51_01

RECURRENT TUMORS

Marie E. Taylor
Carlos A. Perez
Seymour H. Levitt

Principles & Practice of Radiation Oncology, Third Edition; edited by
C. A. Perez and L. W. Brady. Lippincott-Raven Publishers, Philadelphia
Copyright 1997. Chapter 51, pp 1415-1448.

Locally advanced breast cancer is defined by 1992 American Joint
Committee (AJC) staging criteria as stage IIIa and IIIb disease. Stage
IV disease also includes ipsilateral supraclavicular nodal involvement
in the absence of other sites of distant disease. The 1992 AJC staging
system is shown in Chapter 50. Any of the following clinical or
pathologic findings at presentation represents locally advanced
carcinoma staging: tumor size greater than 5 cm with clinically or
pathologically positive axillary lymph nodes; tumor of any size with
direct extension to ribs, intercostal muscles, or skin; edema
(including peau d'orange) and/or ulceration of the skin of the breast,
or satellite skin nodules confined to the same breast; inflammatory
carcinoma (T4d); metastases to ipsilateral axillary lymph nodes fixed
to one another or to other structures; metastases to the ipsilateral
internal mammary lymph nodes; and metastases to the ipsilateral
supraclavicular lymph nodes.

The proportion of patients presenting with locally advanced breast
cancer varies. Haagensen [ref: 105] described 11.6% of cases as having
locally advanced breast cancer. Tumor Registry data at Washington
University Medical Center indicate that approximately 6% of all newly
diagnosed clinically staged breast cases from 1990 through 1994 were
stage III (40% stage IIIa, 60% stage IIIb). When pathologic staging
criteria were used, the proportion of breast cancer cases at our
institution was 8% DCIS; 48% stage I; 35% stage II; 6% stage III; and
3% stage IV. Patients with stage III disease have 55% to 60% survival
at 5 years and 35% to 40% at 10 years. Optimal therapy for these
patients aims to maximize overall and disease-free survival and to
prevent the often serious morbidity of locally recurrent disease.

Natural History

Locally Advanced (T3 and T4) Tumors

Locally advanced breast cancer may evolve from a mass to an
infiltration of the deep lymphatics of dermis, causing edema of the
skin. More pronounced edema (peau d'orange) usually indicates that the
superficial lymphatics as well as the deeper ones are involved. Later
in the course of disease, fixation of the skin over the tumor and
localized redness occur, followed by ulceration and infiltration of
the overlying skin. Skin retraction may be caused by tumor invasion of
Cooper's ligament. Further indications of extensive involvement are
the appearance of satellite nodules and carcinoma en cuirasse, in
which the skin becomes plaquelike and yellowish, red, or gray. [ref:
105]

Lymphatic spread to the axillary, internal mammary, and
supraclavicular lymph nodes frequently occurs. The supraclavicular
lymph nodes most often become involved after high axillary or internal
mammary nodes are involved. However, if there is a large tumor burden
within the breast or if skin in the upper half of the breast is
involved, direct spread to the supraclavicular lymph nodes can occur.
The most common initial sites of hematogenous spread, in order, are
bone, lung, and pleura [ref: 54]; initial involvement of the liver and
brain can occur but is less commonly seen. [ref: 32,256]

Inflammatory Carcinoma

Inflammatory carcinoma of the breast is defined by both clinical and
pathologic criteria. The clinical definition is the presence of
warmth, erythema, and peau d'orange in the involved breast. The
alternative pathologic criterion is the presence of tumor emboli in
dermal lymphatics. [ref: 75,147] Inflammatory carcinoma of the breast
is a biologically aggressive tumor that presents a major challenge to
all oncologists. [ref: 17] Henderson and McBride [ref: 111] emphasized
the difference between primary and secondary inflammatory carcinoma of
the breast according to Haagensen's criteria. [ref: 105] Primary
lesions have acute presentation with erythema over a significant
portion of the breast and concomitant edema and ridging. Often, there
is no palpable mass. This description is most consistent with a
diagnosis originally described by Taylor and Metzler. [ref: 236] In
contrast, secondary inflammatory carcinoma of the breast is more
characteristically a neglected, locally advanced breast cancer with
inflammatory-like changes resulting from underlying necrosis or
ulceration. Making the distinction between these two lesions is often
difficult; dermal lymphatic invasion is present in both, suggesting
that the mechanism for the inflammatory skin changes is similar in
both groups.

Historically, inflammatory breast cancer has been characterized by a
high rate of locoregional recurrence after a mastectomy and also by
the rapid development of metastatic disease.

Clinical Presentation

Patients present with a heterogeneous collection of tumor and nodal
findings. Symptoms include local or regional pain, bleeding,
paresthesia, and paresis. A common presentation includes clinically
palpable findings in the breast, axilla, and supraclavicular region,
skin changes, and possibly arm edema or findings of brachial
plexopathy.

Diagnostic Workup

Physical examination must give special attention to documenting the
locoregional extent of disease and carefully checking potential sites
of hematogenous spread. Areas of erythema, edema, subtle subcutaneous
infiltration, and small subcutaneous nodules are easily missed and
could lead to marginal failures if not taken into account when
irradiation fields are planned. Fixation to the pectoralis muscle or
fascia can be determined by assessing the mobility of the mass with
the pectoralis muscle relaxed and contracted. Regional lymph nodes
should be thoroughly evaluated by careful clinical examination and by
chest computed tomography (CT) if radiographic evaluation of the
internal mammary lymph nodes appears indicated.

Laboratory studies should include a complete blood cell count and
serum chemistry profile with full liver function tests. In the event
of anemia, leukopenia, or thrombocytopenia, morphologic evaluation is
indicated, and bone marrow biopsy is necessary. Radiographic studies
should include chest x-ray, bone scan, and plain radiographs of
symptomatic regions or suspicious areas of increased uptake on bone
scans. Bone scans are generally recommended for patients with stage
III or IV disease even if the alkaline phosphatase level is normal
[ref: 79]; about 35% of patients with stage III cancer have an
abnormal bone scan. [ref: 13,35,48] If the patient has abnormal liver
function values, a CT scan of the abdomen should be obtained. If there
are any neurologic symptoms suggestive of cerebral metastases, a
contrast-enhanced CT scan or gadolinium-enhanced magnetic resonance
imaging (MRI) scan of the brain should be obtained.
Gadolinium-enhanced MRI is the preferred study if leptomeningeal
carcinomatosis is suspected.

Prognostic Factors

Prognostic factors can be intrinsic (related to the initial inherent
condition of the tumor itself, such as lymph node involvement and
histology) or extrinsic (related to the type and adequacy of therapy).
Donegan and colleagues [ref: 69] noted that the following extrinsic
factors were most frequently related to local recurrence.

Gross Characteristics of the Primary Tumor

Larger tumors, more diffuse tumors, and presence of edema are
associated with increased local recurrence. [ref: 69]

Histologic Tumor Grade

Incidence of local recurrence increases with more poorly
differentiated tumors.

Meyer and co-workers [ref: 160] showed a significant correlation of
the thymidine labeling index (TLI), an indication of growth fraction,
with survival rate. TLI below the median of 4.55 carries a 20%
probability of relapse at 4 years, in contrast to 52% probability of
failure with TLI above the median. Sylvestrini and associates [ref:
232] concluded that pretreatment TLI does not correlate with tumor
response to primary chemotherapy. However, it provides information on
the course of the disease, in that a high TLI significantly predicts a
higher progression rate after completion of therapy, a shorter time to
disease progression, and poorer probability of 4-year survival in
comparison with low TLI. This information can allow selection of cases
requiring more aggressive therapy.

Hormonal Receptors

Patients without hormonal receptors have a significantly lower
survival rate and are not likely to respond to hormonal therapy. [ref:
156,257] Conflicting reports have been published on the correlation
between the presence or absence of hormone receptors and chemotherapy
response. [ref: 124,143]

Fisher and colleagues [ref: 84] found that the most significant
pathologic features related to estrogen receptor (ER) and progesterone
receptor (PR) positivity and concordance (ER+PR+, ER-PR-) were
well-differentiated nuclear and histologic grade, slight or absent
tumor lymphoid infiltration, slight or absent necrosis, and moderate
or marked elastica. All of these factors were directly or indirectly
related to tumor differentiation. Disease-free survival decreased from
concordant ER+PR+ to discordant receptors (ER+PR- or ER-PR+), with the
worst prognosis in patients with concordant ER-PR-.

Likewise, McGuire and colleagues [ref: 49,157] observed a strong
correlation between hormone receptor content, percent of cells in S
phase of the cell cycle and aneuploidy and suggested that these
measurements in concert may identify a subset of patients who are at
increased risk for recurrence. Arnarlov and colleagues [ref: 7]
reported that aneuploidy correlated significantly with histologic
grade, axillary perinodal growth, and disease-free survival.

Chevallier and associates [ref: 43] noted that ER and PR have their
own prognostic value and should be considered among other classic
prognostic factors.

DNA Index

Von Rosen and co-workers [ref: 251] found nuclear DNA distribution
patterns to be of prognostic value in node-negative patients but not
in node-positive patients. Fallenius and associates [ref: 77] also
noted that tumors with normal DNA content (diploid or euploid) had a
better prognosis. Dressler and colleagues [ref: 72] noted that
aneuploidy and high S-phase fraction correlated with absence of
steroid receptors.

Involvement of Axillary Nodes by Tumor

Local recurrence developed in 26% of patients with involved axillary
nodes, compared with 6.5% of patients without it; the greater the
number of nodes involved, the more likely the local recurrence. Fisher
and associates [ref: 83] also reported diminishing survival rate with
greater number of metastatic axillary lymph nodes. Baldursson [ref:
15] described 5-year relapse-free survival rates of 91% for pN0, 50%
for pN1, and 19% for pN2-3. The 5-year survival rates for the groups
were 86%, 53%, and 30%, respectively.

Response to Chemotherapy

Tumor and axillary nodal response to neoadjuvant chemotherapy has been
shown to be an indicator for disease-free survival. [ref:
29,154,201,241,245] McCready and co-workers, [ref: 154] in 136
patients with locally advanced breast cancer who received preoperative
chemotherapy followed by mastectomy and axillary dissection, noted on
univariate analysis that number of metastatic lymph nodes, clinical
tumor stage at presentation, clinical and pathologic response, and
menopausal status were significant variables directly associated with
overall and disease-free survival. When evaluated by multivariate
regression, surgical lymph node staging was the most important factor
for disease-free survival. Borger and associates [ref: 29] found
primary tumor size and clinical nodal status to be independent
prognostic factors for locoregional tumor control; patients whose
mastectomy specimen had no gross residual disease had a 93% survival
at 5 years, compared with 30% in patients with less marked response to
therapy (P < 0.01).

General Management

This heterogeneous group of breast cancers includes large primary
tumors, tumors with fixed ipsilateral axillary nodes (N2), tumors with
ipsilateral supraclavicular or infraclavicular nodes, and tumors
producing arm edema. [ref: 1,92] In general, life expectancy is
primarily determined by the high probability (about 80%) that these
patients will develop bloodborne metastasis. [ref: 16,184] Because of
the compelling need for systemic therapy, chemotherapy has assumed a
principal role in most treatment trials of patients with locally
advanced and inflammatory breast cancer. Radiation and surgery each
have important roles to optimize locoregional tumor control.

General management for patients with locally advanced and inflammatory
breast carcinoma is typically defined by operability. There is a
consensus that all patients with technically resectable disease should
have surgery. Historically, borderline resectable or unresectable
locally advanced breast cancers have been treated by radiation therapy
alone. More recently, the use of chemotherapy or hormonal treatment
(or both) before surgical resection and irradiation has played a more
prominent role. In about 80% of cases, borderline inoperable or
inoperable tumors treated with chemotherapy regress sufficiently to
become operable. [ref: 114,179] Chemotherapy may also sensitize the
tumor to local irradiation. Indicators for postoperative radiation
therapy include tumor size greater than 5 cm in diameter; skin,
fascial, or skeletal muscle invasion; poorly differentiated tumors;
lymphatic permeation, four or more positive axillary nodes; matted
nodes at presentation; gross extracapsular tumor extension; and
inadequate surgical margins. Relative total doses of radiation are
determined by the bulk of the residual tumor (primary and/or nodal)
and the adequacy of the surgical margins.

In inflammatory breast carcinoma, chemotherapy followed by mastectomy
and irradiation provides the best tumor control and survival data.
[ref: 176]

Radiation Therapy Techniques

Irradiation of the Inoperable Breast

Patients whose breast cancer is technically inoperable should receive
radiation therapy to the breast, supraclavicular nodes, and axillary
nodes. Treatment of the ipsilateral internal mammary lymph nodes may
be indicated for medial chest wall involvement or if there is clinical
or radiographic involvement of the internal mammary node chain. The
breast is treated with photons through tangential fields with borders
similar to those used in early breast cancer, ensuring that all
potential tumor-bearing tissues are adequately covered. Treatment of
the intact breast and draining lymphatics in patients with advanced
breast cancers presents several technical challenges:

1. Homogeneous irradiation of the breast tissue despite its
half-oblate geometric shape.

2. Adequate skin and dermal dose. In contrast to early breast cancer,
in which there is no clinical need to treat the skin and underlying
dermis, in advanced breast cancer gross or subclinical skin or dermal
involvement is frequently present and must be treated.

3. Precise matching between the plane of the inferior border of the
supraclavicular field and the plane of the superior border of the
medial and lateral breast tangential fields. With the sharp beam edges
of the modern linear accelerator, inadequate matching may result in a
thin cold plane of tissue that may harbor a nidus for recurrence or a
thin plane of matchline fibrosis.

4. Minimal beam divergence into the lung from the medial and lateral
breast tangential fields and minimal dose as low as possible to the
opposite breast from the lateral breast tangential.

5. Adequate coverage of the internal mammary nodes. Inclusion of the
internal mammary nodes in the breast tangential fields often results
in irradiation of too much lung. On the other hand, use of a separate
single anterior internal mammary field that matches the medial border
of the medial breast tangential field produces a cold wedge of breast
tissue that is often unacceptable; this cold wedge may harbor gross or
subclinical deposits of cancer cells.

Adequate Skin and Dermal Dose

Adequate skin and dermal dose is of particular importance.
Technically, it is easily achieved by adding bolus over the entire
breast using a box-shaped bolus that provides an adequate skin dose as
well as homogeneous irradiation of the breast tissue.

The more difficult clinical issue is determining the necessary number
of bolus treatments, given the limits of normal skin tolerance.
Empiric clinical experience has been to use bolus in 40% to 60% of
treatments, either consecutively or every other day. If a separate
internal mammary field is used, it is very important to ensure
adequate coverage of the skin at the junction with the medial
tangential portal.

Irradiation of the Chest Wall

Irradiation of the chest wall after mastectomy can be accomplished
with tangential photon fields, as used in the treatment of the intact
breast, or with appositional electron beams. If tangential photon
fields are used, the technical challenges are similar to those
encountered in treatment of the intact breast. Adequate skin and
dermal dose are achieved by adding bolus over the entire field for one
third of the treatments and adding bolus to the scar alone for an
additional third of the treatments. In treatment of chest wall
recurrences, bolus to the entire field should be used for about one
third of the treatments and bolus with a generous margin around the
recurrence for most, if not all, of the remaining treatments.

Several electron-beam techniques can be used as an alternative to
tangential photon treatment; the simplest is a single appositional
field using 9- to 12-MeV electrons. Bolus should be used for part of
the treatment to increase the surface dose beyond the 80% to 90%
typically given with these beams and to minimize the lung dose.

The difficulty with a single appositional electron field is poor
lateral coverage (Fig. 51-1); a single, medially angled electron field
may be added to boost the depth coverage in the lateral region. [ref:
122] Otherwise, use of more sophisticated electron arc [ref:
133,134,158,174] or pseudoarc technique [ref: 31] may be required.
Electron arc therapy is a technically demanding and time-consuming
treatment. Field shaping requires the use of custom lead strips or a
Cerrobend cast placed on the patient's skin. It is often unappreciated
that the surface dose in the electron arc technique is much lower than
that for the single electron beam of the same energy, and the
effective depth of coverage is less. Use of bolus to bring up the
surface dose becomes mandatory.

Field Borders

Anatomic landmarks defining the field borders for treatment of
breast/chest wall tangentials, supraclavicular nodes, internal mammary
nodes, and axilla are similar to those used in the treatment of early
breast cancer (Fig. 51-2A, Fig. 51-2B, and Fig. 51-2C).

Examples of various field arrangements for irradiation of the chest
wall and regional lymphatics are shown in Figure 51-3. Table 51-1
describes the advantages and disadvantages of the four techniques. The
dose distributions for two tangential fields to the chest wall and
internal mammary portal are illustrated in Figure 51-4.

Matchline Technique

Many methods have been used to achieve an ideal match of the anterior
oblique supraclavicular field caudal edge and the cephalad edge of the
tangential field. A nondivergent supraclavicular field edge is always
achieved by blocking the inferior half of the field. A variety of
methods achieve a nondivergent edge from the tangential beams,
including blocking and table angulation in combination with collimator
angulation. Multiple reports describe techniques using custom
Cerrobend blocking for the cephalad tangential border. [ref:
47,117,135,212] Another technique uses a gravity orbited block to
achieve a nondivergent edge independent of gantry angle. [ref: 228]
Use of a rotatable half-beam block is another matching technique.
[ref: 213,247]

The Mallinckrodt Institute of Radiology Radiation Oncology Center has
used a matchline technique adapted from Lichter and co-workers [ref:
142] that mathematically yields nondivergent edges for the tangential
cephalad border by calculation of appropriate table and collimator
angle combinations depending on gantry angle. Both adaptations are
needed to avoid the trapezoid effect of beam divergence that is
prominent for gantry angles other than directly lateral. This method
has been effective, and field edge feathering has been unnecessary;
however, it requires a great deal of technical time and multiple
calculations at the time of simulation. Precision in day-to-day setup
requires careful technical attention. The matchline technique was
recently changed to employ asymmetric jaws to beam split all portals
along the central axis plane. This technique uses one isocenter to
treat the opposed tangential breast field, the supraclavicular portal,
and the posterior axillary field. With the precision matchline, the
patient does not have to move in any direction on the treatment couch.
Details of this technique have been described. [ref: 126] Dosimetric
studies show that asymmetric jaws provide nearly equivalent field edge
definition and superior absorption in comparison with Cerrobend
blocks. The use of one isocenter results in a reduction of room
treatment time by a factor of two. A composite portal film, which
includes the medial tangential and supraclavicular portals, shows a
perfect matchline for all portals (Fig. 51-5A). Similar composite
portal films, taken with the previous method of geometric matching
with collimator and table angulation, exhibit slight overlap or
underdose regions in many cases (Fig. 51-5B).

Doses

A total dose of 50 to 60 Gy in 1.8- to 2-Gy daily fractions should be
given to the entire breast or chest wall. If no surgery is feasible,
the breast should be boosted an additional 10 to 25 Gy with external
irradiation (electrons or photons) using shrinking fields or with an
**192Ir implant to a total dose of 75 to 80 Gy. The boost dose is
determined by the volume of residual disease. The internal mammary
nodes, supraclavicular fossa, and axillary nodal areas should receive
45 to 50 Gy over 5 to 6 weeks if no macroscopic tumor is present. Any
gross nodal disease should then be boosted with an additional 10 to 15
Gy; an appositional electron-beam field may be used.

Treatment Results

Locally Advanced Breast Carcinoma

Technically Resectable Locally Advanced Breast Cancer

Three published randomized prospective trials [ref: 103,125,172,220]
specifically studied the efficacy of post-mastectomy irradiation in
locally advanced patients. Spangenberg and associates [ref: 220]
randomly assigned 131 patients with technically resectable stage III
breast cancer to receive either radiation therapy and single-agent
chemotherapy (cyclophosphamide) or multiagent chemotherapy alone. The
multiagent chemotherapy regimen included CMF (cyclophosphamide,
methotrexate, and 5-fluorouracil [5-FU]) and vinblastine. The 5-year
relapse-free survival rate was 40% for patients in the former group
and 27% in the latter group. This difference was not statistically
significant.

Papaioannou and co-workers [ref: 172] treated patients with induction
chemotherapy (cyclophosphamide, doxorubicin, vincristine,
methotrexate, and 5-FU with tamoxifen or oophorectomy) followed by
surgery; 205 patients were randomly assigned to receive or not receive
radiation therapy. Both groups received postoperative chemotherapy
(cyclophosphamide, doxorubicin, vincristine, methotrexate, 5-FU, and
tamoxifen). The relapse-free survival was 79% with irradiation,
compared with 73% without it. This difference was not statistically
significant. The trial was flawed by 78 exclusions from the analysis.

In the Helsinki University Stage III Trial, 120 patients with operable
stage III breast cancer were randomly assigned to receive
postmastectomy irradiation, postmastectomy chemotherapy, or a
combination of both. [ref: 103,125] The first 60 patients assigned to
each arm also received levamisole. The incidence of distant metastasis
among patients receiving chemotherapy was 25% in the
chemotherapy-alone arm and 28% in the combination arm, compared with
70% among patients treated with irradiation alone. The incidence of
local failure among patients receiving radiation therapy was 20% in
irradiation-alone arm and 8% in the combination arm, compared with 58%
among those treated with chemotherapy alone. The reduction of both
distant metastases and local failures in patients receiving both
adjuvant chemotherapy and irradiation led to a statistically
significant improvement in overall freedom from relapse and survival
(Fig. 51-6). This important study strongly suggests that in patients
at high risk for locoregional failure, postoperative irradiation added
to mastectomy and adjuvant chemotherapy can have a significant impact
on survival as well as on locoregional tumor control.

Borderline Resectable Or Unresectable Locally
Advanced Breast Cancer

Radiation Therapy Alone

Five-year survival rates for series reporting results with radiation
therapy alone range from 10% to 25%. [ref: 16,33,44,87,102,195,242]
With moderate doses of radiation, local tumor control rates of 35% to
65% can be expected. [ref: 10,16,33,95,102] Doses of 75 to 100 Gy can
yield local control rates of 70% to 100% (Table 51-2) [ref:
3,58,87,168,231] but carry with them a significant risk of soft
tissue, rib, and lung injury. [ref: 221]

Table 51-3 shows local tumor control correlated with radiation dose by
use of mastectomy versus no mastectomy, with or without systemic
therapy. [ref: 177] Noted is the general trend of improved local tumor
control with systemic therapy. Local failure, with or without systemic
therapy in the absence of mastectomy, remains a large portion of
breast recurrence despite higher doses of radiation.

Radiation Therapy Alone Versus Irradiation
Followed by Chemotherapy

The high rates of distant failure in patients with borderline
resectable or unresectable breast cancer highlight the need for
adequate systemic therapy in these patients. [ref: 16,184] Several
published studies have suggested an improved freedom from relapse in
patients treated with chemotherapy and irradiation, compared with
historic results in patients treated with irradiation alone. [ref:
30,39,63,64,78,175,192,196]

Two randomized trials compared radiation therapy alone or combined
with chemotherapy. Between 1977 and 1980, 118 patients at the
Netherlands Cancer Institute were randomly assigned to receive
irradiation alone, irradiation followed by 12 cycles of CMF, or
irradiation followed by alternating cycles of AV (Adriamycin and
vincristine) and CMF. Patients in both chemotherapy arms also received
tamoxifen. The 5-year survival rate was 37% in all arms of the study.
No statistically significant difference in overall freedom from
relapse or locoregional recurrence occurred among any of the arms.
[ref: 203]

Between 1978 and 1985, 231 patients with locally advanced
noninflammatory carcinoma of the breast were randomly assigned to
receive either radiation therapy alone or irradiation plus CMF
chemotherapy; the randomization included both low- and high-dose CMF.
Parenthetically, 120 patients had technically resectable lesions and
underwent mastectomy before randomization. For the combined group of
technically operable and inoperable patients, a significant
improvement in freedom from relapse was seen in premenopausal patients
receiving chemotherapy. No difference occurred in overall survival.
For postmenopausal women, no difference occurred in either freedom
from relapse or overall survival. [ref: 65]

Chemotherapy Followed by Irradiation Versus
Chemotherapy Followed by Surgery

In two published prospective trials, patients with tumors made
resectable with initial chemotherapy were randomly assigned to either
mastectomy or breast irradiation; one study was from the National
Cancer Institute of Milan [ref: 63,246] and the other from the Cancer
and Leukemia Group B (CALGB) study group. [ref: 179] Neither trial
showed any statistically significant difference in freedom from
relapse or survival rate between the surgery and irradiation arms.

Chemotherapy, Surgery, and Radiation Therapy

Extrapolating from the superiority of the triple-modality arm
(surgery, radiation therapy, and chemotherapy) in the Helsinki trial
for operable stage III breast cancer, [ref: 103,125] it may be
persuasively argued that triple-modality treatment is most likely to
yield the best results in marginally unresectable or unresectable
advanced breast cancer. Several single-arm protocols also suggest that
this may be the case. [ref: 14,34,172,205] As yet, no randomized
prospective trial has compared such triple-modality treatment with
lesser treatment.

Perez and co-workers [ref: 177] described the retrospective outcome of
treatment for 281 women with locally advanced breast carcinoma. Median
follow-up was 6.2 years. Thirty-five patients were treated with
irradiation alone, 33 with irradiation and adjuvant chemotherapy, 81
with mastectomy and irradiation, and 32 with mastectomy, irradiation,
and chemotherapy (triple-modality). Disease-free survival correlated
with T stage and N stage is shown in Figure 51-7A. Cause-specific
survival paralleled disease-free survival in these groups.
Locoregional tumor control at 5 years 31%, 54%, 80%, and 91% in the
four groups, respectively (P < 0.0001) (Fig. 51-7B). More than 80% of
locoregional failures occurred within 3 years of initial treatment.
Systemic therapy and/or irradiation given before mastectomy yielded
better locoregional tumor control, disease-free survival, and
cause-specific survival, although these results were not statistically
significant (P = 0.1). The addition of any type of surgery versus no
surgery improved both local tumor control (P < 0.0001) and
disease-free survival (P = 0.0002).

Patterns of failure correlated with treatment modality are shown in
Figure 51-8; mastectomy substantially decreased locoregional
recurrence, but distant metastases were a major component of failure.
Noted is the lower overall proportion of total failure with
trimodality therapy, compared with single or bimodality treatment.
When locoregional tumor control and disease-free survival were
correlated with type of systemic therapy, no statistically significant
difference was found between chemotherapy plus hormones, chemotherapy
alone, or hormones alone. However, there was a statistically
significant improvement in locoregional tumor control and in
disease-free survival with use of any type of systemic therapy,
compared with no systemic therapy (P = 0.022 and P = 0.0013,
respectively). With respect to timing of adjuvant systemic therapy, a
suggested improvement in both locoregional tumor control and
disease-free survival was seen with neoadjuvant chemotherapy versus
adjuvant chemotherapy after irradiation and/or mastectomy, although
these differences were not statistically significant (P = 0.35 and
0.22, respectively). The authors concluded that trimodality therapy is
superior to any other treatment combination and that additional
clinical trials are needed to further define the roles and optimal
uses of the various treatment modalities.

Breast Conservation Therapy for Patients with
Locally Advanced Operable Breast Cancer

The advantages of neoadjuvant chemotherapy have been emphasized by
many authors [ref: 89,120,144,225,229]; these include reduced size of
the primary tumor and regional lymph nodes and eradication of distant
micrometastases. Jacquillat and associates [ref: 120] reported a 100%
complete tumor response in 98 patients after neoadjuvant chemotherapy
plus hyperfractionated irradiation (23 Gy in four fractions) and an
interstitial implant (22 to 30 Gy); the breast was preserved in most
patients, and only 13% locoregional relapse was reported. However,
follow-up was short at the time of the report, and long-term
complications were not assessed.

Bonadonna and colleagues [ref: 27] described results in 165 women with
breast cancer (tumors 3 to < 7 cm). Candidates for mastectomy were
distributed in groups of 33 patients to receive three or four cycles
of neoadjuvant CMF, CAF (cyclophosphamide, doxorubicin, and 5-FU), or
FEC (5-FU, epirubicin and cyclophosphamide). Tumor regression to less
than 3 cm was demonstrated in 127 (81%) of 157 patients who were
subjected to wide local tumor excision (quadrantectomy), allowing
preservation of the breast. Postoperative irradiation to the breast
(60 Gy in 6 weeks) was started 4 to 6 weeks after surgery. Among 75
women with minimum follow-up of 12 months, only 1 had failed in the
breast and 11 at other sites (including 3 in the ipsilateral
supraclavicular nodes).

Singletary and associates, [ref: 214] in 143 patients receiving
neoadjuvant chemotherapy, identified by review of mastectomy specimens
33 patients (23%) who were candidates for breast conservation therapy.
The observed incidence of multiple-quadrant involvement for the
following factors was found: persistent skin edema plus other factors,
65%; tumor greater than 4 cm with other factors, 56%, intramammary
lymphatic invasion with other factors, 20%; and multicentric tumor
plus other factors, 16%. The authors reported that persistent skin
edema was an ominous sign; 25% of patients with this finding
experienced chest wall recurrence after mastectomy and adjuvant
irradiation.

Touboul and colleagues [ref: 240] reported on 97 patients with locally
advanced nonmetastatic, noninflammatory breast cancer treated with
neoadjuvant chemotherapy (four courses of doxorubicin, vincristine,
cyclophosphamide, and 5-FU) followed by preoperative irradiation to
the breast and regional lymph nodes (45 Gy) and a fifth course of
chemotherapy after irradiation. In 37 patients with residual tumor
larger than 3 cm, mastectomy and axillary dissection were performed;
the other 60 patients were treated with conservation therapy. In 33
patients, there was complete tumor regression, and no surgery was
done; an additional boost of 20 Gy was delivered at the 85% level of
the basal dose with an interstitial **192Ir boost of 25 to 30 Gy in 15
fractions with external irradiation. Twenty-seven patients who had
residual mass 3 cm or smaller were treated with wide tumor excision
and axillary dissection followed by a boost of 20 Gy with **192Ir
brachytherapy. After completion of local therapy, all patients
received a sixth course of chemotherapy. Maintenance adjuvant
chemotherapy without an anthracycline was prescribed (12 monthly
cycles). The 5-year locoregional relapse rates were 16% with
irradiation alone, 16% with wide tumor excision and irradiation, and
5.4% with mastectomy (P = 0.04). The 5- and 10-year overall survival
rates were 80% and 69%, respectively, and disease-free survival rates
were 73% and 61%, respectively. There was no significant influence of
local treatment on outcome. Arm lymphedema was noted in 12.5% of
patients treated with axillary dissection and in 3% without axillary
dissection. Cosmetic results were satisfactory in 79% of patients
after wide excision and irradiation and in 71% of patients treated
with irradiation alone.

At the present time, breast conservation therapy for patients with
locally advanced breast carcinoma may be an option in selected cases.
However, longer-term follow-up is needed before this approach may be
accepted as standard of care.

Inflammatory Breast Cancer

The Institut Gustave-Roussy [ref: 192] retrospectively compared the
results of treatment with irradiation and hormonal manipulation (group
C: 60 patients treated from 1973 to 1975); induction AVM (doxorubicin,
vincristine, and methotrexate) plus radiation therapy, and maintenance
VCF (vincristine, cyclophosphamide, and 5-FU) (group A: 91 patients
treated from 1976 to 1990); and induction AVCMF (doxorubicin,
vincristine, cyclophosphamide, methotrexate, and 5-FU) plus radiation
therapy, and maintenance VCF (group B: 79 patients treated from 1980
to 1982). The 4-year disease-free survival rates were 15% for group C,
32% for group A, and 54% for those receiving the most intense
chemotherapy regimen, group B. The corresponding survival rates at 4
years were 42%, 53%, and 74%, respectively. The authors concluded that
a combination of intensive induction chemotherapy and maintenance
chemotherapy can improve both freedom from relapse and survival
compared with irradiation alone.

At Washington University, 179 patients received radiation therapy as
part of the treatment of nonmetastatic inflammatory breast cancer.
[ref: 176] All patients had histologic confirmation of carcinoma of
the breast; 22 patients had only pathologic evidence of dermal
lymphatic permeation, 56 had clinical findings only, and in 98 the
diagnosis was based on both clinical and pathologic findings. There
was no significant difference in cause-specific survival based on the
use of clinical or pathologic criteria for diagnosis. Patients were
treated with four different regimens: 33 were treated with irradiation
alone, 35 with a combination of irradiation and chemotherapy, 25 with
irradiation plus surgery, and 86 with trimodality therapy.
Triple-modality therapy consisted of three cycles of CAF and total or
modified radical mastectomy as feasible. This was followed by chest
wall and regional lymphatic irradiation (50 Gy in 5.5 to 6 weeks) in
combination with cyclophosphamide and 5-FU. Patients with positive
hormonal receptors were treated with hormonal therapy. If the lesion
was unresectable after three cycles of neoadjuvant chemotherapy or if
there was tumor progression, irradiation to the intact breast and
regional lymphatics was given and was followed, if possible, by
mastectomy. Patients with unresectable lesions received additional
doses of radiation to 75 Gy to the tumor volume.

The 5-year disease-free survival rate was 40% for patients treated
with trimodality therapy, 24% for those receiving irradiation and
surgery, and 6% for those treated with irradiation alone or
irradiation combined with chemotherapy without surgical treatment. The
corresponding 10-year disease-free survival rates were 35%, 24%, and
0%, respectively (P < 0.01) (Fig. 51-9). Clearly superior locoregional
tumor control was observed in patients undergoing surgical procedures:
79% with three modalities, 76% with irradiation and surgery, and only
30% with irradiation alone or combined with chemotherapy (P < 0.01).
Distant metastases occurred in 57% of the group treated with
triple-modality therapy, 60% of those treated with irradiation and
surgery, and 85% of those treated with irradiation alone or combined
with chemotherapy. There was no significant correlation between type
of mastectomy or dose of radiation and locoregional tumor control and
survival. Significant morbidity of trimodality therapy (10%), although
somewhat higher than that seen with other modalities (3.2%), was
considered acceptable (P = 0.1). Trimodality had a less pronounced,
although still statistically significant, impact on disease-free
survival and cause-specific survival. The authors found no
statistically significant difference in locoregional tumor control or
disease-free survival based on the sequence of triple modality therapy
(neoadjuvant therapy versus chemotherapy, radiation therapy, and
surgery; surgery, radiation therapy, and chemotherapy; or surgery,
chemotherapy, and radiation therapy). The addition of any
chemotherapy, with or without the use of hormonal therapy, was shown
to result in a statistically significant improvement in locoregional
control (P = 0.01) and a strong although not statistically significant
trend for improved disease-free survival (P = 0.08) (Fig. 51-10A and
Fig. 51-10B). Clearly, prospective randomized studies documenting
therapy outcome by treatment sequence would be helpful.

Postoperative Radiation Therapy in Stage I and II Breast Cancer

Postoperative radiation therapy is recommended for patients with
lesions larger than 5 cm in diameter; any skin, fascial, or skeletal
muscle involvement; poorly differentiated tumors; positive or very
close surgical margins; lymphatic permeation; matted lymph nodes; four
or more positive axillary lymph nodes; or gross extracapsular tumor
extension. It appears that adjuvant irradiation can be effectively
given before, concurrent with, or after chemotherapy.

general management 02

Management of Lobular Carcinoma in Situ

LCIS appears to arise from the terminal duct-lobular complex and is
considered to be a marker of increased risk for invasive carcinoma.
[ref: 273] It is estimated to comprise approximately one third of all
in situ lesions of the breast. Women with a diagnosis of LCIS have a
10% to 37% risk of later developing invasive breast cancer
(Table 50-17).

The diagnosis is usually made in premenopausal women between 35 to 55
years of age (10 to 15 years younger than patients with invasive
breast cancer). [ref: 308] LCIS is invariably an incidental histologic
finding in breast biopsies, because the lesions are not palpable in
breast tissue and are generally invisible to the naked eye. [ref: 747]

LCIS has a high incidence of multicentricity ranging from 48% to 80%
[ref: 621] and bilaterality ranging from 18% to 67%. [ref:
100,202,742] The incidence of multifocal carcinoma of the breast was
reported to be about 20% in DCIS and 7% in 496 patients with invasive
ductal carcinoma, in contrast to 45% to 54% in 92 patients with
invasive lobular carcinoma. [ref: 302] In more than half of the
patients, the foci were in the same quadrant as the primary tumor, and
they were in the adjacent quadrant in 30%. In Farrow's study [ref:
202] of 270 women with LCIS, 33 (12%) had simultaneous bilateral LCIS,
16 (6%) developed subsequent contralateral LCIS (for a total of 18%
with bilateral disease), and 70 (26%) had previous, simultaneous, or
subsequent cancer.

Simkovich and associates [ref: 693] reported that 17% of contralateral
biopsies in lobular breast cancer (in situ or invasive ductal
carcinoma) were indicative for further therapy.

For many years, management of LCIS generally consisted of biopsy of
the involved and the contralateral breast. [ref: 706] In 1974, Wheeler
and associates [ref: 841] argued against mastectomy and proposed
careful lifelong follow-up for patients with LCIS, an approach that is
favored by most surgeons at the present time. [ref: 299,543,829]

The risk of a second breast cancer developing averages about 0.7% per
year, with a cumulative incidence of bilateral breast cancer of 4% to
26%. [ref: 433] Two thirds of the second breast cancers are
metachronous.

An important issue is whether patients with LCIS warrant therapeutic
intervention at initial diagnosis, because some may subsequently
develop invasive carcinoma, the risk in either breast being fairly
equal. In the series from Haagensen and associates [ref: 308] of 211
untreated cases of LCIS with a mean follow-up of 14 years, the overall
probability of developing invasive cancer in the ipsilateral breast
was 10%, and in the contralateral breast, 9%. In patients monitored
for 16 to 25 years, the risk of invasive cancer increased to 22% in
the ipsilateral breast, compared with 15% in the contralateral breast.
These authors recommended close monitoring for patients with LCIS
because of the equal risk of cancer in both breasts and the long
interval to the development of invasive cancer.

Confirming these findings, Rosen and colleagues [ref: 619] reported on
84 patients with LCIS with an average follow-up of 24 years after
biopsy only. The incidence of subsequent invasive carcinoma was 14% in
the ipsilateral breast, 14% in the contralateral breast, and 8% for
bilateral cancers. Overall, 36% of patients later developed invasive
cancer. In contrast, in a study of 149 patients with carcinoma of the
breast in whom a diagnosis of in situ carcinoma was confirmed, among
those treated by local excision, ipsilateral cancers were subsequently
seen in 12% of patients with DCIS and 13% of patients with LCIS. [ref:
761] Contralateral metachronous invasive cancers were seen in 6% of
the DCIS patients and 3% of the LCIS patients. No lymph node
involvement was seen in any of these patients, either with
prophylactic dissection or in follow-up. Seventeen of 109 patients
(15%) with intraductal lesions treated with local excision had local
recurrence; in contrast, in 7 of 40 patients with LCIS treated by
local excision, 2 had recurrences and 1 additional patient developed
invasive tubular carcinoma in the ipsilateral breast. The authors
concluded that both in situ lesions had similar clinical courses, that
treatment (ranging from local excision to modified or radical
mastectomy) did not appear to affect prognosis, and that lymph node
dissection was not necessary.

In a review of 3327 breast carcinomas, 47 were diagnosed as LCIS.
[ref: 461] Lesions were an incidental mammographic finding. Among 33
patients observed after local excision of the tumor, 1 (3%) developed
invasive carcinoma in the contralateral breast.

Lee and colleagues, [ref: 433] in a review of 419 women, found
bilateral lobular breast cancer in 36 patients (8.6%), with a
cumulative risk of 10% after 10 years. Prophylactic contralateral
surgical procedures were performed in 105 women and led to detection
of 70 in situ and 4 invasive cancers. Patients undergoing
contralateral prophylactic mastectomy had a better prognosis than
those with unilateral tumors not treated prophylactically. However,
there was no survival benefit, and the authors believe that
contralateral prophylactic mastectomy should not be recommended.

Page and associates [ref: 554] noted that the incidence of invasive
carcinoma after LCIS at 15 years was 17% and the recurrence rate was
8%. LCIS is considered a marker for malignancy rather than a
premalignant lesion per se; because this applies to both breasts, the
appearance of an invasive lesion can be a slow process (80% to 85%
disease-free survival at 15 years). Since many recurrences can be
successfully treated, observation is an acceptable form of management
in most patients. [ref: 140] Partial mastectomy has been used in the
treatment of these patients. [ref: 75,299,606] Recurrences (either
LCIS or infiltrating carcinoma) were observed in 1.5% to 21% of
patients, and long-term mortality after mastectomy for recurrence
ranged from 0% to 8.1%.

Temple and colleagues [ref: 761] reported on 109 patients with DCIS,
38 patients with LCIS, and 3 with both, who were treated with surgery
ranging from local excision to radical mastectomy. Among the patients
treated with local excision, ipsilateral breast recurrences were seen
in 12% of those with DCIS and 13% of those with LCIS. No lymph node
involvement was seen in any patient, either in node dissection or on
follow-up. Contralateral metachronous invasive breast cancers were
seen in 6% of those with DCIS and 3% of those with LCIS.

Singletary [ref: 696] described her experience in 45 women who had
LCIS alone and 30 women who had LCIS associated with an ipsilateral
stage I or II breast carcinoma but no abnormality in the opposite
breast. There was a 1% incidence of synchronous and 5% incidence of
metachronous nonlobular carcinoma in situ breast cancer in the
contralateral breast. Another breast cancer developed in a different
quadrant in 3 (16%) of 19 women with ipsilateral residual or intact
breast tissue. The author endorsed the recommendation of lifelong
careful observation of patients with LCIS. Surveillance should consist
of monthly self-examination, physical examination of the breast every
3 or 4 months, and annual mammography.

A metaanalysis of 13 reports concerning surgical management of LCIS
and 12 reports on DCIS after local excision showed recurrence rates of
17% with DCIS and 8.4% with LCIS. [ref: 75] The overall mortality
rates after mastectomy for patients with postexcision recurrence were
2.3% for DCIS and 2.8% for LCIS, not statistically different from the
rates for patients initially treated with mastectomy (1.7% for DCIS
and 0.9% for LCIS). Therefore, treatment options for LCIS are complete
local excision of the lesion and close follow-up, ipsilateral total
mastectomy with or without contralateral "mirror" biopsies, bilateral
mastectomies, or hormonal manipulation in investigational protocols.
[ref: 745] However, because 20% to 30% of women with LCIS later
develop invasive carcinoma in either breast, often after a great
number of years, excisional biopsy with close follow-up appears to be
the most reasonable approach. Nevertheless, Singletary [ref: 696] and
Morrow and associates [ref: 517] note that prophylactic bilateral
total mastectomies with breast reconstruction may be indicated in
highly anxious patients. Bilateral mastectomy for LCIS should be
considered only in women who have a strong family history of breast
cancer consistent with a hereditary pattern. [ref: 541]

Currently, there is no information regarding the use of breast
irradiation in LCIS.

Invasive Breast Tumors

Management of invasive breast cancer varies considerably among
institutions and physicians. However, treatment should be based on the
clinical extent and pathologic characteristics of the tumor, in
addition to the age of the patient (menopausal status), some biologic
prognostic factors, and the preference and psychological profile of
the individual patient. Patients with lesions smaller than 5 cm in
diameter and some specific characteristics to be discussed later
should be offered available options, with each modality thoroughly
discussed. In some states, legislation has been enacted requiring
treating physicians to comply with this practice.

For many years radical or modified radical mastectomy was considered
the standard therapy for operable patients. [ref: 216,306,785] Total
mastectomy combined with irradiation to the chest wall and regional
lymphatics has been used with comparable results. [ref: 229,380,491]

Alternatively, for more than 40 years, stage I and II tumors (T1, T2,
N0 or N1) have increasingly often been treated with local excision of
the tumor and breast irradiation. In recent years, several reports
have shown encouraging results with neoadjuvant chemotherapy, which
allows breast-conserving surgery and irradiation to be used in a large
proportion of these patients, because they exhibit substantial tumor
regression. [ref: 67,91,250,371] In response to a questionnaire mailed
to 2405 oncologists from all three disciplines and 60 oncology nurses
in the United States, Tannock and Belanger [ref: 752] noted that more
than 60% thought that modified radical mastectomy and conservation
surgery plus irradiation were equal options for patients with stage T1
and T2 carcinoma of the breast; 31% of the surgical oncologists
favored the former and 35% of the radiation oncologists favored the
latter approach. Medical oncologists were equally divided between the
two procedures (14% and 18%, respectively).

A modified radical mastectomy should be recommended, even for small
tumors, in patients with any of the following characteristics: (1)
patient preference -- cosmesis is not important, or desire to avoid 5
to 7 weeks of irradiation; (2) larger tumors in small breasts in which
a lumpectomy would remove so much tissue that the cosmetic outcome
would be severely compromised; (3) tumors with high risk for local
recurrence (e.g., EIC when negative margins are not attained); (4)
diffuse microcalcifications; (5) presence of skin or connective tissue
diseases that could be complicated by irradiation; and (6) patient
unreliable for follow-up.

Occasionally, after radical or modified radical mastectomy,
postoperative irradiation of the chest wall and peripheral lymphatics
is indicated in selected patients having high-risk characteristics,
regardless of the initial clinical stage or the administration of
adjuvant chemotherapy. [ref: 257,471] Although a few have claimed that
postoperative irradiation is harmful, [ref: 733] several analyses have
demonstrated that postoperative radiation therapy has no deleterious
effect and may even have a modest impact on overall survival. A study
in which adequate doses and fields of irradiation were used
demonstrated higher disease-free survival in patients treated with
irradiation before or after radical mastectomy compared those
undergoing mastectomy alone. [ref: 740] Postoperative irradiation is
discussed in detail in Chapter 51.

Premenopausal and postmenopausal patients who have poor prognosis are
treated with adjuvant chemotherapy. Bonadonna and associates [ref: 66]
described a beneficial effect in postmenopausal women who received at
least 85% of the prescribed doses of adjuvant chemotherapy. Levitt
[ref: 440] noted problems with interpretation of data in some major
clinical studies of adjuvant chemotherapy. Essentially, he concluded
that only premenopausal patients with fewer than three positive nodes
showed a benefit from this treatment.

Hormonal therapy has been extremely useful in many patients,
particularly those with positive ER or PR assay results. [ref: 453]
Procedures have included oophorectomy and adrenalectomy. Estrogens
have been used in premenopausal and postmenopausal patients as well.
Among women who are 4 years or more past menopause, estrogen
administration was successful in approximately 35% of patients.
Adrenal hormones, corticosteroids, progestins, and antiestrogens such
as tamoxifen are beneficial. [ref: 180,453,632]

Breast Conservation Therapy in Microinvasive Carcinoma

Microinvasive ductal carcinoma is defined as predominantly intraductal
carcinoma with microscopic or early invasion. Solin and co-workers
[ref: 717] analyzed 39 women with microinvasive ductal carcinoma of
the breast who were treated with breast-conserving surgery and
irradiation. Axillary lymph node staging showed 37 patients with
pathologically negative nodes and only 2 with positive nodes (each
with one positive lymph node). With median follow-up of 55 months, the
5-year actuarial overall and cause-specific survival rates were 97%.
Nine patients developed a recurrence in the breast; of 8 patients with
local-only first failure, 7 were salvaged with further treatment and
remained free of disease at last follow-up; 1 patient died of
subsequent distant metastasis. Median follow-up after salvage
treatment was 29 months (range, 0 to 54 months).

Bilateral Carcinoma of the Breast

Among factors reported to be associated with increased risk of
bilateral breast carcinoma are younger age, [ref: 63,104,351,445]
family history of breast cancer, [ref: 63,103,351,528,673] lobular
carcinoma, [ref: 351,359,445] multicentric disease, [ref: 528]
histologic differentiation of the primary tumor, parity status, [ref:
501] and positive PR assay. [ref: 359]

Patients with bilateral carcinoma have been treated with total or
modified radical mastectomy. [ref: 395] Breast irradiation combined
with tumor excision is an acceptable alternative therapy for
appropriately selected women with bilateral carcinoma of the breast.
Solin and associates [ref: 714] reported on 30 women receiving
radiation therapy after breast conservation surgery (11 with
concurrent and 19 with bilateral carcinoma). A dose of 45 to 50 Gy was
delivered to both breasts with tangential fields, in addition to a
boost of 10 to 15 Gy with either iridium implant or electrons. The
tangential fields were matched in the midline in 17 patients and
overlapped by up to 3 cm in 10 patients. Adjuvant chemotherapy was
given to 10 patients. The 5-year disease-free survival rate after
treatment of the first breast cancer was 79%, and the overall survival
rate was 72%. In the 60 treated breasts, the 5-year actuarial local
failure rate was 6%. In 25 treated breasts with a minimum of 2 years
of follow-up, 68% had excellent and 24% had good cosmetic results. The
incidence of arm edema was 6%, similar to that reported in patients
with unilateral disease.

In a review of 26 patients with stage I or II synchronous bilateral
breast cancer treated with breast-conserving therapy, the 5-year
actuarial cause-specific survival rate was 92%, and the disease-free
survival rate was 68%. [ref: 295] The incidence of local recurrence as
first failure site was 15% in the right breast and 4% in the left
breast. Distant metastases were the first site of failure in 12%.

Ninety-five patients with bilateral carcinoma of the breast treated
with mastectomy (60 patients), conservation of the breast (17
patients), or both (18 patients) were studied by Gustafsson and
associates. [ref: 303] Cumulative 5-year local tumor control rates
were 94% for the 138 mastectomy patients and 90% for the 52 patients
treated with breast conservation therapy. Distant metastases appeared
in one third of the patients; 50% occurred more than 60 months after
treatment. The 5-year distant disease-free survival rate from
treatment for the second carcinoma was 74%. Twenty-eight percent of
the first carcinomas were stage I, compared with 43% of second
carcinomas (P < 0.05); this result probably reflects the close
follow-up after initial treatment. The interval between the diagnosis
of first and second carcinomas had a profound effect on outcome; the
5-year distant recurrence-free survival rate when second carcinomas
were diagnosed within 5 years was 58%, compared with 95% for patients
diagnosed more than 5 years after the first carcinoma.

De la Rouchefordiere and co-workers [ref: 151] reported on 149
patients with simultaneous bilateral breast cancer (diagnosed within 6
months); of 298 tumors, 40% were T0 or T1, 45% were T2, and 15% were
T3 or T4. The majority (83%) were clinically node-negative. Treatments
were bilateral mastectomy in 43%, irradiation in 16%, and both in 41%
of the patients. Fifty-one patients had bilateral breast-conserving
therapy, and 24 were treated exclusively with irradiation. The 5-year
disease-free survival rates were 70% to 86%, similar to those observed
at the same institution in patients with unilateral tumors. Cosmesis
was assessed in 48 patients and was acceptable in 37 (77%). Special
attention should be paid to any possible overlap of the
supraclavicular and internal mammary fields over the spinal cord; one
patient developed spinal cord myelopathy at T6.

Paget's Disease of the Breast

This rare form of breast cancer comprises 1% to 4% of all breast
tumors. [ref: 23,251] Paget's disease of the breast is a carcinoma
that initially grows within the milk sinuses of the nipple, extending
into the ducts beneath the nipple in an apparently intraductal but not
necessarily in situ manner. [ref: 665] This disease may be
multicentric within the breast and may be associated with axillary
node metastases, although this is rare. Mammograms may demonstrate the
retroareolar spread of Paget's disease. Occasionally, calcifications
in a branching distribution may outline the intraductal spread. [ref:
667] Forty percent of patients have no evidence of a breast mass on
clinical or radiographic examination.

Many of these patients have been treated with mastectomy with
favorable results. [ref: 465,522] The role of limited surgery in
management of these patients has not been established; Lagios and
associates [ref: 424] noted one recurrence among five selected
patients treated with limited surgery only, and Paone and Baker [ref:
555] observed no recurrences in five patients treated with nipple
excision and wedge resection of the underlying breast.

Some patients have been treated with irradiation alone, without even
excision of the nipple-areola complex. This choice is feasible only if
the biopsy of the nipple indicates a relatively confined tumor. [ref:
667]

A few reports have been published on conservation surgery and
irradiation. Rissanen and Holsti [ref: 609] described results in 12
patients with Paget's disease of the breast without palpable mass
(minimum follow-up, 5 years); 9 patients had conservative treatment
with limited surgery, and 8 had radiation therapy. Radiation therapy
was delivered with 200-kV x-rays, and doses were lower than actual
standards. Tumor recurred in 3 patients, and 1 died of breast cancer.
Fourquet and colleagues [ref: 251] reported on 20 patients treated
conservatively with radiation therapy alone (17 patients) or limited
surgery and irradiation (3 patients). Most patients received 50 to 55
Gy to the breast and axillary lymph nodes and 40 to 45 Gy to the
external mammary and supraclavicular lymph nodes. The 7-year actuarial
disease-free survival rate was 81%, and the overall survival rate was
93%. Three patients had recurrence in the treated breast (nipple or
areola). One patient was treated with modified radical mastectomy and
the other two with simple mastectomy; all three patients were alive
with disease at 28, 77, and 98 months after salvage surgery. No
axillary failures were observed. Cosmetic results were satisfactory in
approximately 70% of patients. Although Fourquet and associates
irradiated the axilla, at the present time this is not recommended
because of the low incidence of axillary lymph nodes. [ref: 522]

Pierce and colleagues [ref: 574] reviewed 30 patients with Paget's
disease treated at several institutions with breast conservation
therapy (50 Gy) with a boost (14 Gy). None of the patients who
underwent axillary dissection had positive lymph nodes. With a median
follow-up of 62 months, three women developed a recurrence in the
treated breast, and two additional patients also failed in the breast
as a component of first failure. The 8-year local-only failure rate
was 16%; as a component of any failure, it was 23%. Four of five
patients were salvaged by mastectomy. The 8-year disease-free survival
rate was 95%.

Results of breast-conserving treatment for Paget's disease are shown
in Table 50-18.

Cystosarcoma Phyllodes

These rare fibroepithelial tumors account for fewer than 0.5% of
breast neoplasms in women. [ref: 605] Diagnostic workup is similar to
that used for breast masses.

Liberman and associates [ref: 446] reported mammographic findings in
46 women with 51 phyllodes tumors (32 benign and 19 malignant), 30 of
whom were also evaluated with preoperative sonography. On mammogram,
nonspiculated soft tissue masses were seen in 49 cases (96%); masses
were solitary in 48 and multiple in 1 case of recurrent benign
phyllodes tumor. Tumors 3 cm or larger had a greater relative risk
(3.87) of being malignant; six tumors larger than 8 cm were malignant.
Only 4 of 49 lesions (8%) contained calcifications visible on the
mammogram; 3 of them were benign, and 1 was malignant. Sonography
revealed cystic areas in 4 (44%) of 9 malignant tumors and 3 (14%) of
21 benign lesions (P = 0.15). In five tumors, doubling time was
calculated, and the median doubling time was 116 days (36 days for
malignant tumors, 105 to 638 for benign tumors).

Treatment for cystosarcoma phyllodes is either mastectomy or generous
wide local excision, depending on the degree of malignancy or size of
the lesion. Histologic grade appears to be the most important factor;
only about 10% of benign tumors recur. Size of the tumor has not been
shown to be a significant prognostic factor, but it may influence
surgical margins. However, the presence of underlying or clearly
malignant changes is associated with a higher recurrence rate and
distant metastases. [ref: 824]

Although few data support the use of radiation therapy in patients
with malignant cystosarcoma phyllodes, [ref: 824] from experience with
other types of histologic tumors in the breast it appears that
adjuvant radiation therapy may decrease the incidence of chest wall
recurrences, yet it may not have a significant impact on survival.
Patients with malignant cystosarcomas and positive or close surgical
margins and those who have local recurrence should be offered
radiation therapy to the breast or chest wall (50 Gy) to be followed
by a boost (10 to 15 Gy), depending on the presence of residual
microscopic or gross disease. Because of the low incidence of axillary
lymph node metastases, we do not advocate irradiation of the regional
lymphatics.

Cohn-Cedermark and associates [ref: 122] studied 77 patients with
primary cystosarcoma phyllodes of the breast; median age was 50 years.
Local recurrence was more common among patients treated with local
excision alone (13 of 24, 54%) than among patients treated with
mastectomy (2 of 53, 4%) (P < 0.001). However, there was no
significant difference between these two groups in terms of distant
metastasis-free survival (13% versus 25%, respectively). Sixteen
patients (21%) had distant metastases and subsequently died of
cystosarcoma phyllodes. With median follow-up of 8 years, 43 patients
(56%) were alive without disease, 6 (21%) were dead with disease, and
18 (23%) were dead with no evidence of disease. In a Cox multivariate
analysis, the only independent prognostic factors were tumor necrosis
(P < 0.05) and presence of stromal elements other than fibromyxoid
tissue (P < 0.01).

In a review of 170 cases, 92 (54%) were found to be benign, 59 (34.7%)
malignant, and 19 (11.2%) borderline. [ref: 605] Benign and borderline
tumors were usually treated with wide local excision, whereas the
malignant tumors were treated with simple (44%), modified (15%), or
radical mastectomy (34%). Mastectomy was recommended for recurrences
of borderline tumors and initially for malignant tumors. Lymph node
dissection was performed in 49% of the 59 patients with malignant
tumors; only 1 had node metastases. Postmastectomy irradiation to the
chest wall (50 Gy in 25 fractions) was administered in 4 patients with
infiltration of the pectoralis major muscle. Local recurrence was
found in 4 (4.3%) of 92 patients with benign tumors, 3 (15.8%) of 19
with borderline lesions, and 7 (11.9%) of 59 patients with malignant
tumors. With a mean follow-up of 8 years, the 5-year disease-free
survival rates in patients with malignant tumors were 65.5% (36 of 55)
after mastectomy and 75% (3 of 4) after wide local excision.

Angiosarcoma of the Breast

Angiosarcoma is a rare tumor comprising 3% to 9% of all breast
sarcomas; it accounts for about 1% of all breast primary malignancies.
[ref: 96] These tumors tend to be large, soft, and ill defined,
frequently with focal areas of hemorrhagic necrosis; they are somewhat
difficult to diagnose histologically because of the normal appearance
of some of the blood vessels. [ref: 106] In a report by Chen and
co-workers, [ref: 106] 37% of angiosarcomas were misinterpreted as
benign tumors at initial biopsy. The tumors are anaplastic and are
composed of unevenly dilated vascular channels lined by flat
endothelial cells. Some authors have correlated survival with the
degree of histologic tumor differentiation.

Simple excision of the tumor is frequently followed by local
recurrence [ref: 106]; therefore, total mastectomy or generous wide
local excision and radiation therapy are the recommended methods of
treatment. More radical procedures may be required for larger tumors,
particularly those invading adjacent structures. Lymph node metastases
are uncommon; according to Bundred and associates, [ref: 84] axillary
node sampling or dissection appears to be unnecessary because the
primary mode of tumor spread is hematogenous. Sellke and colleagues
[ref: 677] stated that, of 164 cases reported in the literature, 21
patients survived longer than 5 years and 8 survived longer than 10
years without evidence of disease.

Radiation therapy has occasionally been used in treatment of these
patients, but no specific reports are available. If a patient requires
irradiation for this lesion, we recommend doses and techniques similar
to those outlined for the treatment of breast carcinoma. Several
authors, including Donnell and associates [ref: 171] and Rosner, [ref:
622] have reported better survival with adjuvant chemotherapy. Antman
and colleagues [ref: 16] described five patients with angiosarcoma
treated with a combination of surgery, post-operative irradiation, and
adjuvant chemotherapy (doxorubicin alone or cyclophosphamide combined
with dacarbazine), some of whom survived 20 to 38 months after
therapy. Bundred and associates [ref: 84] believe that adjuvant
chemotherapy cannot be justified even in women with poorly
differentiated angiosarcoma of the breast, because the behavior of the
tumor is so unpredictable.

Adenocarcinoma in Axillary Lymphadenopathy Without Detectable Breast
Primary Cancer (Stage T0N1b)

The radiation oncologist is sometimes faced with the puzzling clinical
presentation of isolated axillary lymphadenopathy with adenocarcinoma,
with no clinical or radiologic evidence of a primary tumor in the
breast or any other anatomic site. [ref: 816] In addition to a careful
physical examination including the breast, bilateral mammograms and
chest x-ray films should be obtained. An exhaustive radiographic
workup, including CT scan of the chest, upper gastrointestinal
studies, barium enema, and intravenous pyelogram, is not warranted.
[ref: 97] Often, after careful examination of mastectomy specimens,
the primary tumor in the breast cannot be demonstrated (Table 50-19).
These patients have a relatively good prognosis, with survival rates
of 50% to 80% at 5 and 10 years. [ref: 212,550]

Whereas some authors advise mastectomy and axillary dissection, [ref:
24] an alternative is irradiation of the breast and the regional
lymphatics. Although a primary tumor often is not identified on
mammograms and boost doses cannot be delivered to the breast in
general, the breast is treated with doses of 50 Gy and the
axillary/supraclavicular lymph nodes with 50 Gy, with a boost of 10 to
15 Gy to the axillary fossa, where the initial lymph nodes were
palpated. Because of the excellent prognosis of these patients and the
infrequency of distant metastases, adjuvant chemotherapy may not be
warranted. [ref: 816]

Campana and associates [ref: 97] reported on 31 patients, some of whom
were previously included in publications by Vilcoq and colleagues.
[ref: 816] Initial treatment consisted of resection of the involved
lymph nodes followed by irradiation in 14 patients, axillary
dissection and radiation therapy in 8 patients, radiation therapy
followed by axillary dissection in 6 patients, irradiation and
modified radical mastectomy in 1 patient, and irradiation alone in 2
patients. The radiation therapy techniques included breast irradiation
in 30 patients using standard techniques; dose delivered to the
midpoint of the breast was 62 Gy. All 31 patients received radiation
therapy to the regional lymphatics (axillary, supraclavicular, and
internal mammary chains) to doses of 60 Gy or higher; 10 patients
received adjuvant chemotherapy. All 31 patients were monitored, with a
median follow-up of 9 years. By the time of the report, 8 locoregional
recurrences had been detected; all were histologically proven
infiltrating ductal carcinoma. Four of the 8 patients had isolated
breast recurrences at 63, 90, 136, and 162 months after treatment; 3
of these 4 patients were salvaged by modified radical mastectomy, and
another was given further irradiation for a recurrence in an initially
underdosed region. Three of 8 patients had breast recurrences
associated with axillary relapse, and another developed low cervical
lymph node recurrence. Two contralateral breast tumors occurred 1 and
9 years after treatment of the ipsilateral lesion. The overall 5- and
10-year survival rates were 76% and 71%, respectively. The risks of
locoregional recurrence at 5 and 10 years were 14% and 25%,
respectively, and the risks of developing distant metastases were
23.5% and 29%, respectively. Cosmetic results were evaluable in 24
patients; cosmesis was reported to be good in 19 (75%) and fair in 5
patients. Six patients had limitation of shoulder movement, 2
experienced arm edema, and 2 had asymptomatic focal pulmonary
fibrosis.

Baron and co-workers [ref: 38] described results in 35 patients with
axillary metastasis without evidence of a breast cancer, including
mammographic evaluation. Twenty-eight patients were treated with
mastectomy, and 7 were managed by limited resection, axillary
dissection and irradiation, or both. Twenty-two (67%) of 33 breast
specimens contained carcinoma (invasive in 18, in situ in 4). The
actuarial 5-year survival rate was 77% after mastectomy and 65% after
breast-preserving surgery and irradiation. Five patients had
contralateral breast cancer treated either before or after therapy for
the axillary carcinoma with occult breast tumors. The 5-year survival
rate was comparable for patients receiving or not receiving adjuvant
chemotherapy (79% and 77%, respectively).

Sixty cases of axillary histologic metastasis were analyzed by Merson
and associates. [ref: 496] The number of pathologic metastatic nodes
was one in 13 patients, two to three in 10 patients, and four or more
in 23 patients; the number of metastatic nodes was not evaluable in 14
cases. Extranodal invasion was seen in 92% of cases. Thirty-three
patients underwent breast surgery at the time of histologic diagnosis
of the axillary metastases; 6 patients were treated with radiation
therapy to the breast, and 17 patients did not receive any immediate
treatment to the breast (9 of these later developed a primary breast
carcinoma). Thirty-seven of 60 patients received adjuvant therapy (29
chemotherapy, 8 tamoxifen). The 5- and 10-year survival rates were 77%
and 58%, respectively. Survival was equivalent for patients treated
with immediate surgery plus radiation therapy and for patients who
were observed without treatment to the breast. Adjuvant treatment did
not improve prognosis.

general management

General Management

Since the beginning of this century, based on Halsted's hypothesis of
tumor spread, the prevalent surgical treatment for carcinoma of the
breast was radical mastectomy (with several modifications). [ref: 216]
Halsted [ref: 319] postulated that breast cancer spreads by direct
extension rather than by tumor cell embolization; based on this
premise, it was incorrectly thought that the wider the surgical
extirpation, the greater the chance for cure. McWhirter [ref: 491]
popularized a lesser surgical procedure (total mastectomy) in
combination with irradiation to the chest wall and regional
lymphatics, a technique that yielded results comparable to those of
radical mastectomy. [ref: 220,380]

Another treatment of breast cancer, initially described by Keynes
[ref: 393] in 1929 and 1937, combined conservation surgery (ranging
from biopsy to wide local tumor excision to segmental mastectomy or
quadrantectomy) and definitive irradiation. [ref: 9,95,330,564,817]
This approach, popular in Europe since 1950, has progressively gained
acceptance in the United States in the past 20 years and will be
extensively discussed in this chapter.

Patients with positive axillary lymph nodes and selected high-risk
node-negative patients are treated with adjuvant chemotherapy.
Patients with positive hormonal receptors receive tamoxifen.

Von Rueden and Sessions [ref: 821] treated 32 women over the age of 65
who had invasive breast cancer with quadrantectomy and tamoxifen as
the sole treatment (no axillary dissection, irradiation, or
chemotherapy) and compared them with 110 women of similar age treated
with standard therapy (total mastectomy or segmental resection with
axillary dissection and breast irradiation). With a mean follow-up of
52 months, disease-free survival in the patients treated with breast
conservation therapy (quadrantectomy) and tamoxifen was 92%, and
overall survival was 67%, with no patient developing a recurrence. In
the 110 women treated with standard therapy, disease-free survival was
83%, overall survival was 82%, and 5% developed local recurrence.
Thus, selected groups of older women may be treated with less
aggressive therapy with satisfactory results. [ref: 515]

Ductal Carcinoma in Situ and Lobular Carcinoma in Situ

Ernster and colleagues [ref: 193] observed a substantial increase in
diagnosis of DCIS in the United States (200% higher in 1992 than in
1983). During the same years there was a marked decline in the
proportion of these patients treated with mastectomy (from 71% to
43.8%) and an increase in the proportion treated with lumpectomy (from
25.6% to 53.3%). In 1992, 30% of patients were treated with lumpectomy
alone and 23.3% with lumpectomy and breast irradiation. Nevertheless,
the number of patients treated by mastectomy may be inappropriately
high (> 50% in many stages).

Autopsy studies show that up to 16% of asymptomatic women have DCIS.
[ref: 525] Two series reported a 15% to 20% incidence of intraductal
carcinoma in women undergoing mammography for screening. [ref: 33,798]

As defined by Broders, [ref: 79] the essential histologic feature of
DCIS is confinement of malignant epithelial ductal cells within their
natural basement membrane. Frykberg and Bland [ref: 274] describe four
distinct histologic DCIS patterns, probably representing an evolution
toward invasive carcinoma: papillary, cribriform, solid, and classic
comedo carcinoma. A micropapillary pattern has been described.

Estrogen receptors have been documented in 60% of cases of DCIS, the
same incidence as in invasive breast cancer. [ref: 274]

Natural History

Despite fragmented studies in a small number of patients, with
selection bias and inclusion of some invasive lesions, a few long-term
follow-up reports document a 30% to 50% risk of subsequent development
of invasive breast cancer after untreated DCIS that generally develops
within 10 years (Table 50-12). Patients with LCIS also have a
propensity to develop invasive lesions (35% to 45% in 10 to 20 years).
[ref: 55,202,241,423,552,617] Microscopic study of breast specimens in
patients with DCIS treated by mastectomy have shown residual
noninvasive DCIS in the original biopsy site in up to 77% of cases.
[ref: 621,855]

Patients with DCIS who present with a large mass (> 2.5 cm) have
significantly higher potential for occult invasion, multicentricity,
axillary lymph node metastases, local recurrence, and ultimately worse
survival than patients who have nonpalpable microscopic lesions at
diagnosis of DCIS. Multicentricity (occult malignancies located
outside the quadrant of the primary tumor) should be differentiated
from multifocality (malignant foci within the same quadrant as the
primary tumor or residual disease). [ref: 241] Rates of
multicentricity in patients with DCIS range from 15% to 78% (average,
35%). [ref: 356,668,770,822] Fisher and associates [ref: 241] reported
multicentricity in 27 (39.7%) of 68 patients with intraductal
carcinoma treated with total mastectomy in NSABP Protocol B-06.

In serial analysis of 82 mastectomy specimens (5-mm whole-organ
sections), Holland and co-workers [ref: 356] found that DCIS involved
one breast quadrant in 66%, extended over more than one quadrant in
23%, and was centrally located in 11% of patients. The nipple or
subareolar area was more frequently involved by comedo carcinoma (32
of 50, 64%) than by micropapillary and cribriform (11 of 32, 34%).
Nine patients had manifestations of Paget's disease.

Bellamy and associates, [ref: 50] in a review of 130 specimens,
observed more frequent single-quadrant involvement with the comedo,
solid, or cribriform subtypes of DCIS than with micropapillary DCIS,
which tended to have more diffuse involvement.

Schwartz and colleagues, [ref: 671] in an analysis of 50 breasts with
nonpalpable DCIS, noted that the solid and cribriform patterns were
rarely multicentric or microinvasive, whereas papillary and
micropapillary patterns were often multicentric and more diffuse but
rarely microinvasive; others have pointed out similar findings. [ref:
333] Comedo carcinoma was more likely to be both microinvasive and
multicentric. Comedo carcinomas also have a high incidence of
HER-2/neu-protein overexpression [ref: 790] and a high proliferative
index on thymidine-labeling studies. [ref: 498] Analysis of the
pathologic data in NSABP B-17 protocol showed that the only predictors
for ipsilateral breast recurrence were comedo-type necrosis and
involved or uncertain margins. [ref: 237] Patchefsky and colleagues
[ref: 557] also found that papillary and cribriform tumors were less
frequently associated with microinvasion.

Clinical significance of multicentricity or bilaterality in breast
cancer is controversial. As many as 96% of all local breast
recurrences after DCIS are found in the same quadrant as the primary
lesion, implicating residual disease rather than multicentricity,
[ref: 241,254,422,552] and this feature has not been shown to have an
impact on survival. [ref: 423,621] Synchronous occurrence of two or
more clinically evident breast malignancies in one patient is
extremely rare (approximately 0.1%). [ref: 238]

Clinical Presentation and Diagnosis

DCIS may present as a palpable mass, although more frequently these
lesions are diagnosed on screening mammography in a nonpalpable stage.
Microcalcifications are seen in up to 34% of cases. [ref: 293,668,678]
Almost 95% of all in situ breast cancers detected by mammography are
associated with microcalcifications. In addition, 90% of all
carcinomas that present as mammographic calcifications are in situ,
and 80% are DCIS. [ref: 293,678]

Patchefsky and colleagues [ref: 557] showed that microcalcifications
are more frequent in comedo carcinoma and may be associated with
microinvasion in 63% of cases. Holland and associates [ref: 356] noted
that mammography did not adequately predict extent of micropapillary
and cribriform tumors, with histologic extent more than 2 cm beyond
that predicted on the mammogram in 47% of cases. Microcalcifications
were present in 40 (80%) of 50 cases of predominantly comedo DCIS,
compared with only 5 (16%) of 32 micropapillary or cribriform tumors.

Ikeda and Andersson, [ref: 367] in a study of 190 women with
biopsy-proven DCIS, found 117 patients (62%) who showed suspicious
clustered microcalcifications on mammograms; 30 (16%) had negative
mammograms. Forty-three patients (23%) showed other abnormalities: 15
had circumscribed masses, 12 had focal nodular patterns, and the
remaining 16 patients showed dilated retroareolar ducts, ill-defined
round tumors, subareolar mass, or focal architectural distortion.

Treatment of DCIS

The optimal treatment of patients with DCIS must be individualized,
based on the natural history of the disease, tumor extent, histologic
features, and patient preference. Important prognostic factors to be
considered in therapeutic decisions include tumor size, pathologic
subtype, nuclear grade, presence of necrosis, extent of microscopic
tumor, and status of surgical margins. [ref: 237,422,687,722]

Silverstein and colleagues, [ref: 687] in an analysis of 181 women,
noted that larger tumor size (>/= 2.5 cm) and comedo carcinoma subtype
carried a greater tendency toward positive initial histologic margins
and residual intraductal breast carcinoma, although this trend was not
statistically significant (P < 0.1).

The incidence of residual DCIS after biopsy is high, an important
consideration if local excision only is advocated for treatment. Rosen
and associates [ref: 621] observed a 57% incidence of residual DCIS
after biopsy in 53 women treated with mastectomy. Wanebo and
co-workers [ref: 831] and Carter and Smith [ref: 100] described
residual DCIS after biopsy in 60% and 66% of mastectomy specimens,
respectively. However, it is uncertain whether the surgeons had
attempted to obtain negative margins at the time of biopsy.

In several reports of patients treated with biopsy only, recurrences
ranged from 8% in selected patients [ref: 423] to 75% [ref: 404,444]
(Table 50-13).

Updated NSABP results in 48 patients treated with lumpectomy
(follow-up, 83 months) showed breast relapses in only 2 (7%) of 27
patients receiving irradiation, compared with 9 (43%) of 21 patients
treated with lumpectomy alone, and none in 27 patients treated with
mastectomy. [ref: 240]

Page and colleagues [ref: 552] reported that 7 (28%) of 25 women
monitored for longer than 3 years after biopsy only of DCIS of the
breast developed invasive lesions during a follow-up period of
approximately 15 years.

On the other hand, Lagios [ref: 420] reported a 12.6% incidence of
local recurrence in 79 selected patients with DCIS treated with local
excision alone. Arnesson and associates [ref: 19] in Sweden observed a
13% recurrence rate in 38 women treated with local excision alone
(mean follow-up, 60 months). Schwartz and colleagues [ref: 670]
treated 70 patients with DCIS with local excision and surveillance
alone. With a mean follow-up of 49 months (maximum, 168 months), 11
patients (15.3%) developed a recurrence (8 to 85 months after initial
diagnosis). Ten of the 11 patients with recurrent tumor had comedo
carcinoma initially. Half of the recurrences in Lagios' series were
invasive, as frequently reported, compared with only 2 of 11 in the
Schwartz report.

More effective treatment options for DCIS include total mastectomy or,
more recently, wide tumor excision followed by breast irradiation.
Surgeons should make every attempt to excise the entire lesion with
clear histologic margins to lower the incidence of posttreatment
recurrence.

As in invasive carcinoma, Coleman and associates [ref: 124] showed
regional variations in the use of breast-conserving therapy for DCIS:
54% in Seattle and San Francisco, 48% in Connecticut, 26% in Iowa, 29%
in Atlanta, and 31% in Hawaii. Winchester and associates, [ref: 852]
in a review of the data from the National Cancer Data Base, noted an
increased use of partial mastectomy without lymph node dissection from
12.6% of patients in 1985 to 23.2% in 1991 in the treatment of
carcinoma in situ of the breast. The use of total and modified radical
mastectomy was essentially unchanged (5.8% to 8.7% and 41.8% to 42%,
respectively).

Routine node dissection has been eliminated for DCIS because so few
patients have positive nodes. A possible exception is the large or
extensive intraductal cancer (e.g., >/= 4 cm), which is known to have
a small but finite incidence of axillary spread, probably the result
of pathologic sampling error (undetected invasive carcinoma).
Silverstein and associates [ref: 688] analyzed axillary node
positivity, disease-free survival, and breast cancer-specific survival
in six breast cancer subgroups by tumor stage (T) category. Nodal
positivity for DCIS was 0% in 189 patients. In multiple series
totaling more than 1000 patients, the incidence of positive axillary
lymph nodes was less than 1% (Table 50-14). [ref: 21,100,241,395,692]
Therefore, only the breast needs to be addressed when considering
treatment, whether by total mastectomy or tumor excision with or
without irradiation. Results described with mastectomy are shown in
Table 50-15. Occasionally chest wall recurrence of DCIS has been
reported after mastectomy. [ref: 148,234]

Several reports have described satisfactory outcome in women treated
with local excision and irradiation for intraductal cancers. Kuske and
colleagues [ref: 417] described results in 76 women with 77 DCIS
treated with breast conservation surgery and irradiation. With median
follow-up of 4 years, overall 5-year survival was 99%, disease-free
survival 89%, and the breast tumor-control rate 93%. Comedocarcinoma
had a 5-year actuarial tumor control rate of 88% in 70 irradiated
patients compared with 98% for all other histologic subtypes of DCIS
(P < 0.03). All six patients with local failure were successfully
salvaged by further surgery.

Fourquet and associates [ref: 253] treated 67 patients with DCIS with
breast-conserving therapy. With a follow-up of 104 months, the 10-year
breast recurrence rate was 10.4%. Five of seven recurrences were
invasive ductal carcinoma. The only factor associated with risk of
local recurrence was total dose of irradiation to the tumor bed; all 7
recurrences occurred in 41 patients who received less than 63 Gy,
whereas 26 patients treated with higher doses had no recurrences (P <
0.02).

Solin and co-workers [ref: 720] reported 261 breasts with DCIS in 259
women from nine institutions in Europe and the United States treated
with local excision and irradiation. The 10-year actuarial rate of
local failure was 16%, and of distant metastasis, 4% (Figure 50-9).
The authors updated the results with 15-year follow-up. [ref: 718] The
cause-specific survival rate was 96%, the breast relapse rate was 19%,
and the incidence of distant metastasis was 3%. Although the comedo
carcinoma plus nuclear grade 3 lesions had a greater incidence of
local failure in the earlier years, by 10 years there was no
significant difference from other histologic subtypes (18% and 15%,
respectively). Of 45 patients with local recurrence, 24 (53%) had
invasive ductal carcinoma, and 21 (47%) had DCIS (one associated with
Paget's disease). Treatment of local recurrence consisted of
mastectomy in 42 patients, local excision and axillary dissection in
1, local excision alone in 1, and other treatment in 1 patient with
associated distant metastases. Of 26 patients who had assessable lymph
nodes, one had positive axillary lymph nodes.

Solin and colleagues [ref: 717] also described results of a large
multiinstitutional study of 110 women with unilateral nonpalpable
mammographically detected DCIS of the breast who were treated with
breast-conserving surgery and irradiation. With a median follow-up of
9.3 years, the local recurrence rates were 7% (3 of 42) in patients
with negative final pathologic margins, 29% (5 of 17) in those with
positive or close margins, and 14% (7 of 51) in those with unknown
final margins. Local recurrence developed in 14 (25%) of 56 women
younger than 50 years of age but only in 1 (2%) of 54 older women;
median interval to local recurrence was 4.9 years in the younger group
and 8.7 years in the older group. There was no significant difference
in 5-year local recurrence according to histologic subtype (8% for
comedo and 2% for other). The 10-year actuarial cause-specific
survival rate was 96%. A subgroup of 21 patients was identified with
characteristics similar to those detailed by Lagios and colleagues
[ref: 422] (mammographic detection with microcalcifications alone,
pathologically confirmed negative margins of excision, tumor size <
2.5 cm); these patients were treated with breast-conserving surgery
(local excision) and irradiation, and there were no local recurrences
during a median follow-up period of 8.7 years.

Fowble and associates, [ref: 258] in 110 women with mammographically
detected DCIS treated with breast conservation therapy (50 Gy) with
boost to the primary site (10 Gy), reported a 10-year actuarial breast
recurrence rate of 15% and a cause-specific survival rate of 100%.
Three ipsilateral recurrences were invasive ductal carcinoma. Two
patients developed contralateral invasive breast cancer. Vicini and
colleagues, [ref: 812] in 104 breasts in 102 women treated in a
similar manner with median follow-up of 6.1 years, noted a 10-year
actuarial tumor control rate of 94%. Three recurrences were invasive,
and two were DCIS. All patients failing in both series were treated
with mastectomy.

Treatment of DCIS (Continued)
(1 of 1)

In 172 women with DCIS who were treated with breast-conserving surgery
and definitive breast irradiation at multiple institutions in Europe
and the United States, the only pathologic parameter that correlated
with the rate of local recurrence was the presence or absence of the
combination of comedo carcinoma plus nuclear grade 3; the 8-year
actuarial rate of local recurrence was 20% with this combination
versus 5% for other lesions (P = 0.009 on univariate analysis; P =
0.017 on multivariate analysis). [ref: 722] None of the pathologic
parameters evaluated correlated with overall survival (all P >/=
0.16), cause-specific survival (all P >/= 0.13), or freedom from
distant metastasis (all P >/= 0.13).

Silverstein and co-workers [ref: 686] reported on 227 selectively
treated patients with DCIS without microinvasion. The 98 patients with
least favorable lesions (large tumors with involved biopsy margins)
were treated with mastectomy, and the 129 with most favorable lesions
(small tumors with clear margins) were treated with breast
preservation. The preservation group was further subdivided into those
who received excision plus irradiation and those received excision
alone. One local invasive recurrence and no deaths occurred in the
mastectomy group; the 7-year actuarial disease-free survival rate was
98%. In the excision and irradiation group (103 patients), 10 patients
(10%) had local recurrence (5 invasive and 5 noninvasive), and 1
patient died; the 7-year actuarial disease-free survival rate was 84%
(P = 0.038). In the excision-alone group (26 patients with an average
lesion size of 1 cm), there were two local recurrences (8%), one of
which was invasive; there were no deaths, but only 3 patients had been
followed for longer than 4 years at the time of the report. A total of
163 axillary node dissections were done; all were negative. Six (50%)
of 12 local recurrences in the conservatively treated patients were
invasive. There was no difference in overall survival in any subgroup
regardless of treatment.

In another study of 70 cases of DCIS, 34 were treated with modified
radical mastectomy; 36 were treated with either local excision (2
patients), lumpectomy (26 patients), or quadrantectomy (8 patients),
with complementary irradiation in 34 of the 36 cases (and boost in
32). [ref: 137] The main histologic subtype was comedo carcinoma (25
of 70 cases). One local relapse (3%) in the radical surgery group and
three (9%) in the conservation treatment group were noted at 55
months. The obvious factor influencing local recurrence was inadequate
surgical excision.

NSABP Protocol B-17 randomly assigned 391 women to be treated with
lumpectomy alone; 64 (16.4%) of this group developed ipsilateral
breast cancer, compared with 28 (7%) of 399 women treated with
lumpectomy and breast irradiation. [ref: 223] The 5-year cumulative
incidence of second cancers in the ipsilateral breast was reduced by
irradiation from 10.4% to 7.5% for noninvasive cancer and from 10.5%
to 2.9% for invasive lesions (P = 0.055 and P < 0.001, respectively)
(Fig. 50-10A). Among the 64 women who developed breast relapse after
lumpectomy alone, 32 had invasive and 32 had noninvasive carcinoma. Of
the 28 women treated with lumpectomy and breast irradiation who
developed a recurrence, 8 (28.6%) had invasive and 20 (71%) had
noninvasive carcinoma. The 5-year event-free survival rate was better
in women receiving breast irradiation (84.4%) than in those treated
with lumpectomy alone (73.8%) (P = 0.001; Fig. 50-10B). Among 63 women
whose primary DCIS was treated by lumpectomy only and in whom a second
breast tumor developed, 28 (43.8%) were treated by repeat lumpectomy
and 36 (56.3%) by mastectomy. The percent of women treated by second
local excision was similar regardless of whether the second tumor was
invasive (42.8% in both groups). Of 28 women receiving irradiation who
subsequently developed tumors in the ipsilateral breast, 10 of the 20
whose second tumors were noninvasive were treated with lumpectomy and
10 with mastectomy; all 8 women whose second tumors were invasive were
treated with mastectomy. Eight women in the lumpectomy group and 10 in
the lumpectomy plus irradiation group developed a contralateral breast
cancer as the first failure event (11 were invasive, 7 were DCIS).
Fourteen of 18 contralateral tumors were treated by lumpectomy, 3 by
mastectomy, and 1 (an inflammatory cancer) by systemic therapy and
mastectomy. Longer follow-up is needed before a final analysis of
these data can be carried out. Swain [ref: 746] pointed out that 42%
to 45% of tumors in Protocol B-17 were microscopic (< 0.1 cm), and
another 30% were 1 cm or less in diameter, so the results of this
trial cannot be extrapolated to larger tumors, which tend to be
multicentric and may have areas of invasion.

Patient selection in clinical trials was documented by Fentiman and
associates. [ref: 210] They identified 207 women with pure DCIS on
biopsy who were eligible for European Organization for Research on
Treatment of Cancer (EORTC) Trial 10853 (randomization to observation
or breast irradiation of 50 Gy), only 77 of whom (36%) were registered
for the study. Major reasons for nonentry were too-extensive DCIS (76
of 139, 55%), history of previous breast cancer (18%), a lump larger
than 3 cm (4%), and patient refusal (4%). These exclusions may make
the results of the trial applicable only to a minority of patients
with DCIS.

Silverstein and associates [ref: 691] published a review of 425
patients with DCIS who were treated with either mastectomy or
breast-conserving surgery and irradiation; among the latter group,
they noted a recurrence rate of 3.9% in patients who had
non-high-grade DCIS without necrosis, 11.1% in those who had
non-high-grade DCIS with necrosis, and 26.5% in those who had
high-grade DCIS. The 8-year actuarial disease-free survival rates were
93%, 84%, and 61%, respectively (P treated with excision and breast irradiation, the 10-year disease-free
survival rate was 75%, compared with 20% in 31 patients treated with
excision alone.

The reported incidence of local recurrence after treatment of DCIS
with conservation surgery and breast irradiation is higher but not
significantly different from that seen after mastectomy (Table 50-16).
More importantly, survival rates are comparable, and patients who have
recurrences after lumpectomy and irradiation can be salvaged with
mastectomy.

Solin and associates [ref: 711] analyzed 42 local failures in the
breast after initial treatment of DCIS with breast-conserving surgery
and irradiation in 274 women; 23 patients (55%) had invasive ductal
and 19 (45%) had intraductal carcinoma, 1 with associated Paget's
disease. Overall, 20 recurrences (48%) were detected by mammography
only, 11 (26%) by physical examination only, and 11 (26%) by
mammography and physical examination. Of the 19 intraductal
recurrences, 14 (74%) were detected by mammography only. Thirty
recurrences (71%) were at the initial tumor excision site; 10 (24%)
were elsewhere in the breast, and 2 (5%) were diffuse or multifocal.
Of the recurrences, 20 (48%) occurred within 5 years, 17 (40%) between
5 and 10 years, and 5 (12%) 10 years after treatment of the original
lesion. Surgical treatment at the time of local recurrence included
mastectomy (39 patients), excision (2 patients), and other (1
patient). Adjuvant systemic therapy at the time of local recurrence
included chemotherapy in 2 patients and hormonal treatment in 7; 1
patient received both; 32 patients received none. Median follow-up
after salvage treatment was 3.7 years (mean, 4 years; range, 0.1 to
9.5 years). The 5-year actuarial survival rate after salvage treatment
for the 42 patients with local recurrence was 78%, the rate of
cause-specific survival was 84%, freedom from distant metastases was
86%, and freedom from chest wall recurrence after salvage mastectomy
was 92%. All three patients who developed chest wall recurrence after
salvage mastectomy also developed distant metastatic disease.

Fisher and colleagues, [ref: 237] in analysis of 573 patients of 790
with DCIS enrolled in NSABP Protocol B-17, identified 53 patients who
had second ipsilateral breast tumors; 38 occurred in the group treated
with lumpectomy only and 15 in patients treated with lumpectomy and
irradiation. The overall incidence of breast relapse was 13.9% and 5%,
respectively, in the two treatment groups. Significant prognostic
factors in the patients treated with lumpectomy and irradiation were
moderate to marked comedo necrosis and indistinct or involved
lumpectomy margins.

In contrast to LCIS, with risk of contralateral carcinoma of the
breast with intraductal lesions is small. The rate of bilateral breast
cancer associated with DCIS is 10% to 15%, although some have reported
rates as high as 30%. [ref: 742] Most of these occult contralateral
malignancies are also carcinoma in situ. [ref: 268,582] Webber and
co-workers [ref: 834] showed a risk of contralateral breast cancer of
3.4% with an average follow-up of 9 years (4 of 116 patients). Kinne
and associates [ref: 395] reported on 25 patients with bilateral in
situ carcinoma of the breast; 8 patients developed metachronous
invasive carcinoma, including 2 with positive axillary lymph nodes,
probably the result of undetected invasive carcinoma.