for CSS, RFR, or LRFR. The comparison of positive
versus negative margins showed a 5-year LRFR of 33% versus 80% (p = 0.009).
Pairwise comparisons of the 5-year CSS of 91% for > 48 Gy versus either 50% for 48 Gy or 50% for no radiation
showed p-values of 0.03 and 0.06, respectively.
The authors propose
that if negative surgical margins can be achieved, breast sarcoma should be managed by
conservative surgery with postoperative irradiation to a microscopic tumoricidal dose (50
Gy) to the whole beast, and at least 60 Gy to the tumor bed. The decision to
treat should be preceded by a preoperative multidisciplinary assessment. It is also
recommended that an axillary lymph node dissection is not indicated, with the possible
exception of patients with carcinosarcoma.
The CSS, RFR, and LRFR experience, and median age at
diagnosis found in this group of patients are consistent with the results reported
elsewhere. With a 10-year CSS of 48%, a significant number of patients with this disease
will survive; therefore, it would be helpful to be able to define a group of patients in
whom a policy of breast conservation may be attempted. We have attempted to look carefully
at the effect of completeness of resection, and adjuvant irradiation, as these have proven
to be significant factors for sarcomas in other anatomic sites. Grade and size are such
well-accepted prognostic factors for sarcomas that they are an integral part of the
staging system for sarcoma as defined by the AJCC and UICC. In a historic series, we found
grade to be significant for distant, but not local failure however, in more recent series,
we, and others, have found that grade is an important prognostic factor for local and
distant recurrence, as well as cause-specific survival. In the present series, we found
that grade was statistically significant for CSS but not local control, which reflected a
higher rate of distant failure for high-grade tumors. Despite the wide range of tumor
sizes, size was not found to be an independent prognostic factor, for any endpoint on
univariate analysis in this series, although it has been found to be an important
prognostic factor in other anatomic sites.
Many authors have noted that virtually all recurrences and deaths from this disease
occurred within 5 years. There was one patient who failed locally and distantly 21 years
after presentation with what was felt to be recurrent disease. Excluding this one patient
with a late failure, most patients recurred, and died within 5 years, and all did so by
7.5 years. All local failures were seen within 5 years.
We have elected to exclude benign cystosarcoma phyllodes, but to include all malignant
histologic subtypes in our analysis. Some authors have excluded angiosarcoma, or
cystosarcoma phyllodes from an evaluation of breast sarcoma, while others have
stated that malignant cystosarcoma phyllodes should be included in any review. The
decision to include or exclude these specific histologies seems to be due to a belief
about their behavior, rather than an observation that histology is a prognostic indicator
independent of other factors such as grade. However, a prospective review of localized
soft tissue sarcomas of the extremities by Pisters et al. has reported that the
histologic subtypes leiomyosarcoma, liposarcoma, and malignant peripheral nerve tumors
were independent prognostic factors for some end points. In our study, we found no
prognostic significance to the histologic subtype, which is consistent with the generally
held belief that histologic subtype is of minimal prognostic significance for sarcomas
elsewhere; we would not advocate treating breast sarcoma patients differently based on
histologic subtype, when grade and surgical margins are considered.
It is widely accepted that axillary lymph node involvement is uncommon in this group of
patients, with most series reporting only occasional cases. Of 36 patients who had a
nodal dissection, only 3 were found to have involved nodes, 2 of those 3 had
carcinosarcoma, and the third had fibrosarcoma. In the patients with carcinosarcoma, only
the epithelial component was found in the lymph node. In a separate series of 30 cases of
breast sarcoma, 7 patients had clinically enlarged nodes, of which only 1 had pathologic
involvement on node dissection. In a patient with a pure breast sarcoma, an axillary node
dissection does not appear necessary. There may be some merit in performing an axillary
dissection in patients with carcinosarcoma in light of our observation that 2 of 5
patients with this subtype had nodal metastases. One of 6 patients with carcinosarcoma had
involved nodes in another series.
It is difficult to separate the issues of tumor size, extent of surgery, adequacy of
margins, and postoperative irradiation in an evaluation of the most appropriate local
therapy. In those patients with larger tumors, one might expect that clear margins would
be more difficult to obtain, and it is those patients in whom it is likely that more
aggressive surgery would be performed, and postoperative irradiation given. It was not
possible to ascertain with accuracy if this sort of selection took place in our series.
However, if that was the case, then it is reasonable to assume that those patients with
larger tumors, those patients who received a higher dose of postoperative irradiation, and
those patients who had a mastectomy would all do worse on univariate analysis. As would be
expected, those patients who had positive margins had much worse local control. This
relationship underscores the importance of obtaining negative surgical margins; however,
it may be that the inability of a surgeon, using good surgical techniques, to obtain a
negative margin, reflects the biology of the disease. The answer to the question of
whether or not negative surgical margins are desirable is intuitively obvious, and
consistent with our previous findings. However, the issue of the extent of surgery
necessary, or the requirement for adjuvant irradiation, is not. There is no consensus
on the recommendation for the extent of surgical therapy in the literature. Many authors
have not found radiation to be of benefit, or have not commented on its use. while others
have advocated mastectomy and radiation, or breast-conserving therapy in selected cases.We
have not found any statistically significant difference in local control or survival
between those patients treated with complete mastectomy or conservative surgery. We
feel that in patients in whom negative surgical margins can be achieved, that breast
conserving surgery is indicated. In our experience, a positive surgical margin predicted a
higher rate of local failure; however, we did note that 2 of 3 patients who had
microscopically positive margins after conservative surgery were controlled locally with
radiotherapy. Two of 4 patients who had positive margins and radiotherapy failed locally,
and 2 of 2 patients with positive margins who did not receive radiotherapy failed. This is
compared to an overall local control rate of 80% in those patients with negative surgical
margins. These results suggest that radiotherapy may be effective in reducing the risk of
local relapse in patients with positive surgical margins, but that negative surgical
margins give the best opportunity for local control. However, given the small number of
patients in the positive margin group, conclusions must be interpreted with caution.
Our experience is consistent with the current understanding that sarcomas are no less
radiocurable than epithelial tumors. Adjuvant radiation to a dose of at least 60 Gy in 6
weeks is a standard part of the therapy of many soft tissue sarcomas.. At our center,
radiation is used for close margins, but is not routinely prescribed for margins of
greater than or equal to 2 cm, or if a fascial plane separates the margin from the tumor.
Given the fact that the standard dose of radiation delivered to patients with soft-tissue
sarcomas is at least 60 Gy, it was not unexpected that there was little difference in the
outcome between patients who did not receive radiation and those who received less than
the normalized median dose of 48 Gy. When one examines the curves for local control versus
dose, it can be seen that they separate, with the higher radiation dose patients having
the best local control. However, the p-value was only 0.10 for the comparison of
greater than 48 Gy versus less than 48 Gy. Statistical significance might have been
achieved if there were more patients in the treatment groups, or if patient selection
could be truly accounted for.Based upon the analysis presented in this paper, and the
general acceptance of the principle of conservative management of soft tissue sarcoma
where appropriate, we propose that if negative surgical margins can
be achieved, breast sarcoma should be managed by conservative surgery with postoperative
irradiation. The decision to treat should be preceded by a preoperative
multidisciplinary assessment. Postoperative radiation should deliver a microscopic
tumoricidal dose to the whole breast, and at least 60 Gy to the tumor bed. An axillary
lymph node dissection is not indicated, except for patients with carcinosarcoma, due to
the low incidence of metastatic lymph nodes observed in breast sarcoma patients.