50_01

Anatomy

Breast anatomy is described in great detail in numerous publications
[ref: 154,305] and is covered only briefly here.

The breast is made up of the mammary gland, fat, blood vessels,
nerves, and lymphatics (Fig. 50-1). [ref: 394] The surface of the
breast has deep attachments of fibrous septa, called Cooper's
ligament, which runs between the superficial fascia (attached to the
skin) and the deep fascia (covering the pectoralis major and other
muscles of the chest wall). [ref: 154]

The mammary gland lies over the pectoralis major muscle and extends
from the second to the sixth rib in the vertical plane and from the
sternum to the anterior or even midaxillary line. [ref: 154] An
additional layer of mammary tissue extends laterally into the axilla
to a variable degree. This axillary projection, often called the tail,
is sometimes quite prominent and should always be kept in mind during
examinations and in planning radiation therapy.

The retromammary bursa lies between the deep layer of the superficial
fascia and the deep fascia; it contains loose areolar tissue that
allows for mobility over the chest wall. [ref: 305] It is crossed by
projections of the deep layer of the superficial fascia that join with
the deep pectoral fascia to form the posterior suspensory ligaments of
the breast. Deep projections of mammary parenchyma may extend between
the muscle bundles of the pectoralis major muscle. The mamma consists
of glandular tissue arranged in multiple lobes composed of lobules
connected in ducts, areolar tissue, and blood vessels. The smallest
lobules consist of clusters of rounded alveoli that open into the
small branches of the lactiferous ducts (Fig. 50-2); these unite and
form larger ducts that eventually converge into single canals in the
nipple, corresponding to each lobe of the gland (15 to 20
galactophori). [ref: 398]

A network of lymphatics is formed over the entire surface of the
chest, neck, and abdomen and becomes dense under the areola. Mammary
gland lymphatics begin in the interlobular or prelobular spaces,
follow the ducts, and end in the subareolar network of lymphatics of
the skin. [ref: 154,305] The following lymphatic pathways originate
mostly in the base of the breast: (1) the axillary or principal
pathway passes from the upper and lower halves of the breast to the
lateral chain of nodes situated between the second and third
intercostal space; (2) the transpectoral pathway passes through the
pectoralis major muscle to the supraclavicular lymph nodes; and (3)
the internal mammary pathway passes through the midline, through the
pectoralis major and intercostal muscles, usually close to the
sternum, to the nodes of the internal mammary chain. The main
lymphatic channels of the breast are illustrated in Figure 50-3.

Epidemiology

In the United States in 1997, approximately 181,600 new cases of
breast cancer were diagnosed, and 43,900 women died of the disease.
[ref: 10] Breast and lung cancer are the foremost causes of cancer
death in women. Breast cancer mortality has been stable for the last
50 years; however, in recent years there has been a slight overall
decline in breast cancer age-adjusted mortality, in contrast to
pronounced peaks in incidence. Almost all of the increase (83%) is
accounted for by early diagnosis of in situ or invasive lesions less
than 2 cm in diameter with screening mammography. Reduced mortality is
a reflection of earlier diagnosis and more effective treatment,
including adjuvant chemotherapy, particularly in premenopausal
patients.

Although 1 of every 10 wo men in the United States is projected to
develop carcinoma of the breast, 80% to 92% of all breast masses are
benign. [ref: 452] Approximately 1% of breast cancers occur in men.

The risk factors for breast cancer in women are well documented: age
greater than 50 years, personal or family history of breast cancer
(e.g., mother, aunt), nulliparous, or first child after 30 years of
age. [ref: 850] Madigan and colleagues [ref: 462] identified 193 women
with breast cancer; the most important risk factors were age at first
birth and nulliparity, higher socioeconomic status, and family
history. Most of the increased incidence rates over the last two
decades have occurred among women aged 45 to 74 years, with a further
increase in black women.

Breast cancer is more common in Jewish than in non-Jewish women and in
black women compared with white women. Low incidence and mortality
rates for female breast cancer are found in most Asian and African
countries, intermediate rates in southern European and South American
countries, and high rates in North America and northern European
countries. [ref: 389] Among the possible explanations for this
variability are environmental factors, including diet.

Japanese women show lower rates of breast cancer than Caucasian women,
a difference accounted for by an increased incidence of this cancer in
postmenopausal Caucasian women. Postmenopausal breast cancer is also
less common among Japanese women who have migrated to a Western
country. After two or three generations, the incidence of female
breast cancer among descendants of Japanese immigrants to Hawaii or to
the mainland of the United States approached that of Caucasian
residents. [ref: 389]

Newcomb and associates [ref: 527] noted that relative breast cancer
risk among premenopausal women was reduced by a history of lactation
(relative risk [RR], 0.78) and by a cumulative lactation time
exceeding 24 months.

Several large-scale studies have failed to demonstrate a correlation
between the prolonged use of oral contraceptives and breast cancer.
[ref: 453,624,775,810] White and associates, [ref: 844] in an analysis
of 961 women born in 1944 or later who took oral contraceptives all of
their reproductive life, reported no increase in breast cancer
incidence. However, there was a small increased risk, particularly for
patients aged 35 years or younger, among those who used oral
contraceptives for a long time, used them in early reproductive life,
or used high-progestin oral contraceptives for at least 1 year.

Stanford and co-workers [ref: 729] found no increased breast cancer
risk among middle-aged women taking estrogen or combined estrogen and
progestin as hormone replacement therapy in a case-control study of
537 women with breast cancer and 492 control women without breast
cancer. Long-term use (>/= 8 years) of the combined therapy regimen
was associated with, if anything, a reduction in risk of breast
cancer. On the other hand, Colditz and colleagues [ref: 123] noted
that the addition of progestins to estrogen therapy in postmenopausal
patients did not reduce the risk of breast cancer (RR, 1.41), compared
with postmenopausal women who had never used hormones.

Exposure to ionizing radiation during or after puberty increases the
risk of developing carcinoma of the breast. Land and associates [ref:
426] reviewed reports on three populations of patients: a report by
Tokunaga and colleagues [ref: 772] on survivors of the atomic bombings
in Hiroshima and Nagasaki; a report by Boice and Monson [ref: 65] on
women in Massachusetts who had multiple fluoroscopic examinations of
the chest for pulmonary tuberculosis; and a study by Shore and
co-workers [ref: 683] of patients with postpartum mastitis who were
exposed to multiple x-ray examinations, in which nonexposed civilians
were used as a control group. They concluded that the risk of
radiation-induced cancer of the breast increased approximately
linearly with increasing doses and was heavily dependent on age at
exposure. In an update, Land [ref: 425] reported that most of the 54
cancers observed among women receiving doses higher than 100 rem (1
Sv) were probably caused by radiation. These observations were later
confirmed by other investigators. In a study of 31,710 women who had
tuberculosis and were examined with repeated fluoroscopic studies, a
substantial proportion (26.4%) received doses to the breast of 10 cGy
or more [ref: 503]; the breast cancer risk was greatest among women
who had radiation exposure between the ages of 10 and 14 (RR, 4.5 per
1 cGy and an additive risk of 6.1 per 104 person-years per 1 cGy);
there was substantially less excess risk with increasing age at first
exposure.

In another study of 1030 women with scoliosis who had multiple
radiographic examinations over a period of 8.7 years, 11 cases of
breast cancer were reported, compared with 6 expected (risk factor,
1.82) [ref: 353]; risk also increased with the number of x-ray
examinations and estimated irradiation dose to the breast (mean, 0.13
Gy). Furthermore, in a cohort of 1201 women who received x-ray
treatment in infancy for enlarged thymus gland (estimated mean
absorbed dose of irradiation to the breast, 6.9 Gy), after an average
of 36 years of follow-up, 22 breast cancers were diagnosed, compared
with 12 in 2469 nonirradiated sisters (adjusted risk factor, 3.6).
[ref: 350] The dose-response relation was linear, with an RR of 3.48
per 1 Gy and an additive excess risk of 5.7 per 104 person-years per 1
cGy.

A high risk of solid tumors, especially breast cancer, has been
described in women treated with radiation therapy at young age for
Hodgkin's disease. In a review of 1380 women treated at 15
institutions before the age of 16 years, 17 women developed breast
cancer -- 7 after radiation therapy alone and 10 after irradiation and
chemotherapy. Sixteen breast cancers appeared within or at the margin
of the irradiation fields. The cumulative probability of breast cancer
at 40 years of age was 35%. Women in this cohort of survivors had a
risk of breast cancer 70 times higher than that of the general
population. [ref: 56]

It should be stressed that the risk of breast cancer associated with
radiation exposure decreases sharply with increasing age at exposure,
and even a small benefit to women of screening mammography would
outweigh any possible risk of radiation-induced breast cancer. [ref:
204,206,503] For women between 50 and 75 years of age, the benefit of
annual screening mammography exceeds the radiation risk by a factor of
almost 100, and for women aged 35 to 75 years the benefit of reduced
mortality is projected to exceed the radiation risk by a factor of
more than 75. [ref: 497]

Loomis and associates [ref: 458] examined the relation between breast
cancer and exposure to low-frequency magnetic fields in 68 women with
breast cancer and 199 controls, all electrical workers, compared with
27,814 women with breast cancer and 110,750 controls, all with other
occupations; the electrical workers demonstrated excess mortality from
breast cancer (odds ratio, 1.38). There was no excess breast mortality
for seven other occupations with potential for increased electrical
exposure, including telephone operators and computer operators. Six
other studies have not shown an etiologic correlation between women
exposed to electromagnetic fields and the risk of breast cancer. [ref:
777]

Higher alcohol consumption has been correlated with increased risk of
breast cancer. [ref: 847] A study of 62,573 women from the Netherlands
reported an RR for breast cancer of 1.3 among women who consumed up to
30 g of alcohol daily and 1.72 among those who consumed greater
amounts. [ref: 789] Longnecker and associates [ref: 457] also noted an
RR of 1.39 with about one drink daily, 1.69 with two, and 2.3 with
three drinks daily.

A study of more than 100,000 women showed no association between
smoking and risk of breast cancer. [ref: 456]

Byrne and associates, [ref: 88] in a review of 280,000 women, noted
that women with 75% or greater breast density parenchymal patterns on
the mammogram, as measured by the proportion of breast area composed
of epithelial and stromal tissue, had a fivefold greater risk of
breast cancer. This parameter was independent of other prognostic
factors, such as family history, age at first birth, or alcohol
consumption.

Analysis of 2201 women between the ages of 30 to 62 years showed no
significant association between degree of adiposity and the incidence
of breast cancer but suggested that increased central-to-peripheral
body fat distribution may be a more specific marker than premalignant
hormonal pattern predisposing to this disease. [ref: 36]

Natural History

The most common site of origin of breast cancer is the upper outer
quadrant (38.5%), followed by the central area (29%), the upper inner
quadrant (14.2%), the lower outer quadrant (8.8%), and the lower inner
quadrant (5%). [ref: 305] These rates correlate with the amount of
breast tissue in the various quadrants. Cancer is somewhat more common
in the left than in the right breast; it is unusual for cancer to
appear in both breasts simultaneously (1% to 2%). Metachronous
bilateral carcinoma of the breast has been observed in 5% to 8% of
patients.

The growth rate of a tumor in the breast is thought to be constant
from the date of origin. [ref: 305] Using estimates of doubling time,
it would take an average of approximately 5 years for a tumor to reach
palpable size, and those lesions with slower doubling time would have
an even longer latent period. [ref: 305]

As the cancer grows, it travels along the ducts, eventually breaking
through the basement membrane of the duct, invading adjacent lobules,
ducts, fascial strands, and the mammary fat, spreading through the
breast lymphatics and into the peripheral lymphatics. The prognostic
and therapeutic implications of tumor multicentricity and
multifocality are discussed later in this chapter (see Prognostic
Factors). The tumor can grow through the wall of blood vessels, spread
into the deep lymphatics of the dermis, and eventually produce edema
of the skin (peau d'orange), which usually indicates that the
superficial as well as the deep lymphatics are involved. Ulceration
and infiltration of overlying skin, which may develop late in the
course of the disease, are usually preceded by fixation and localized
redness of the skin over the tumor [ref: 305] and are less frequently
seen because of current emphasis on screening and early diagnosis.

A common route taken by breast carcinoma as it metastasizes is first
through the axillary lymph nodes; the incidence of lymph node
metastasis increases with larger tumors. [ref: 168] About 20% to 40%
of newly diagnosed stage T1 and T2 breast cancers have pathologic
evidence of axillary nodal metastases, which are most frequent from
lesions of the upper outer quadrant of the breast. Table 50-1
summarizes incidence of axillary metastases according to the size of
the primary tumor.

In 263 patients with T1 unilateral invasive breast cancer, 72 patients
(27%) had nodes positive for metastasis. [ref: 102] Univariate
analysis showed that lymph node metastases were associated with tumors
larger than 1 cm (P = 0.001), moderate or poorly differentiated
nuclear grade (P = 0.005), high fraction of cells in the growth phase
(S phase) of the cell cycle (P = 0.041), presence of lymphatic
vascular invasion (P < 0.001), and age younger than 60 years (P =
0.01). The number of involved axillary lymph nodes definitely
influences prognosis and therapy outcome. [ref: 221]

In a study of 135 patients who underwent axillary lymph node
dissection in conjunction with either modified radical mastectomy or
lumpectomy (median follow-up, 6.9 years), there was no difference in
overall or disease-free survival between patients whose highest or
only level of axillary involvement was level I and those whose highest
or only level was level II. [ref: 43] Although patients whose highest
level of nodal involvement was level III had significantly worse
overall and disease-free survival rates, when patients were stratified
by the total number of positive nodes (1 to 3 versus >/= 4), there was
no difference in overall or disease-free survival rates in levels I,
II, and III.

Metastases to the internal mammary nodes are more frequent from medial
half and central lesions; these metastases occur more frequently when
there is axillary node involvement (Table 50-2). [ref: 324] The
supraclavicular lymph nodes may be the target of metastatic deposits,
usually after the high axillary or internal mammary lymph nodes are
involved by tumor, depending on the location of the primary lesion.

Vascular invasion by tumor and hematogenous metastases to the lungs,
pleura, bone, brain, eyes, liver, ovaries, and adrenal and pituitary
glands may be observed, even with small tumors. A highly significant
correlation was found between tumor size and incidence of distant
metastases. The distribution of tumor sizes and metastatic spread was
log-normal with a median diameter of 3.5 cm. The proportion of grade 1
tumors was higher in small tumors than in large ones, while the
reverse was observed for grade 3 tumors; these data suggest that,
during their growth, tumors progress toward higher grades. [ref: 779]

Patterns of Failure After Treatment

Although the frequency of recurrence is correlated with tumor stage,
the failure patterns after mastectomy are similar in patients with
various stages of disease. Valagussa and colleagues, [ref: 786] in
their study of operable breast cancer patients treated with radical
surgery, noted that "node-negative" patients had fewer failures than
"node-positive" patients, but that the proportion of locoregional and
distant metastases was essentially the same in both groups.

Fowble and associates, [ref: 257] in an analysis of node-positive
patients after mastectomy and chemotherapy who did not receive
postoperative radiation therapy, noted that isolated locoregional
recurrence correlated with the presence of four to seven positive
lymph nodes, T3 tumor stage, positive surgical margins, and high
nuclear grade. Patients with more than seven positive nodes tended to
fail systemically as well as locally, minimizing the frequency of
isolated locoregional recurrence.

DePietro and co-workers [ref: 156] found, in 800 patients with first
recurrence at various sites after mastectomy, that visceral metastases
were more frequent in patients younger than 50 years of age, whereas
local recurrence was more common in patients older than 50 years.
Survival after recurrence for patients with first metastasis confined
to the soft tissues was higher than for those with bone or visceral
metastases. A study by Hagemeister and colleagues [ref: 318] found
more tumor involvement than had been clinically suspected in 166
patients who died of breast cancer and had autopsy; most of these
patients had received treatment including chemotherapy. There were 325
unsuspected metastases; areas of tumor involvement included the
endocrine organs (40%), liver (30%), lungs (28%), cardiovascular
system (21%), and genitourinary system (21%). Major causes of death
were pulmonary insufficiency (26%), infection (24%), cardiac disease
(15%), and hepatic insufficiency (14%). The most common cause of death
was metastatic disease to various organs, accounting for 42% of all
deaths; infection was the second most common cause of death.

Fisher and associates, [ref: 218] in an analysis of patients treated
in National Surgical Adjuvant Breast Project (NSABP) Protocol B-06,
concluded that ipsilateral breast tumor recurrence was a harbinger of,
but not a cause of, distant metastases. While mastectomy or breast
irradiation after lumpectomy prevented expression of the marker
(breast relapse), neither lowered the risk of distant metastases,
which was determined by a host of prognostic factors.

The patterns of failure in intraductal, lobular in situ, and stage T1
and T2 breast cancer after breast conservation surgery are analyzed in
detail in later sections of this chapter.

Clinical Presentation

The majority of patients with carcinoma in situ, T1, or T2 breast
cancers present with a painless or slightly tender breast mass or have
an abnormal screening mammogram. Patients with more advanced tumors
may have breast tenderness, skin changes, bloody nipple discharge, or
occasionally change in the shape and size of the breast. Rarely
patients may present with axillary lymphadenopathy (which occasionally
may be painful) or even distant metastasis.

Screening Mammography

Approximately 40% of lesions reported by Austin and associates [ref:
28] were detected by mammography only. In another series,
approximately 50% of lesions detected by mammography only and 36% of
those detected by mammography and physical examination were invasive
carcinomas smaller than 1 cm. [ref: 34]

A pessimistic attitude toward the curability of breast cancer and a
skeptical view on the effectiveness and high economic cost of
mammography were expressed by Skrabanek. [ref: 700] However, there is
increasing evidence that early treatment after screening including
mammography is associated with reduced breast cancer mortality for
women 50 years of age and older. [ref: 537] Sixteen-year results from
the Health Insurance Plan (HIP), which involved two systematically
selected, randomly sampled groups of about 31,000 women aged 40 to 64
years, demonstrated that mortality was reduced by about one third in
screened women 50 to 59 years of age. [ref: 681,682] The HIP study
women were offered screening examinations; the control group was
observed and monitored. Death from breast cancer among cases detected
starting 3 to 3.5 years after screening ended became very similar for
study and control groups of women.

Figure 50-4 demonstrates the complex changes that occur in cumulative
case survival rate in a screening program. The survival difference
between mammography-only and clinical examination-only cases appeared
in years 7 to 10 after diagnosis; therefore, only long periods of
observation determine whether this represents a real separation of the
curves or chance variation. Although the greatest difference in
mortality between screened and control groups was detected in women 50
to 59 years of age when they entered the study, the differences are in
favor of the study group at all ages. However, in the age 40 to 49
subgroup the difference (46 versus 61 deaths at 14 years) was not
statistically significant with the available follow-up.

Several authors [ref: 400,705] agree that screening mammography in
women 40 to 49 years of age may reduce mortality from breast cancer,
although the last word on this controversy has not been written. [ref:
245] A Swedish study reported on 40,000 women aged 40 to 64 years who
were invited for screening mammography (single view) and 20,000 women
who served as controls. [ref: 272] Of 128 breast cancers detected in
the first round, 37 were stages II to IV, and in the second, 21 of 95
were stage II or greater, indicating a reduction in the number of
advanced cancers detected in the screening group from the first to the
second round. After 5 years, the number of tumors found in the study
population exceeded the number in the control population by 45%, with
a tendency toward more favorable stages in the screened women between
the ages of 40 and 64 years. A breakdown by age suggests a beneficial
effect of mammography in the 50- to 59-year-old age group but not in
women 40 to 49 or 60 to 64 years old. Several reports have shown the
cumulative incidence of stage II or worse tumors detected during a
screening trial as an early indicator of mortality outcome. [ref: 748]

Andersson and associates [ref: 12] reported on a study conducted in
Malmo with about 21,000 women invited for screening and 21,195 in the
control group. A total of 588 women in the study group and 447 in the
control group were found to have carcinoma of the breast, a large
proportion of which (26%) were noninvasive carcinomas; the proportion
of advanced cancers in women who did not attend screening was 72%
compared with 50% in the control group. Overall, women in the study
group age 55 or older had a reduction in breast cancer mortality (35%
versus 44%).

A large trial was conducted in the United Kingdom involving 45,841
women aged 45 to 64 who were offered annual screening by clinical
examination and mammography; 63,636 were taught breast
self-examination, and 127,117, for whom no extra services were
provided, constituted a control population. [ref: 783] No significant
reduction in mortality from breast cancer was observed during the
first 5 years among the various groups. However, thereafter there was
a 14% reduction in mortality, which rose to 20%, in the women
undergoing mammographic screening.

Peeters and co-workers [ref: 560] reported on a mammographic screening
program carried out in Nijmegen in 23,000 women age 35 to 64 years in
the first round, 7700 women in the second group, and 3900 younger
women in the third group. At the time of first screening examination,
21% of tumors were 10 mm or less, while 20% were over 20 mm. At later
examination 33% of the cancers were less than 10 mm and 20% over 20
mm. Detection rates of carcinoma at first examination were 9.5 per
1000 elderly women screened, 5.6 per 1000 for women age 50 to 64
years, and 2.3 per 1000 for women younger than 50 years. This study
was not randomized, and mortality reduction of 50% in the screened
population was evaluated by means of a case-control study.

Faulk and colleagues [ref: 203] compared mammographic screening in
32,140 women aged 50 years and older (10,914 elderly, 21,226 younger).
The cancer detection rate was substantially higher in elderly women
(9.2 per 1000 women) than in younger women (5.7 per 1000 women). The
median size of cancers in elderly women was 11 mm (versus 12 mm in
young women). Cancer stage was earlier in elderly women than in
younger women (84% versus 75% for stage 0 or I). Axillary nodal status
was 93% node negative in elderly women and 88% in younger women. They
concluded that mammographic screening is at least as effective in
detecting cancers for which there is a favorable prognosis in women
aged 65 years and older as it is in women aged 50 through 64 years.

Kerlikowske and associates, [ref: 390] in a metaanalysis of 13
selected studies, concluded that the overall RR for breast cancer
mortality for women aged 50 to 74 years undergoing screening
mammography compared with those who did not was 0.74; in contrast, the
relative risk in women aged 40 to 49 years was 0.93. Screening
mammography significantly reduced breast cancer mortality in women
aged 50 to 74 years after 7 to 9 years of follow-up, regardless of
screening interval or number of mammographic views per screen, whereas
there was no reduction in breast cancer mortality in women aged 40 to
49 years after 7 to 9 years of follow-up. In the Kopparberg part of a
Swedish trial, for the 40- to 49-year-old age group, predicted
survival was calculated from the size, node status, and grade of
cancers detected during the trial in comparison with those found in
two later series of tumors, one from the 1989-1992 Kopparberg
screening program and the other from the British Columbia Screening
Program that began in 1988. [ref: 749] Both of the Kopparberg programs
used a 2-year interval between mammograms. The effects of grade, node
status, and size on survival in the 40- to 49-year age group were
similar to the effects in older age groups, with mortality reduction
in the former group in the Kopparberg trial being 26%. These data
suggest that the most likely way to achieve further reduction in
mortality is to reduce the interval between screens, possibly to 1
year.

Screening Mammography (Continued)
(1 of 1)

A large screening study with different results was carried out in
Canada in women with no history of breast cancer and no mammography in
the previous 12 months who were randomly assigned to undergo either
annual mammography and physical examination (MP group) or annual
physical examination only (PO group) [ref: 502]; 39,405 women were
enrolled and monitored for a mean of 8.3 years. The rate of
screen-detected breast cancer on first examination was 7.20 per 1000
in the MP group and 3.45 in the PO group; more node-positive tumors
were found in the MP group than in the PO group. At subsequent screens
the detection rates were a little less than half the rates at screen
one. Of the women with invasive breast cancer through 7 years, 217 in
the MP group and 184 in the PO group had no node involvement, 66 and
56 had one to three nodes involved, and 32 and 34 had four or more
nodes involved. There were 38 deaths from breast cancer in the MP
group and 39 in the PO group; survival rates were similar in the two
groups. Women whose cancer had been detected by mammography alone had
the highest survival rate. The authors concluded that yearly
mammography in addition to physical examination of the breasts
detected considerably more node-negative, small tumors than screening
with physical examination alone, but it had no impact on the rate of
death from breast cancer up to 7 years of follow-up.

The results of the Canadian National Breast Screening Study have been
interpreted by some to suggest that mammography might be detrimental.
Comparison of women who had cancers detected by mammography with women
who did not have mammography and had palpable cancers suggests there
is no detrimental effect; rather, the use of mammography results in
improved survival. The poor results of the Canadian study likely are
due to the unbalanced allocation of women with advanced cancers (large
tumors, four or more positive nodes) to the screened group, the poor
quality of the mammography in the trial, and an insufficient sample
size. [ref: 399,754]

Using Breast Cancer Detection Demonstration Project data, Byrne and
associates [ref: 89] calculated 14-year breast cancer survival rates
among 4051 women diagnosed with breast cancer. Because of the higher
breast cancer survival rates among women aged 40 to 49 with poorer
prognostic characteristics, the breast cancer survival advantage for
having a smaller tumor, no positive lymph nodes, or breast cancer
detected by mammography alone was lower for women aged 40 to 49 than
women aged 50 or older at diagnosis.

Stacey-Clear and co-workers [ref: 728] analyzed 117 women younger than
50 years of age who were diagnosed with breast cancer based only on an
abnormal mammogram. Ductal carcinoma in situ was found in 47 (40%),
while 70 (60%) had infiltrating ductal or infiltrating lobular
carcinomas. During the same interval, 928 women in this age group
presented with palpable breast cancer. Five-year survival for all
mammographically detected cancer patients was 95%, whereas for women
with palpable cancers the survival was 74% (P < 0.00005).

De Koning and colleagues, [ref: 149] in a computer simulation analysis
of five Swedish randomized trials on mammographic screening, concluded
that the improvement in prognosis in women aged 40 to 49 years is much
smaller than for women aged 50 years or older. Approximately 70% of
the 10% observed reduction in breast cancer mortality (7% for women in
the 40 to 49 age group) may be attributable to mortality reduction
resulting from screening of these women after they reach age 50.

Feig [ref: 205] reviewed the published literature and various issues
concerning the potential benefit with mammographic screening of women
40 to 49 years of age. He concluded that, based on a metaanalysis by
Smart and associates, [ref: 705] with improved mammographic
techniques, mortality could be reduced by 35% in women younger than 40
to 49 years of age with two-view annual mammographic screening.

Wald and co-workers, [ref: 826] in a metaanalysis of six randomized
trials, observed a 15% reduction in mortality in women 40 to 49 years
of age for whom a screening mammography was performed, compared with a
reduction of 25% in women aged 50 to 74 years.

Solin and colleagues [ref: 719] analyzed 206 newly diagnosed and
treated breast cancers in 201 women from a health maintenance
organization. Eligibility for local treatment with breast-conserving
surgery and breast irradiation was significantly increased in women
who had undergone mammographic screening compared with women who had
not undergone mammographic screening (88% and 60%, respectively) (P <
0.0001). There was no difference in the use of breast-conserving
therapy in the two groups (44% and 37%, respectively) (P = 0.40);
however, there was a statistically significant difference for the
subgroup of women aged 50 years or older (49% and 21%, respectively)
(P = 0.16).

No randomized trial has been carried out comparing the effectiveness
of different screening intervals. Despite some controversy, [ref: 820]
the National Cancer Institute, [ref: 820] the American Cancer Society,
and the American College of Radiology recommend a baseline mammogram
at the age of 35 (30 in high-risk groups). Repeat examinations should
be carried out every 2 years beginning at age 40. In women older than
50, mammograms should be performed annually. Risk factor information
could be used to determine the optimal frequency of screening.

Hendee and Kellie [ref: 342] reviewed the economic aspects of
mammographic screening. They concluded in 1984 that the net increase
in cost associated with annual mammographic screening, initial
treatment, and subsequent terminal care charges for those dying of the
disease would be offset by the cost of medical care in the women who
died of other causes. Eddy and associates [ref: 183] reviewed the
value of mammography screening in women under 50 years of age and
concluded that evidence did not strongly support mammographic
screening for all 40- to 49-year-old women. They recommended a
flexible policy for asymptomatic women who are willing to bear the
cost, in consultation with their physicians.

Moskowitz [ref: 518] carried out a detailed cost/benefit analysis of
breast cancer screening and concluded that the cost/benefit range for
mammographic screening of a large population of asymptomatic women is
within the range accepted for other areas in health care.

Cancer incidence data from three U.S. metropolitan areas suggest that
both black and white cancer patients living in census tracts with
lower median education or income values are diagnosed in later disease
stages than are patients with higher median education or income.
Within education and income strata, black women had a less favorable
stage of disease at diagnosis than whites. The exception was in upper
education/income levels, where the disadvantage for blacks
disappeared. [ref: 201,431,434] These data provide additional evidence
that women of low socioeconomic status could benefit from education
and targeted breast screening.

Screening by Physical Examination

No controlled studies have evaluated the effectiveness of screening by
breast self-examination alone, except the United Kingdom trial. Using
Breast Cancer Registry data, Constanza and Foster [ref: 130,247] found
fewer deaths from breast cancer (14% versus 26%) and improved
estimated 5-year survival rates (75% versus 59%) among women who
reported performing breast self-examination compared with those who
did not. In the Breast Cancer Detection Demonstration Project, the
estimated overall sensitivity of breast self-examination in detecting
breast cancer was 26% compared with 75% for the combination of
clinical breast examination and mammography. [ref: 675]

Baines and associates [ref: 32] reviewed the potential value of breast
physical examination in 89,835 women participating in the Canadian
National Breast Screening Study, 50% of whom did not have mammography
and had only physical examinations performed by nurses. There was no
significant difference between nurse and physician examiner findings;
the authors concluded that physical examination of the breast by
trained nurses was useful and cost-effective.

Clinical and breast self-examination may be complementary, enabling
palpation of interval cancers and detection of tumors not visualized
by mammography. In 729 patients treated by mastectomy and axillary
dissection for primary breast carcinoma, disease-free survival at 10
years was significantly higher after detection by mammography (77% of
30 patients) or clinical screening breast examination (78% of 101
patients) compared with self-detection (64%). In addition, annual
clinical breast examination, compared with less frequent clinical
palpation, reduced the risk of recurrence, controlling for detection
modality (P = 0.03).

Diagnostic Workup

The workup of a patient with a breast mass including complete clinical
and family history is summarized in Table 50-3. The patient should
always be examined both sitting up and lying down (to confirm masses
felt on the sitting-up examination and to better detect lesions
lurking deeper in the breast or against the chest wall). Careful
inspection of both breasts should be made, including size, form and
symmetry, changes in pigmentation, scaling or discharge from the
nipple, and dilated veins or edema of the skin in a nonpregnant
patient. The location, size, consistency, tenderness, and mobility of
the palpable tumor should be recorded. It is useful to draw and
photograph the projection of any suspicious or palpable masses on the
skin of the breast or nodal areas.

In addition to examination of the breast, careful evaluation of the
axilla and supraclavicular node areas is mandatory. The number,
consistency, tenderness, mobility or fixation, and size of lymph nodes
should be noted. Clinically node-negative patients have pathologic
involvement in 10% to 40% of cases (depending on primary tumor size),
whereas no tumor is found in 25% to 30% of patients with clinically
palpable axillary nodes. Examination of the abdomen for liver
enlargement and evaluation for bony pain are also essential. Finally,
a complete pelvic examination should be part of the overall evaluation
of the patient, if not recently performed by the patient's other
physicians.

Laboratory studies include a complete blood count and chemistry
profile, with particular emphasis on alkaline phosphatase level as
well as liver function tests (e.g., SGOT, serum glutamic oxaloacetic
transaminase; SGPT, serum glutamate pyruvate transaminase; LDH,
lactate dehydrogenase; biliurubin).

Imaging Techniques in Breast Cancer

Radiographic studies include chest x-ray, bilateral mammograms, and
plain radiographs of symptomatic bones, if clinically warranted.

Kopans and associates [ref: 401] reviewed diagnostic imaging
techniques for evaluation of patients with breast cancer and their
advantages or disadvantages. Mammography performed with dedicated
units and high-sensitivity screen film (or Xerox) techniques has been
described. [ref: 165] Classically, breast carcinoma is seen as an
ill-defined mass that may have spiculated margins (Fig. 50-5),
although rarely cancers may also be seen with a knobby, lobulated, or
even a smooth contour (ultrasound may distinguish them from cystic
masses). Architectural distortion of the breast tissue may be present.
The appearance of linear, radiated, or spiculated changes about a
central focus should always be considered suspicious for carcinoma.
The tumor may be hidden by dense parenchyma; review of previous
mammograms is very important in detecting subtle interval changes in
the appearance of the breast.

Calcifications can be associated with either benign or malignant
conditions of the breast. However, calcifications associated with
malignant tumors are typically 100 to 300 micrometers in size and are
rodlike, tubular, branching, or punctate. Clusters of
microcalcifications (more than five) are suggestive of intraductal
disease, and in nonpalpable lesions excisional biopsy done after
needle localization will aid in the diagnosis (Fig. 50-6). About 30%
of biopsies (51 of 173 patients) with clusters of at least five
microcalcifications without palpable findings showed malignancy; 56%
of the lesions were noninvasive, and 37% were invasive ductal
carcinoma. Solin and associates, [ref: 712] in a review of 1507 cases
of breast cancer, noted that mammographic needle localization breast
biopsy increased from 3% in 1977 to 1978 to 26% (111 of 421) in 1987
to 1988. The incidence of intraductal carcinoma referred for breast
conservation therapy increased from 6% (2 of 32) to 13% (53 of 421).

The average sensitivity of mammography is about 90% (60% to 95%), and
the specificity is 94% (50% to 98%). The positive predictive value is
approximately 8% to 14% for screened patients, but is significantly
higher for patients with symptoms or palpable masses. [ref: 401] If
microcalcifications were initially present, radiographs of the
surgical specimen and posttylectomy mammography are important to rule
out residual disease for patients considering breast conservation
therapy. [ref: 157,172]

Mammograms were performed in 120 consecutive patients treated with
conservation surgery and breast irradiation. [ref: 759] In six
patients (5%), calcifications or masses proved to be residual tumors.
This led to reexcision in two, mastectomies in two, and a higher
irradiation boost dose to the tumor in two. In 39 patients (32%), the
preirradiation mammogram provided information helpful for the
interpretation of postirradiation mammograms.

Ultrasonography has a reported sensitivity of 73% and specificity of
95% [ref: 704]; it is very helpful in differentiating cysts from solid
tumors. Major disadvantages of ultrasonography include inability to
image microcalcifications and great difficulty in visualizing lesions
less than 1 cm in diameter. Medullary cancer, with its often sharp
margins, may simulate a benign fibroadenoma. In addition, fatty breast
tissue is poorly examined because of its high echogenicity. [ref: 850]

Transillumination, the use of infrared or near-infrared light for
screening tests for breast disease, has a low sensitivity because of
the inability to detect small or deep tumors. Monsees and associates
[ref: 506] reported a 58% sensitivity, and Gisuold and co-workers,
[ref: 290] 67%; others have reported somewhat higher figures (85% to
88%). [ref: 174,827]

Thermography has no significant value for either screening or
diagnosis of breast disease because of its low sensitivity and
specificity. [ref: 848] In a study of 420 patients, an abnormal
thermogram was found in 18.6% of women with invasive ductal carcinoma
[ref: 732]; this finding was associated with large tumors, high grade,
or positive lymph nodes. However, thermography was not an independent
prognostic factor.

Magnetic resonance imaging (MRI) is being evaluated for imaging breast
diseases, and some authors suggest a sensitivity and specificity
comparable to those of mammography [ref: 145,347]; however, cost and
availability are major deterrents at the present time. Use of
gadolinium DTPA, a paramagnetic agent, increases the contrast between
normal and malignant tissues. Heywang and associates [ref: 348] were
able to visualize 20 small carcinomas with MRI, 3 of which would have
been missed by non-contrast-enhanced imaging, and 2 carcinomas in
dense breast that were not visible on mammography. Magnetic resonance
spectroscopy provides information about cellular biochemical processes
by measuring the concentration of energy systems and metabolism,
particularly of organic phosphorus. [ref: 466] Degani and colleagues
[ref: 147] studied postoperative specimens from five benign and nine
malignant breast tumors and showed that the concentrations of
nucleosides, triphosphates, and phosphomonoesters are consistently
higher by a factor of 3 in carcinomas than in benign lesions. The
studies have been conducted in vitro, and there is no foreseeable
clinical application.

Computed tomography (CT) outlines tumors or lymph nodes greater than 1
cm in size. The need for iodine contrast material to differentiate
benign from malignant conditions, high radiation dose, cost per study,
and inability to detect small lesions preclude the use of CT for
initial evaluation, except under special circumstances.

Bone scan is frequently used to evaluate asymptomatic bony metastases.
In patients with stage I disease, the incidence of abnormal bone scan
is about 2%, but a significantly greater incidence of abnormalities is
found in stages II (10%) and III (over 30%). [ref: 85,112] In a group
of 7604 patients who had bone scan, out of over 20,000 women operated
for breast cancer in Denmark, [ref: 765] about 5% had abnormal
studies. The incidence of abnormal scans was greater in patients over
60 years (8%) than in the younger group (3%), most likely because of
the many benign bone disorders frequently seen in older women. Routine
bone scan at the time of initial treatment of stage I and II breast
cancer is of limited value and should be reserved for patients with
bone pain [ref: 626,764]; Alcazar and colleagues [ref: 7] reported one
true-positive result in a study of 112 patients.

Internal mammary lymphoscintigraphy (procedure in which
**99mTc-labeled colloid is injected subcutaneously into the abdominal
wall) has been advocated on the basis that 15% of patients demonstrate
cross-drainage between parasternal lymphatics, and in 30% of patients
parasternal lymph nodes lie outside usual radiation therapy portals.
[ref: 186] McLean and Ege [ref: 490] reported a sensitivity of 76% and
a specificity of 67% in 62 patients with axillary lymphoscintigraphy
when compared with the findings at axillary dissection. A very
critical factor of this technique is that it indicates disease by the
absence of normal tissue uptake rather than by demonstrating the
disease itself. [ref: 630] Ege and Clark, [ref: 186] in 524 patients
with carcinoma of the breast treated with partial mastectomy and
irradiation who had internal mammary lymphoscintigraphy, observed a
statistically significant difference in actuarial survival in patients
with normal versus abnormal scanning of internal mammary lymph nodes.

Positron emission tomography (PET) using **18F-labeled
fluorodeoxyglucose is being more frequently used in breast cancer. In
51 patients with 72 suspicious lesions, 41 (57%) were found to have
malignant tumors. PET sensitivity ranged from 68% to 94% and
specificity from 84% to 97%. Sensitivity to detect lesions smaller
than 1 cm was limited due to partial volume effects. [ref: 29] This
test may be of value in the presurgical detection of regional lymph
node or distant metastases. [ref: 5]

Pathologic Studies

Histopathologic diagnosis may be obtained by fine-needle aspiration of
cystic or solid masses or biopsies of solid masses; any fluid
aspirated from the breast should be examined for malignant cells.
Fine-needle aspiration of the breast is a simple, low-cost, accurate
diagnostic technique that has been used for many years in Europe,
particularly Sweden. [ref: 870,871] This technique is gaining
increasing acceptance in the United States. [ref: 49,830,850]

Breast biopsy of any suspicious mass is mandatory. The biopsy usually
can be done using local anesthesia; the patient should be informed of
the nature of the lesion to allow for her greater participation in
therapeutic decisions. There has been no evidence that delay in
treatment up to 2 weeks after biopsy significantly worsens prognosis.
[ref: 216]

Stereotactic needle biopsy is being increasingly used to obtain
histologic diagnosis with high accuracy. It is more frequently used in
the diagnosis of small breast lesions. [ref: 556] In 6152 lesions
biopsied at multiple institutions, 817 (13.3%) had infiltrating breast
cancer, 167 (2.7%) had intermediate- or high-grade ductal carcinoma in
situ (DCIS), and 213 (3.5%) had atypical hyperplasia or low-grade
DCIS. Complete agreement between the core biopsy and subsequent
histologic sections with ultrasound guidance was reached in 89.7% of
lesions and partial agreement in 9.2%. Clinically significant
complications occurred in only 6 of 3765 cases (0.2%) for which
follow-up was available.

In nonpalpable lesions, needle localization and radiographic
techniques are necessary to identify the tissue to be removed. Failure
to remove the mammographic abnormality has been reported in 2% to 8%
of patients undergoing needle localization. [ref: 427,768] Sometimes
several biopsy attempts are necessary, and close communication between
the diagnostic radiologist and surgeon is critical. The localizing
wire should be left in place in the specimen, and a radiograph should
be obtained to ensure that the area of abnormality has been adequately
excised. If the specimen x-ray does not document complete tumor
removal, an immediate reexcision of the area at the tip of the wire
should be carried out. If there is any question, particularly in
patients with microcalcifications, a postbiopsy mammogram should be
obtained to determine the completeness of the tumor excision. [ref:
172]

Although a one-stage procedure (frozen section and mastectomy) was
advocated in the past, in recent years a two-step approach has gained
wide acceptance, in many instances because of patient preference. If
breast conservation surgery is considered, the surgeon must have as
much knowledge of the nature of the mass as possible, because, if
malignancy is suspected, a wide local excision with adequate margins
should ideally be carried out. The surgeon should prepare (orient) the
specimen accordingly, and the margins of the resected breast tissue
should be identified and inked before processing. The pathologist
should be made aware of the nature of the lesion, for appropriate
processing of the specimen. [ref: 653] Radiation oncologists should be
familiar with the implications of the diagnostic procedures for
carcinoma of the breast as active participants in a breast
preservation therapeutic approach.

Estrogen and progesterone receptor assays are routinely done in the
United States in patients with breast cancer; these parameters are
correlated with prognosis and tumor response to chemotherapeutic and
hormonal agents. [ref: 454,853] Immunohistochemical techniques have
been developed for qualitative (not quantitative) estrogen and
progesterone receptor assessment if no fresh tissue is available.

Cellular assays measure the growth fraction (S-phase fraction, SPF) of
tumors, either by thymidine labeling index (TLI) or flow cytometry
methods. [ref: 501] The growth fraction assay is a prognostic variable
that is independent of tumor stage and hormone receptor status.

Staging Systems

Two staging systems are widely used for breast cancer: the American
Joint Committee (AJC) [ref: 46] and the Union Internationale Contre le
Cancer (UICC) [ref: 782] systems (Table 50-4). The Columbia system
(Table 50-5) is important, both historically and because it clearly
identifies prognostic factors affecting operability. [ref: 305,307]
Figure 50-7 depicts the various clinical stages according to tumor and
nodal characteristics.

Pathologic Classification

The World Health Organization (WHO) has classified proliferative
conditions and tumors of the breast into the following categories:
benign mammary dysplasias, benign or apparently benign tumors,
carcinoma, sarcoma, carcinosarcoma, and unclassified tumors. [ref:
647] The AJC [ref: 46] has developed the alternate system shown in
Table 50-6.

Numerous detailed reports and monographs describe the pathologic
features and clinical implications of carcinoma of the breast. For a
concise review, the reader is referred to publications by Fisher and
associates, [ref: 235,238] Harris and colleagues, [ref: 332] and
Rosen. [ref: 615] The radiation oncologist should be familiar with the
histologic characteristics of breast cancer, because many of them
affect prognosis and may have important therapeutic implications.

Brief descriptions of several types of carcinoma of the breast follow.

Intraductal carcinoma or ductal carcinoma in situ (DCIS) is a
noninvasive lesion with five histologic subtypes. [ref: 202] Some
investigators consider solid DCIS to represent an earlier form of
comedocarcinoma. Papillary and micropapillary DCIS are distinct
entities, with the former having invaginations of a fibrovascular core
within the lumen of the duct that are completely absent in the
micropapillary histology. Cribriform DCIS is histologically
distinctive, with a sieve or "Swiss-cheese" appearance within the
lumen of the duct. Comedo-intraductal carcinoma is more likely to have
a higher nuclear grade, higher TLI, foci of microinvasion, and a
higher local recurrence rate after tylectomy alone. [ref: 422,499,557]
Page and Jensen, [ref: 553] in an incisive editorial, pointed out the
major pathologic and biologic differences in DCIS and the need to
separate the tumors into low, intermediate, and high grade, as this
stratification has important prognostic implications. Table 50-7 lists
the various types of intraductal carcinoma and some important
pathologic features.

Lobular carcinoma in situ (LCIS) is a noninvasive proliferation of
abnormal epithelial cells in the lobules of the breast. To make the
diagnosis of LCIS, Page and colleagues [ref: 554] require that the
tumor be composed of evenly spaced cells that completely fill,
distort, and distend more than half of the acini in a lobular unit.
The cells are generally small and lack atypia. Mitoses are rarely
seen. Because the lobules of the breast atrophy after menopause, LCIS
is primarily an incidental premenopausal finding; most are nonpalpable
and mammographically silent.

Invasive (infiltrating) ductal carcinoma is the most common type of
breast cancer, comprising more than 50% of all cases. It appears as
solid cords or groups of ductal tumor cells varying in size and
cytoplasmic content and degree of differentiation. [ref: 235] Necrosis
is rare, but lymphatic invasion may be present. An associated in situ
component is frequently seen.

Tubular carcinoma is composed of tubular structures typically lined by
a single layer of well-differentiated epithelium. The tubular cells
simulate those of normal ducts or ductules, [ref: 99] arranged in
multiglandular cribriform or adenocystic configurations and are
frequently associated with other in situ carcinomas of the breast.
[ref: 486] Tubular carcinomas have a nonaggressive growth pattern;
axillary lymph node involvement is reported in about 10% of patients.

Medullary carcinoma is composed of cords and masses of large cells
with reticular pleomorphic nuclei containing prominent nucleoli. There
is a scant fibrous stroma, but lymphoid infiltrate is prominent. These
tumors are microscopically and grossly well circumscribed, and lymph
node metastases are infrequent. Prognosis, in general, is better than
for other tumors.

Lobular invasive carcinoma may be interspersed with lobular carcinoma
in situ; the cells appear singly or in small clusters in a targetoid
or single-file pattern. Some scirrhous carcinomas probably are
invasive lobular lesions; these tumors tend to be aggressive and
multicentric and are prone to develop distant metastases. Du Toit and
co-workers [ref: 178] reported five subtypes of lobular carcinomas in
171 cases and observed a 12-year actuarial survival rate of 100% for
the tubulolobular subtype but only 47% for the solid variant. The
other three subtypes had intermediate prognoses.

Mucinous carcinoma, also called mucoid or colloid carcinoma, has been
observed in older women with relatively long duration of symptoms.
[ref: 685] It has a greater likelihood to be devoid of a cellular
reaction; necrosis and lymphatic invasion are very rare. It is slowly
growing with a pushing border and has a low frequency of axillary
lymph node metastasis. [ref: 685] Survival is appreciably better than
with infiltrating ductal carcinoma. [ref: 533]

Adenocystic carcinoma is rarely found in the breast. Histologic
features and clinical behavior are similar to its counterparts in the
salivary gland and the upper respiratory tract.

Papillary carcinoma is a rare neoplasm, histologically characterized
by frondlike projections with fibrous supporting stalks. Its delicate
or nonexistent fibrovascular cord, nuclear hyperchromatism, and
absence of double layer of cells and apocrine changes distinguish
papillary carcinoma from intraductal papillomas. [ref: 404] These
lesions have a low tendency to invade or metastasize to the regional
lymph nodes.

Primary neuroendocrine small cell carcinoma has been reported in seven
cases by Francois and associates. [ref: 269] The histology and
prognosis are identical to those of lung cancer. It is important to
distinguish these lesions from metastatic lung tumors, direct invasion
of breast by Merkel cell tumor, lymphoma, or carcinoid tumor. All of
the patients reported have been treated with radical or modified
radical mastectomy, sometimes combined with irradiation, chemotherapy,
or tamoxifen. It is reasonable to treat these patients with aggressive
multiagent chemotherapy, excision of the primary tumor, and breast
irradiation, although no data are available on the outcome of this
approach.

Apocrine intraductal carcinoma is occasionally diagnosed, and
treatment and prognosis are similar to those of DCIS. Axillary
dissection, as in DCIS, is unnecessary inasmuch as none of 37 women
with the diagnosis of apocrine intraductal carcinoma had evidence of
lymph node metastases. [ref: 755]

Paget's disease describes involvement of the nipple by tumor. Most
investigators agree that it represents extension of neoplasms from
subjacent ducts in the nipple [ref: 368] or metastases from an
underlying carcinoma. [ref: 105] The tumor seems to travel linearly
down the ducts and may appear to be multicentric. There may be an
associated subareolar tumor.

Carcinosarcoma is a rare tumor characterized by nodular,
circumscribed, or irregular masses without encapsulation. Squamous
differentiation with carcinoma was noted in 15 of 70 tumors reported
by Wargotz and Norris. [ref: 832] The sarcoma was predominantly
polymorphous, frequently resembling malignant fibrous histiocytoma in
40 and fibrosarcoma in 28. Cellularity of the sarcoma component was
high (97%). Thirteen of 50 patients (26%) had axillary lymph node
metastasis. Only 1 of 11 patients had hormonal receptors. The 5-year
survival rate was 49% for carcinosarcoma in comparison with 64% for
spindle cell carcinoma.

Cystosarcoma phyllodes is usually a benign lesion; within broad
fibrous beads that look "leaflike" are cystic clefts lined by a single
layer of cells. These tumors are large; usually they are encapsulated,
without invasion of the adjacent breast. The lesions frequently
develop from preexisting fibromas and have a long initial period of
slow growth followed by sudden, rapidly increasing size. The grade
(mitotic rate) and surgical margins have prognostic importance.
El-Naggar and colleagues [ref: 190] performed DNA analysis of 30
cystosarcoma phyllodes and concluded that DNA ploidy, proliferative
index, number of mitoses, and tumor margin were significantly
associated with prognosis. Several authors have reported a few cases
in which these tumors metastasized to the other breast, axillary lymph
nodes, mediastinum, or lungs. In a report by Treves and Sunderland
[ref: 776] of 77 patients, 18 lesions were classified as malignant, 18
as borderline, and 41 as benign. Nine patients with malignant tumors
developed distant metastases.

Primary mammary lymphomas are rare. Thirty-five cases were analyzed
including 16 primary lymphomas. Diffuse large cell lymphoma was
present in 10 of 16 primary and 14 of 18 secondary cases. In a series
of 16 cases, 7 primary in the breast, 15 were non-Hodgkin's lymphoma.
[ref: 18] Lymphoepithelial lesions in ducts and lobules and frequent
vascular involvement were found in both primary and secondary cases.
Immunohistochemistry studies of 13 tumors showed that 12 were B-cell
in origin, and 1 was a primary T-cell lymphoma. Survival was related
to stage and histologic characteristics. Half of the patients with
primary lymphoma had recurrent disease; although local recurrences
were observed, a tendency for recurrence in other extranodal sites was
common. These patients should be treated like others with similar
localized lymphoma.

Other unusual tumors occasionally described in the breast include
sarcoma, pure squamous cell carcinoma, and basal cell carcinoma.

Prognostic Factors

The prognostic factors influencing local relapse and survival can be
divided into intrinsic factors, which are related to the
characteristics of the tumor (e.g., histologic features, lymph node
metastases), and extrinsic factors (e.g., host factors, the type and
adequacy of treatment).

Intrinsic Prognostic Factors

Tumor Size and Clinical Stage

Size and clinical stage are strong prognostic factors influencing
incidence of local recurrence, nodal and distant metastases, and
survival. [ref: 283]

Some series of patients treated with breast conservation therapy show
a correlation between the size of the tumor and the incidence of local
recurrence, whereas this observation is not reported by other authors
(Table 50-8). This difference is probably related to the treatment
techniques used (e.g., completeness of tumor excision, use of
irradiation boost). In general, patients with stage T1 tumors have
better disease-free and overall survival than patients with stage T2
tumors; outcome is strongly influenced by nodal status.

Diffuse, infiltrating tumors have a higher incidence of local
recurrence than localized tumors. In patients treated with mastectomy,
analysis of skin changes (edema, erythema, and ulceration of the
breast) showed that only edema was statistically significant for
increased local recurrence. [ref: 168]

Presence of Multiple Primary Tumors (Multicentricity)

Noguchi and colleagues, [ref: 532] in a careful clonal analysis of 30
breast carcinomas, noted that every tumor was monoclonal in origin,
leading them to conclude that multiple breast cancers are multifocal
(arising from a single primary focus that spreads throughout the
breast) and not multicentric in origin (separate and independent
multiple tumors).

In 657 mastectomy specimens, Gump and associates [ref: 302] noted
multifocal disease in 19% of patients with invasive ductal carcinoma,
as opposed to 81% of 42 patients with intraductal carcinoma and 50% of
92 patients with invasive lobular carcinoma. Although in invasive
ductal carcinoma tumor multicentricity was related to size (12% in
lesions < 2 cm, 23% in those > 2 cm), this was not observed in
invasive lobular carcinoma, in which the corresponding multicentricity
values were 54% and 45%, respectively.

In patients with gross multicentric disease or diffuse
microcalcifications, it is difficult to perform breast-conserving
surgery and have satisfactory cosmetic results. Also, these patients
are at greater risk of breast relapse. In general, depending on the
extent of the tumors, these patients are treated with a modified
radical mastectomy. Fowble and colleagues, [ref: 262] in 88 such
patients with stage I and II breast cancer, described 42% having
positive axillary lymph nodes. After modified radical mastectomy, 15
patients received postoperative irradiation and 35 adjuvant
chemotherapy. The 5-year locoregional recurrence rate was 8%, which
compares well with patients with unifocal disease (7%). The overall
5-year survival and disease-free survival rates were 87% and 76%,
respectively, in the node-negative group and 90% and 71% in the
node-positive group. A contralateral breast primary tumor occurred
either synchronously or metachronously in 6 of 57 patients (10.5%)
with gross multicentric disease and 4 of 31 (12.9%) with diffuse
microcalcifications. Management of patients with multicentric tumors
will be discussed later.

Location of Primary Tumor

Fowble and co-workers [ref: 261] analyzed recurrence and survival in
886 patients with stage I and II breast cancer treated with breast
conservation therapy (median follow-up, 5 years). The patients were
divided into four groups according to the primary tumor location:
outer (495 patients), inner (202 patients), central (119 patients),
and subareolar (70 patients). Subareolar tumors were defined as those
immediately beneath the nipple-areolar complex or within 2 cm of the
areolar margin. There were no significant differences in 5-year
actuarial overall survival (91%, 86%, 92%, and 91%, respectively, P =
0.34), relapse-free survival (75%, 74%, 80%, and 79%, P = 0.77), or
disease-free survival (82%, 78%, 87%, and 84%, P = 0.29) among the
four groups. A separate analysis for pathologic node-negative and
node-positive patients disclosed similar findings.

Haffty and associates, [ref: 317] in a review of 1014 patients with
early breast cancer treated with breast conservation therapy,
identified 98 patients who fulfilled the criteria of having a
central/subareolar tumor. Ten of 98 patients had the nipple-areolar
complex sacrificed at the time of surgery, while the remaining 88
patients had the entire area included in the boost cone-down field.
Reexcision was performed in only 16 patients. Median radiation dose
was 48 Gy to the breast and 64 Gy (adding the boost) to the tumor
site. The 10-year actuarial breast recurrence-free survival rate was
84%, the distant disease-free survival rate was 88%, and and the
overall survival rate was 79%, similar to patients with tumors in
other locations. The nipple-areolar complex could be preserved in most
patients, and there were no significant complications. Thus a
subareolar breast cancer presentation is not a contraindication to
breast-conserving therapy in early-stage disease.

Histologic Features and Tumor Grade

Schnitt and associates [ref: 657] compared the results of tumor
excision and irradiation in 49 patients with infiltrating lobular
stage I and II carcinoma and 561 patients with similar stages of
infiltrating ductal carcinoma. The 5-year actuarial risk of local
recurrence was similar for both groups (12% versus 11%). The 12% local
recurrence rate for infiltrating lobular carcinoma was intermediate
between that for infiltrating ductal carcinoma with or without an
extensive intraductal component (23% and 5%, respectively). In our
experience in 839 patients, the breast relapse rates after breast
conservation therapy are equivalent for invasive ductal or lobular
carcinoma (6% in 712 patients with ductal and 2% in 49 with lobular
carcinoma), and survival is comparable, an observation similar to that
of others. [ref: 239,603]

Bornstein and co-workers [ref: 73] identified 93 patients with
invasive lobular carcinoma, 1089 with invasive ductal disease, and 59
with both ductal and lobular histology in a group of 1863 patients
with T1 or T2, N0 or N1 tumors treated with breast conservation
therapy. The 10-year disease-free survival rates were 56%, 55%, and
42%, respectively. Local recurrence rates were 15%, 13%, and 13%,
respectively. Contralateral breast cancer developed in 4%, 6%, and
13%.

Weiss and colleagues [ref: 837] reported on 879 patients with stage I
and II breast cancer treated with conservation surgery and
irradiation. The patients were divided into seven groups based on
histologic subtype: 368 patients with infiltrating and intraductal
ductal carcinoma, 389 with infiltrating ductal carcinoma, 41 with
infiltrating lobular carcinoma, 23 with combined infiltrating ductal
and lobular carcinoma, 28 with medullary carcinoma, 12 with colloid
carcinoma, and 18 with tubular carcinoma. Patients with tubular and
colloid carcinomas were more likely to present with T1 lesions,
hormone receptor positivity, and node-negative status than with other
histologic subtypes. Most medullary carcinomas were hormone
receptor-negative and most of these patients were younger than 50
years old. There were no significant differences in 5-year actuarial
overall survival, cause-specific survival, or relapse-free survival
among the histologic categories. In addition, patterns of first local
failure were not significantly different among the histologic groups.
There was, however, a difference among the seven groups in distant
metastasis only at first failure, with invasive ductal carcinomas
having the highest rate.

The incidence of local recurrence is greater and the survival rate is
decreased with undifferentiation and necrosis of the tumor, higher
nuclear grade, vascular invasion, and inflammatory infiltrate. [ref:
168,191]

Clemente and associates, [ref: 121] in 506 cases of ductal
infiltrating carcinoma (T1-2N0M0), described peritumoral lymphatic
invasion in 6.9% of routinely evaluated specimens, whereas in a
randomly selected group of 234 cases the frequency was 20%. Patients
with routinely evaluated peritumoral lymphatic invasion had worse
disease-free and total survival rates than those without this feature
(P = 0.0001 for each), as well as more local recurrences (P = 0.0001)
and a higher incidence of distant metastases (P = 0.0576).

Extensive Intraductal Carcinoma

According to the Harvard definition of extensive intraductal carcinoma
(EIC), 25% or more of the primary tumor is intraductal carcinoma, and
intraductal carcinoma is seen outside (adjacent to) the infiltrating
border. [ref: 328,329] EIC involving the primary tumor and adjacent
tissues has been reported by some groups, particularly Harvard
University [ref: 329,656] and Marseilles, [ref: 414] to be associated
with a higher incidence of breast recurrences. In contrast, Clarke and
associates, [ref: 118] Fisher and colleagues, [ref: 242] van Limbergen
and co-workers, [ref: 795] and an analysis of our experience found no
significant impact on local tumor control with extensive intraductal
extension. This difference may be related to the pathologic criteria
used in the definition of EIC, adequacy of tumor excision, and doses
of irradiation delivered to the boost volume. Table 50-9 summarizes
reports of breast relapse correlated with presence of EIC.

Fourquet and associates [ref: 252] reported a 20% incidence of EIC in
185 women younger than 45 years of age compared with 10.4% in 279
older women. As suggested by Kurtz and colleagues, [ref: 410] these
two prognostic parameters may be interdependent. At Washington
University Medical Center, a somewhat higher breast relapse rate in
EIC-positive patients was seen only in women younger than 40 years of
age (Fig. 50-8).

Holland and co-workers [ref: 355] stated that an EIC component is
associated with subsequent breast recurrence because of the presence
of residual intraductal carcinoma in these patients. In a series of
214 women who underwent mastectomy, 71% of those with EIC had residual
intraductal carcinoma, compared with 28% of those without that
pathologic feature. In particular, 44% of the EIC-positive patients
had prominent residual tumor compared with 3% of those who were EIC
negative (P < 0.00001). Carefully assessed negative surgical margins
and adequate irradiation may decrease or eliminate the significance of
EIC for local failure.

Involvement of Axillary Nodes by Tumor

Haagensen [ref: 306] demonstrated a direct relation between tumor
involvement of axillary nodes and chest wall recurrence and an inverse
correlation with survival in patients treated with radical or modified
radical mastectomy. Fisher and associates [ref: 233] described
diminishing survival after mastectomy with a greater number of
metastatic axillary lymph nodes. [ref: 221]

Valagussa and colleagues, [ref: 786] in addition, noted that patients
with involved internal mammary nodes had more recurrences at 10 years
than those who did not (27.9% relapse rate in patients with negative
axillary nodes; 60% in patients with negative axillary nodes and
involved internal mammary nodes). The Primary Therapy of Breast Cancer
Study Group [ref: 584] and others [ref: 186,228,660,867] confirm these
findings; lower survival and a greater incidence of local recurrences
in patients with positive axillary nodes after partial mastectomy were
initially reported. [ref: 429] At the Institut Gustave-Roussy, [ref:
643] among 356 patients, local recurrence was noted in 26% of those
with and 6.5% of those without nodal involvement; the greater the
number of nodes involved, the more likely the occurrence of local
failure and the lower the survival rate. [ref: 168] Increased
incidence of breast relapse was also observed by van Limbergen and
colleagues [ref: 795] in patients with N1b metastasis (8 of 42
patients, or 19%) and in those in whom three or more lymph nodes were
positive or the apical axillary lymph nodes contained metastatic tumor
(4 of 14 patients, or 28.6%), compared with patients with N0 or N1a
lymph nodes (14 of 187, 7.5%). However, more recent reports of several
authors noted lower survival but fewer breast relapses after breast
conservation therapy in patients with positive nodes, a result of the
interaction of irradiation to the breast with adjuvant chemotherapy.
[ref: 219,562]

S-Phase Thymidine Labeling Index

Meyer and colleagues [ref: 499] demonstrated a significant correlation
between SPF, as determined by TLI, and survival in 278 patients with
primary breast cancer. TLI below the median of 4.55% carried a 20%
probability of relapse at 4 years, in contrast to 52% with TLI above
the median. These findings were confirmed by Tubiana and co-workers
[ref: 780] in 128 patients with breast cancer who were monitored for
more than 15 years. Patients with positive hormonal receptor status
but high TLI are at a greater risk for early relapse, whereas estrogen
receptor (ER)-negative patients with low TLI have a low probability of
early failure.

Patients with high SPF have a greater tendency to develop distant
metastases. Hatschek and associates [ref: 337] noted decreased
disease-free survival and increased distant metastasis rates in
patients with either negative or positive lymph nodes who had a high
SPF on cytofluorometric studies.

In an update, Meyer and Province [ref: 500] reported on 845 women with
stage I and II tumors treated surgically. Tumor size, axillary lymph
node status, SPF, nuclear size, and ER status all related strongly to
breast cancer-specific survival and relapse-free survival. Lymph node
status and tumor size predicted long-term survival; differences for
other variables largely disappeared by 10 years. By multivariate
analysis, axillary lymph node status, tumor size, and ER were
independently prognostic for disease-specific, relapse-free survival.
Progesterone, DNA ploidy, and SPF did not independently predict
prognosis. Nuclear size was the strongest independent predictor in
patients with negative axillary lymph nodes.

DNA Ploidy Index

Most breast cancers exhibit a bimodal distribution of DNA values.
Diploid tumors tend to have a better prognosis than those with an
aneuploid DNA distribution. [ref: 25,189,381]

Toikkanen and co-workers, [ref: 771] in 351 patients monitored for a
minimum of 22 years, observed a 25-year survival rate of 28% for
patients with nondiploid tumors, in contrast to 48% for those with
diploid DNA pattern.

Ploidy was found to be associated with histologic type, tumor grade,
and SPF values, but not with patient age, menopausal status, tumor
size, axillary nodal status, ER status, or progesterone receptor (PR)
status. [ref: 271]

Diploid tumors also tend to be ER-positive, whereas aneuploid tumors
are frequently ER-negative. [ref: 27] Older patients are more likely
to have hyperdiploid tumors. [ref: 756] DNA ploidy as measured by flow
cytometry correlates with nuclear grade, with low-grade tumors being
diploid and high-grade tumors being aneuploid. [ref: 558] Ewers and
co-workers [ref: 198] assessed DNA content in more than 500 primary
breast cancers; 60% were aneuploid. In patients with stage I or II
tumors, the recurrence rate for aneuploidy was twice that for the
diploid group.

There is conflicting information regarding the value of ploidy
analysis for women with either node-negative or node-positive breast
carcinoma. Data from a few studies seem to indicate that SPF in
combination with ploidy may be more important prognostically than
ploidy alone, although the results have not been uniform. In a study
of 1331 breast tumors reported by Clark and colleagues, [ref: 114] 57%
were aneuploid. The median SPF was 10.3% for aneuploid tumors and 2.6%
for diploid lesions.

In a report by Fallenius and associates [ref: 200] of patients with
negative lymph nodes, the disease-free survival rate was 60% for
patients with aneuploid tumors, compared with 90% for those with
diploid DNA index. On the other hand, Keyhani-Rofagha and colleagues,
[ref: 392] in 165 patients with node-negative adenocarcinomas of the
breast, described 5- and 10-year survival rates of 87.8% and 73.4%,
respectively, in patients with diploid tumors, versus 84.1% and 75.5%
in patients with aneuploid DNA index.

Witzig and associates [ref: 854] examined DNA ploidy in primary tumors
of 265 patients with lymph node-negative breast cancer treated with
surgery and no adjuvant therapy; 130 (49%) had diploid and 135 (51%)
had nondiploid tumors. Ploidy was not significant for either
relapse-free survival (P = 0.20) or overall survival (P = 0.13) or in
any multivariate analysis.

Tumor tissue from 398 patients was analyzed [ref: 226]; in patients
with diploid tumors (43%), the mean SPF was 3.4% +/- 2.3%, and in the
aneuploid population (57%), it was 7.9% +/- 6.3%. Diploid tumors were
more likely than aneuploid tumors to be of good nuclear grade (P <
0.001) and smaller size (P = 0.03). More tumors with high SPF were of
poor nuclear grade than were tumors with low SPF (P = 0.002). No
significant difference in the rates of disease-free survival (P = 0.3)
or survival (P = 0.1) at 10 years was found between women with diploid
or aneuploid tumors. After adjustment for clinical tumor size, the
difference in both disease-free survival and survival rates in
patients with high and low SPF tumors was only 10% (P = 0.04 and P =
0.08, respectively).

Although DNA flow cytometry provides important prognostic information,
standardization of technique and histogram analysis, as well as
quality control, need to be addressed before this technology can be
used reliably on a wide scale. [ref: 540]

Estrogen or Progesterone Hormonal Receptors in Tumor Cells

Several studies have indicated that patients with hormonal receptors
have a significantly higher survival rate. Crowe and associates [ref:
133] studied 1392 patients with carcinoma of the breast who were
treated with modified radical mastectomy. ER-positive tumors (>/= 3
fmol/mg cytosol protein) were found in 1063 patients (76.4%). Their
10-year overall survival rate of 65.9% was significantly better than
the 56% rate in 329 patients with ER-negative tumors (P = 0.0001).

However, this correlation is not consistent. The prognostic
significance of tumor hormone receptor status in premenopausal
patients with node-negative breast cancer was evaluated in 147
patients. [ref: 778] Of the 104 patients (71%) who had ER-negative
disease, 22 (21%) exhibited local or distant recurrence, and 13 died
of breast cancer. Of the 43 patients who had ER-positive tumors, 5
(12%) had recurrences, and 2 died of breast cancer. After median
follow-up of 85.4 months, no statistically significant difference in
disease-free or overall survival rate was demonstrated between the two
groups. Only tumor size continued to have significant prognostic
implications, regardless of ER status.

The apparent discrepancy in some of these reports may be explained by
technical nuances. Gaffney and colleagues [ref: 278] measured estrogen
and progesterone concentrations, DNA ploidy, and SPF by standard
techniques in 124 samples of invasive ductal carcinoma. Suspensions of
tumors cells were examined by immunocytochemical assay for the
percentages of ER- and PR-positive cells. The 20 ER-positive tumors
each contained less than 100 fmol/mg. Sixty-five samples were
considered estrogen positive by radioligand assay. Immunocytochemical
assay showed that tumors with less than 40 fmol/mg did not contain
ER-positive cells. The authors concluded that tumors classified as ER-
or PR-positive based on accepted cutoff values for radioligand assays
may actually be receptor negative, because the tumors do not appear to
contain receptor-positive cells.

Esteban and associates [ref: 194] noted that quantitative
immunohistochemistry of ERs provides results with better predictive
value than the biochemically procured ones.

Patients with negative hormonal receptors have a small probability of
responding to hormonal therapy. [ref: 853] Conflicting reports have
been published regarding their response to chemotherapy. [ref: 454]

Oncogenes

P53

The gene known to be most frequently mutated in sporadic breast cancer
is the tumor suppressor gene p53; alterations of this gene were
identified in 43 of 192 tumors (22%). [ref: 92] Mutations of p53 were
found more often in tumors of younger women (P = 0.002) and
African-American women (P = 0.04) and in tumors lacking ER (P = 0.03),
PR (P = 0.04), or both (P = 0.06). In 843 cases of breast cancer, p53
oncoprotein detected by monoclonal antibody Pab-1801 was not found in
low-grade carcinomas (tubular, mucinous, papillary, and invasive
cribriform types), but it was observed in 4.2% of infiltrating lobular
carcinomas (6 of 140 cases), 15.5% of high-grade invasive ductal
breast carcinomas (99 of 640 cases), and 50% of pure medullary
carcinomas (5 of 10 cases). [ref: 475] The overall survival rates were
not significantly different in patients with mutant or wild-type p53
tumors. In contrast, in another study of 156 patients with primary
invasive breast cancer, overexpression of p53 protein emerged as a
reliable and independent predictor for disease recurrence and reduced
survival. [ref: 270]

Jansson and associates, [ref: 374] in a review of 206 women with
breast cancer, 37 of whom had tumors with mutant p53,] noted better
tumor control and better rates of disease-free and overall survival in
95 women who received locoregional radiation therapy after
quadrantectomy or mastectomy, compared with 73 who did not receive
irradiation. Among the 165 patients with wild-type p53,] the 78
treated with irradiation exhibited better relapse-free survival than
patients without wild-type p53. The locoregional recurrence rate was
6% for the irradiated patients and 23% for those who did not receive
irradiation. The authors postulated that mutant p53 may result in
radio-resistance because of the loss of radiation-induced apoptosis.
The axilla was not specifically irradiated in any of the patients, and
a high rate of regional node failure was observed (46%, primarily in
the axilla).

In another review of 205 cases of stage I breast cancer, 11% showed
accumulation of p53 which correlated with age younger than 50 years,
negative ER/PR status, aneuploidy, high SPF, and high pathologic
grade. [ref: 731] The 5-year disease-free survival without distant
metastasis was 95% in patients with no appreciable p53 and only 73% in
those with significant accumulation of the oncogene. However, on
multivariate analysis p53 was not an independent prognostic factor.

Sjogren and associates [ref: 699] analyzed 316 primary breast tumors
for the presence of mutant p53 protein and found that 69 (22%) had p53
mutations and 64 tumors (20%) had elevated levels of p53 protein
suggesting the presence of mutations. The 5-year relapse-free and
breast cancer-corrected survival rates were significantly lower for
patients with p53 sequence-positive tumors than for those with
negative ones.

In 564 women with breast cancer, there was no evidence that p53 status
by immunochemistry assay could predict response to adjuvant
chemotherapy (cyclophosphamide, methotrexate, 5-fluorouracil [5-FU],
and prednisone); there was a trend for patients with negative p53 to
benefit from chemotherapy. [ref: 188]

HER-2

The HER-2 oncogene (also called c-erbB-2) probably codes for a surface
membrane receptor that interacts with an unidentified growth factor
and is frequently amplified in human breast carcinoma. It is present
as a single copy in normal cells and has been mapped to chromosome 17.
HER-2 amplification or overexpression occurs in almost all cases of
high-grade DCIS but has been reported in only 10% to 40% of
infiltrating ductal carcinoma [ref: 146] and rarely in invasive
lobular carcinoma. Overexpression of HER-2 is believed by some to be
an important independent prognostic indicator in breast carcinoma,
identifying a subset of patients with poor prognosis tumors,
particularly if axillary node metastases are present; however, others
do not share this view. [ref: 701,790]

In 888 node-positive women with breast cancer, 235 of whom received no
additional treatment after surgery, HER-2/neu oncoprotein expression
was associated with poor prognosis. [ref: 763] The study also showed
that HER-2/neu oncoprotein expression is useful in predicting survival
times in patients receiving adjuvant therapy, suggesting that it may
be a marker of drug resistance.

Thor and associates, [ref: 766] in a study of HER-2 oncogene
expression in 313 patients treated for carcinoma of the breast,
described a significant correlation between amplification of the
oncogene and ER status, lymph node status, and survival. For patients
with strongly reactive tumors, the 6-year survival was zero, compared
with 55% in those with negative to weakly reactive tumors. Many
unanswered questions remain regarding HER-2 and its role in breast
cancer development and progression.

Epidermal Growth Factor

Epidermal growth factor (EGF) is a polypeptide hormone that stimulates
cell proliferation through specific binding to a cell surface
receptor. Many oncogenes are closely related to known growth factors
or their receptors, and EGF receptor is homologous to a known oncogene
(c-erbB-1). EGF receptor positivity was present in 2500 of 5232 breast
tumors (48%) in 40 different series of patients. [ref: 396] The EGF
receptor content in breast cancer tissue is inversely related to ER
content and may be a predictor of poor prognosis. [ref: 563,636]

In addition, most studies observed a trend, if no significant
correlation, between higher EGF receptor levels in tumors with the
highest percentages of S phase or Ki-67 expression. With regard to
relapse-free and overall survival rates, five of nine different groups
of investigators showed that patients with EGF receptor-positive
tumors had a poor prognosis (follow-up, 1 to 4 years). However, in
three of five groups with a follow-up of at least 6 years,
investigators found only a tendency for any relation between EGF
receptor status and long-term outcome.

C-myc Gene

Garcia and associates [ref: 284] reported amplification of the c-myc
gene in 18% of 125 primary breast cancers and a significant
association between amplification of this gene and inflammatory
carcinoma.

Bcl-2 Gene

The Bcl-2 gene prevents apoptosis in vitro and in vivo. Moderate to
strong Bcl-2 protein expression (present in 46% of 174 women with
breast cancer) was strongly associated with several favorable
prognostic features, such as a low mitotic count, high histologic
grade of differentiation, lack of p53 protein expression (P < 0.0001
for each), lack of tumor necrosis, low SPF, low cathepsin D
expression, DNA diploidy, and the lobular histologic type, but not
with primary tumor size or axillary nodal status. [ref: 376] Women
with breast cancer and moderate to strong Bcl-2 protein expression had
a more favorable short-term disease-specific survival rate (69% versus
46% at 5 years) but a similar long-term rate (29% versus 33% at 30
years), compared with women with breast cancer and weak or no Bcl-2
protein expression.

CA 15-3 and CA 27-29

CA 15-3 is a widely used tumor marker in carcinoma of the breast, but
its role in the management of patients with early disease is
controversial. O'Hanlon and co-workers [ref: 535] reported on 168
patients with stage I tumors at presentation; mean preoperative CA
15-3 levels at presentation were significantly elevated in patients
with stage I disease compared with patients with benign disease. CA
15-3 levels were not elevated in patients with locoregional disease,
were significantly elevated in patients with bony metastases, and had
a mean lead time of 6.3 months over bone scintigraphy.

In 166 patients with stage II and III breast cancer, CA 27-29 had a
high probability of predicting posttreatment recurrence, with a
5-month average lead time over chemical or imaging detection. Testing
for CA 27-29 was approved for marketing in 1996 (Biomira Diagnostics,
Rexdale, Ontario, Canada).

BRCA Genes

BRCA1 is an important breast cancer susceptibility locus that was
identified in chromosome 17q21 by linkage analysis in familial breast
and ovarian cancers. Of 97 patients studied (including 3 males), 18
had familial breast cancer, 59 had sporadic unilateral breast cancer,
and 20 had bilateral breast tumors and no family history of breast
cancer. A high frequency of allelic imbalance (67%) was manifested at
BRCA1 in the 18 familial breast cancer patients. In the 20 patients
with bilateral tumors, the corresponding percentage was considerably
lower. [ref: 69] A specific BRCA1 mutation, 185delAG, was found to be
associated with breast cancer (21%) in Jewish women younger than 40
years of age. [ref: 243]

Eight breast cancer pedigrees with a high probability of containing
persons with the BRCA1 gene mutation (odds 79.2% to 99.9%) were
identified through genetic linkage analysis using probes located
within q12-22 on the long arm of chromosome 17; 102 female relatives
were successfully typed with one or both of adjacent markers, and 41
were probably not BRCA1 mutation carriers. [ref: 581] Of the remaining
61 women classified as probable BRCA1 carriers, breast cancer was
diagnosed in 35; 13 of these had bilateral disease. Lifetime disease
penetrance of the BRCA1 gene was 88%, and this plateau was reached
earlier (by age 65 years) than estimated by segregation analysis. The
survival curve of patients with breast cancer was less steep for BRCA1
gene carriers than for the general population; survival rates at 5,
10, and 20 years, unadjusted for noncancer deaths, were 83%, 63%, and
41%, respectively. The 5-year survival rate was significantly higher
in BRCA1 carriers than that in an age-matched Scottish population (P <
0.05).

Marcus and associates [ref: 472] concluded that BRCA1-related
hereditary breast cancers are more frequently aneuploid and have a
higher tumor-cell proliferation rate (SPF) than other hereditary
breast cancers. Paradoxically, patients with BRCA1-related hereditary
breast cancer had lower recurrence rates than did other hereditary
breast cancer patients. The latter group had a higher incidence of
tubulolobular group carcinomas, which may be associated with BRCA1
linkage.

In another study, BRCA1 was identified in 10% of 80 women diagnosed
before the age of 35 years who were not selected on the basis of
family history. [ref: 428] Easton and colleagues [ref: 181] estimated
that more than half of women with BRCA1 mutations develop breast
cancer before the age of 50 and approximately 82% by age 70.

One report suggested that women with breast cancer who harbor mutant
BRCA1 alleles may have a longer survival than unselected women with
breast cancer. [ref: 581]

Current techniques on isolation of BRCA1 provide a unique opportunity
to identify persons carrying this gene mutation, estimated to be 1 in
200 to 1 in 400 people in the United States. It is thought that the
BRCA1 gene accounts for 45% of familial breast cancers and, perhaps,
25% of breast cancer cases that occur before the age of 30.

Genetic studies in patients with breast cancer place a second
susceptibility BRCA2 locus in chromosome 13q12-13. Like BRCA1, BRCA2
appears to confer high risk of early-onset breast cancer in women, but
unlike BRCA1, it does not confer a substantial elevated risk of
ovarian cancer. [ref: 860] The risk of breast cancer in men carrying
BRCA2 mutations, although small, is probably greater than in men
carrying BRCA1 mutations.

Approximately 45% of hereditary breast cancer may be explained by
BRCA1, and about 70% of the remaining hereditary breast cancer may be
explained by BRCA2. [ref: 860]

The identification of carriers of this gene will provide new
opportunities for research in genetic and environmental etiology.
Fundamental understanding of the function of BRCA1 and BRCA2 and ways
to exploit that knowledge to diagnose, treat, and ultimately prevent
breast cancer will be important subjects of prospective clinical
trials [ref: 835] and bioethical discussions (with important
implications in eligibility for health care coverage).

Cathepsin D Assay

Cathepsin D is an estrogen-induced lysosomal protease that is
overexpressed and secreted in excess by breast cancer cells, under the
influence of estrogen regulation. Although the level of cathepsin D is
very low or negligible in normal breast tissue, it is elevated in both
benign and malignant ductal proliferative diseases of the breast.
Thorpe and associates [ref: 767] observed that patients with elevated
cathepsin D levels had a shorter disease-free survival time and a
trend toward shorter overall survival. Tandon and colleagues, [ref:
751] in a multivariate analysis, noted that a high level of cathepsin
D was the most important independent prognostic factor in
node-negative breast cancer. Spyratos and co-workers [ref: 727] also
found that the predictive value of cathepsin D was greater in axillary
lymph node-negative patients than in node-positive patients. These
studies, however, used an enzyme-linked immunosorbent assay (ELISA) or
Western Block technique, which in association with the tumor included
inflammatory cells that also contain cathepsin D. Later, Henry and
associates [ref: 345] studied cathepsin D using an immunohistochemical
staining technique, which may be more specific to tumor cells, and
found an opposite prognostic impact, with a significant trend toward
increased overall disease-free and absolute survival times when
cathespin D values were elevated in node-positive but not in
node-negative patients.

In 123 patients with primary breast cancer observed for 5 years, the
cathepsin D level was correlated only with axillary lymph node
involvement. [ref: 586] Univariate analysis showed that high levels of
cathepsin D (> 20 pmol/mg) were correlated with a greater risk of
recurrence and a shorter overall survival time (P < 0.01 and P < 0.03,
respectively). On multivariate analysis, a high cathepsin D level, a
negative PR status, and lymph node involvement were the most important
factors for predicting relapse-free survival (P = 0.02, P < 0.01, and
P < 0.05, respectively).

Apolipoprotein D

Apolipoprotein D is a glycoprotein involved in the human plasma lipid
transport system that is present in large amounts in cyst fluid from
women with gross cystic disease of the breast. Expression of
apolipoprotein D in breast carcinomas was examined in 163 tumors by
immunostaining; 60 carcinomas (36.8%) were negative, 28 (17.2%) were
weakly positive, 33 (20.2%) were moderately stained, and 42 (25.8%)
were strongly positive. Preliminary analysis in 152 women with a mean
follow-up of 42 months confirmed that low values of apolipoprotein D
were significantly associated with a shorter relapse-free survival
time and a poorer survival rate. [ref: 163]

Pepsinogen C

Expression of the proteolytic enzyme pepsinogen C, which is involved
in digestion of proteins in the stomach and is synthesized by a
significant percent of breast cancers (46% of 113 specimens) was found
to be increased in both well and moderately differentiated tumors (89%
and 88.5%, respectively) compared with poorly differentiated cancers
(27.7%). [ref: 818] No prognostic significance has been reported.

Carcinoembryonic Antigen

Carcinoembryonic antigen (CEA) has been reported in 10% to 25% of
breast cancers. In 202 stage I and II breast carcinomas, 113 (56%)
expressed CEA in more than 15% of the cells. [ref: 195] Expression of
CEA was associated with positive ER status (P = 0.003). The only
significant independent predictor of disease-free survival and overall
survival in the ER-negative group was CEA. Multivariate analysis
showed that the association between CEA and ER status was enhanced
further after compensation for other parameters with independent
predictive value.

Prostate-Specific Antigen

Prostate-specific antigen (PSA), which was initially considered to be
specific to epithelial cells of the prostate, was found in breast
tumor cytosol in 30% of 1200 female breast tumor patients. PSA was
produced by cells containing steroid hormone receptors. In a study of
174 women with primary breast cancer treated by modified radical
mastectomy or breast-conserving therapy including irradiation, PSA was
detected in 27% of patients [ref: 866]; the median positive value was
0.12 ng/ml. PSA-positive patients were less likely than PSA-negative
patients to experience breast cancer relapse (11% versus 29%) or to
die of disease (6% versus 19%).

Extrinsic Prognostic Factors

Although the most important parameters in determining prognosis and
patterns of failure are the intrinsic factors, some host factors and
adequacy of treatment are highly relevant.

Age

Young age may be a risk factor for breast recurrence in conservation
surgery and irradiation. Different investigators have used various age
cutoffs such as 50, 40, 35, and 30 years. Vilcoq and colleagues [ref:
817] found a locoregional recurrence rate of 35% versus 4% in women
under or older than 30 years. Kurtz and associates [ref: 415] reported
a 19% incidence of local recurrence in 210 women younger than 40 years
compared with 9% in 1172 older women. This observation correlated with
EIC, high tumor grade, and a major mononuclear cell reaction. The
Harvard Joint Center's inferior results in younger women also
correlated with the presence of EIC. [ref: 595] These findings have
been reported by other authors [ref: 118,153,795,817] (Table 50-10).

De la Rochefordiere and associates, [ref: 150] in a study of 1703
patients with stage I to III breast cancer, noted that younger
patients had significantly lower survival rates and higher local and
distant relapse rates than older patients. A log-linear function
indicated a 4% decrease in recurrence for every year of age.

Fowble and colleagues, [ref: 264] in 980 women with stage I and II
breast cancer treated with breast conservation therapy, reported that
women younger than 35 years of age had a significantly decreased
8-year relapse-free survival rate (53%) than patients 36 to 50 years
(67%) or women older than 50 years of age (74%). Also, the incidence
of breast recurrence was greater in the younger group (24%, 14%, and
12%, respectively; P = 0.001). Regional recurrence rates were 7%, 1%,
and 1%, respectively (P = 0.0002).

Nixon and co-workers [ref: 530] analyzed 1398 women with stage I and
II breast cancer treated with conservation therapy. The rates of
recurrence at 10 years were 51% for 107 women younger than 35 years of
age and 37% for 1026 older women (> 35 to 65 years); the rates of
distant metastasis were 37% and 28%, respectively. However, the
10-year survival rates were 71% and 75%, respectively.

Solin and associates [ref: 721] compared the results of
breast-conserving therapy in 173 women older than 65 years of age with
those of 385 women aged 50 to 64 years. Both groups had a 24%
incidence of positive lymph nodes. Adjuvant systemic chemotherapy was
used in 18% (102 of 558 patients) and tamoxifen in 17% (94 patients).
Median follow-up was 6.2 years. The 10-year relapse-free survival
rates were comparable in the two groups (64% and 70%, respectively).
The local failure rates were 13% and 12%, respectively, but the
intercurrent disease deaths were more frequent in the older age group
(11% versus 2%).

Although age has been studied as a prognostic factor in breast cancer,
little attention has been paid to its role in the selection and
outcome of local therapy. In a review of 42 breast cancer patients
younger than 40 years of age with stage 0, I, or II disease, 37% had
medical contraindications to breast preservation, compared with 25% of
women older than age 40. [ref: 516] Twenty-one percent of younger
women eligible for conservation opted for mastectomy and
reconstruction compared with 9% of their older counterparts; only 4%
of women in either age group selected mastectomy alone as therapy.

Race

Black women are commonly diagnosed with more advanced stages of breast
cancer than white women. Simon and Severson, [ref: 694] in a review of
10,502 women diagnosed with breast cancer (82% white and 18%
African-American), observed that African-American women were more
likely to present with regional or distant disease (45%) than white
women (37%). White women had better survival than African-American
women during the first 4 years after diagnosis (P < 0.0001), but there
were no significant differences in survival by race in women who lived
longer than 4 years (P = 0.64). Black women are more likely than
whites to report that they have not received a mammogram within 3
years before diagnosis. However, history of mammographic screening
accounted for less than 10% of the observed differences in stage at
diagnosis. [ref: 379]

In 887 black women and 265 white women with breast cancer analyzed by
Ansell and associates, [ref: 15] black women had lower 5-year breast
cancer survival rates than white women (50.2% versus 60.2%, P = 0.05),
and survival was lower when adjusted for stage and age. However, when
adjusted for income in addition to stage and age, the effect of race
on survival was reduced (from relative risk of 1.26 to 1.17). Black
women were younger and poorer than white women. There were no
significant differences between blacks and whites with regard to
stage, ER status, or type of treatment.

In 75 black and 615 white women with stage I and II breast cancer
treated with breast conservation therapy and CMF (cyclophosphamide,
methotrexate, and 5-FU), with or without prednisone and tamoxifen, the
5-year actuarial local-only first failure rates were 5% for black
women and 6% for white women (P = 0.53); regional-only failure rates
were 9% for blacks and 1% for whites (P = 0.002), and regional
recurrence as any component of first failure was 16% and 4%,
respectively (P = 0.001). [ref: 573] Distant metastases as the only
site of first failure were significantly greater in the black
population (20% at 5 years versus 11% in white patients) (P = 0.01).
The 5-year overall survival rate for the black patients was 82%,
versus 91% for the white patients (P = 0.01); the disease-free
survival rates being 64% and 83%, respectively (P = 0.0002). Black
patients younger than 40 years of age or with pathologically positive
axillary nodes had significantly worse disease-free, relapse-free, and
overall survival compared with similarly staged white patients.

Eley and associates [ref: 187] reported on a study of 612 black and
518 white women aged 20 to 79 years with primary invasive breast
cancer. After controlling for geographic site and age, the risk of
dying was 2.2 times greater for blacks than whites. Adjustment for
stage reduced risk from 2.2 to 1.7; further adjustment for
sociodemographic variables had no effect. They concluded that
approximately 75% of the racial difference in survival was explained
by the prognostic factors studied. Sociodemographic variables appeared
to act largely through racial differences in stage at diagnosis.

On the other hand, Heimann and colleagues, [ref: 341] in an analysis
of 1277 white and 481 black women with breast cancer treated with
mastectomy, showed that prognosis in white and black women is
comparable when adjusted for tumor size and lymph node status.

Obesity

In a study of 923 women treated by mastectomy and axillary dissection,
those who were obese (25% or more over optimal weight for height) at
the time of primary breast cancer treatment were at significantly
greater risk for recurrence (42%), compared with nonobese patients
(32%), 10 years after diagnosis (P < 0.01). [ref: 680] On multivariate
analysis, obesity remained a statistically significant prognostic
factor after controlling for tumor size, number of positive axillary
lymph nodes, age at diagnosis, and adjuvant chemotherapy with a hazard
ratio of 1.29. Recurrent disease developed in 32% of obese patients
compared with 19% of nonobese women.

Pregnancy

Although in the past it was thought that pregnancy after the diagnosis
of breast cancer was associated with a worse prognosis, recent
evidence suggests the contrary. [ref: 360] This subject is discussed
in detail later.

Multivariate Analysis of Prognostic Factors

Zafrani and associates, [ref: 868] in a study of 434 patients with
infiltrating ductal carcinoma treated with limited surgery and
irradiation, found incomplete surgical excision, lymphatic invasion,
and presence of an extensive in situ component to be pathologic
predictors of local breast recurrence. The corresponding predictors of
survival were these same three features, in addition to size of the
primary tumor and histologic high grade of the tumor.

Kurtz and colleagues, [ref: 410] in 496 patients with stage I and II
ductal carcinoma treated with breast conservation therapy, identified
EIC, tumor histologic grade, and mononuclear cell reaction as
significant prognostic factors on Cox multivariate analysis.

In 280 women with node-negative invasive ductal carcinoma, the tumor
diameter, SPF, histologic grade, and nuclear grade were significant
predictors of disease-free survival, and tumor diameter and SPF had
significant associations with cancer-specific survival. [ref: 495] Cox
analysis showed histologic grade to be the only independent predictor
of relapse, whereas tumor diameter and SPF were independent predictors
of mortality.

Favorable prognostic factors include tumor size 2 cm or less, low
nuclear and histologic grades, low SPF, diploid state, low cathepsin D
level, and positive ER and PR status. [ref: 857]

Prognostic features for 10-year survival were determined from 22
pathologic and 5 clinical variables encountered in 1090 node-negative
and 651 node-positive patients enrolled in NSABP Protocol B-06. [ref:
236] Better survival in node-negative patients was noted for whites
rather than blacks, for patients with favorable tumor types (tubular,
mucinous, papillary) rather than intermediate types (lobular invasive,
classic medullary, and not otherwise specified [NOS]) or unfavorable
forms (NOS pure and atypical medullary), and for tumors with good
rather than poor nuclear grade. Number of nodal metastases, degree of
tumor elastosis, and patient age younger than 40 years or older than
64 years, in addition to nuclear grade and race, were prognostically
significant for node-positive patients.

The best predictor of long-term survival in 399 patients with positive
axillary lymph node infiltrating ductal breast carcinomas was extent
of axillary metastasis (number of axillary metastases, size of the
largest metastasis, and lymph node capsular invasion). [ref: 120] The
mitotic count, tumor grade, primary tumor stage, smooth tumor border,
tumor necrosis, and multifocal primary tumors were weaker but
significant survival correlates. Among patients with small tumors (<
1.8 cm in diameter), those with one lymph node micrometastasis (1 to 2
mm) had a worse prognosis than those with uninvolved lymph nodes of
similar size.

Predictors of distant relapse after conservation therapy for breast
cancer were studied by Tubiana-Hulin and co-workers [ref: 781] in 425
women. Among 21 studied variables, 5 were independent risk factors in
predicting metastasis-free survival: tumor size, clinical and
histologic lymph node status, histologic grade, and PR status.

Analysis by Epstein and associates [ref: 191] of 438 patients with
infiltrating ductal carcinoma treated with conservation surgery and
irradiation showed that the predictors of distant relapse were
comparable to those reported in patients treated with radical
mastectomy.

Based on these prognostic factors, it is possible to identify subsets
of patients who have a low risk of recurrence and would not benefit
significantly from adjuvant systemic therapy (Table 50-11). A concise
review of prognostic factors and their impact on selection for breast
conservation therapy was published by Recht. [ref: 591]

Survival in Untreated Breast Cancer

A classic paper by Bloom and associates [ref: 62] outlined the natural
history of breast cancer in 356 patients seen between 1805 and 1933,
not treated by surgery or irradiation, 250 of whom had a pathologic
diagnosis of cancer. There were no patients with stage I disease, 2.4%
with stage II, 23% with stage III, and 74% with stage IV (Manchester
system). [ref: 559] Survival in the untreated group was compared with
that of a later group of patients treated with radical or modified
radical mastectomy, with or without irradiation. They reported an
overall 10-year survival of 34% in the treated patients and 3.6% in
the untreated group; survival in both groups was dependent on the
histologic grade of the tumor.