Pakha EA Masuda s

Invasive breast carcinoma: Allison KH Penault-Llorca F

General overview Ellis lo

Horii B
Tsuda H
Vincent-Salomon A

Definition Despite their overlapping morphological features, these

The term "invasive breast carcinoma (lBC)" refers to a large and biomarker-defined subtypes show distinct outcomes and
heterogeneous group of malignant epithelial neoplasms of the responses to therapy, as well as differences in their global
glandular elements of the breast. genomic and transcriptomic profiles.

ICD-O coding Localization

None The majority of breast cancers (~90°/o) are unifocal and can
occur in any quadrant of the breast, with higher frequency ir
lcD-11 coding the upper-outer quadrant {230}. A synchronous contralateral
None tumour is found in approximately 2°/o of patients. About 0.1°/o of
breast cancers present as an axillary metastasis with no clear
Plelated terminology breast primary {2208}.
None
Clinical features
Subtype(s) ln non-screened populations, a palpable mass is the most com-
lBCs are categorized into morphologically defined subtypes in mon clin.ical sign of lBC, although skin retract.Ion, nipple inver-
the subsequent sections. It .is important to note that all lBCs are sion, nipple discharge, and (less commonly) a change in the size
grouped into the following biomarker-defined subtypes/groups or shape of the breast or a change in the colour or texture of the
for treatment purposes on the basis of EP and EP882 (HEB2) skin may be seen. In extreme cases, skin ulceration may occur
status: Inflammatory breast carcinoma is a clinical presentat.Ion of breast
• EP-positive, HE82-negative cancer defined by diffuse erythema and oedema involving a thirc!
• EP-positive, HEB2-positive or more of the skin of the breast. All the symptoms of breast cart-
• EB-negative, HE82-positive cer may also be caused by benign breast disease, so evaluation
• EB-negative, HE82-negative. with imaging and histological sampling with core biopsy or FNA
are indicated to establish a definitive diagnosis.

< 26.3 No data

All nghtsreservecl.Thedeslgnatlonsemployed and the presentation of the material in this publlcatlon do not Imply the expression of any opinion whatsoever Data source: GLOBOCAN 2018
on the part of the worid Health organizatlon / Intematlonal Agency for Research on cancer concemlng the legal status of any country, territory, city orarea
or of Its authorltles, or conceming the de!imitatlon of Its frontlers or boundarles. Dotted and dashed lines on maps represent approximate borderlines for
Graph productlon: IARC
(hltp //gco iarc.rr/today)
ti;B!rogr:#:ta,I;:
® Intematlonal Ageney fo.
which there may notyet befull agreement. . World Health organizatiori Research an Cf ncer 201I

Flo.2.72 Estimated age-standardized incidence rates (ASRs; World), per 100 000 person-years, of breast cancer in 2018.

82 Epithellal tumours of the breast

 - screened populations, a spiculated mass with or without
esociated calcif ications is the classic appearance of cancer,
a.I cancers may also be visualized as architectural distor-
nicr. well-circumscribed masses, or calcifications alone. About
=--5°/a of palpable cancers are not seen on mammogram, and
tr€ majority of these can be identified with targeted ultrasonog-
=LEny. Ultrasonography can also be added to improve sensitiv-
IT.. 1 women with mammographically dense breasts, and it is
s, these the method of choice for imaging the breast in women aged
les and < J0 years. The false negative rate of combined mammography
ir global FG ultrasonography is quite low, ranging from 0°/o to 3°/o {1424,
€T6}. Unless the presence of an unequivocally benign diagno-
ss such as a cyst is established on the basis of imaging, tissue
s=ripling is usually necessary to determine that carcinoma is
and can - present.
Flo.2.73 Inflammatory breast carcinoma. Clinically, inflammatory breast carcinoma
uency in WBl is the most sensitive (but not the most specific) method
shows diffuse erythema and oedema involving a third or more of the skin of the breast.
tralateral mr detecting breast cancer; therefore, its use is restricted to
t 0 . i O/o of s=:reening women at very high risk (e.g. carriers of Bfi'CA7 or
no clear 3FIfA2mutations) and in some centres after an initial diagnosis
i obular carcinoma or as part of the staging work-up for breast
-er.
•Vhen the results of the physical examination, imaging, and
lost com- |Eedle biopsy/cytology are all benign and concordant, the risk
ple inver- I missing a cancer is extremely low. However, when any one
in the size I these modalities is non-concordant or cannot be evaluated,
:ure of the reoeat biopsy or surgical diagnostic biopsy is indicated.
lay occur.
i of breast Egivemiology
ing a third i-=ast cancer is the most commonly diagnosed cancer in
reast can- Erales (accounting for 24°/o of all female cancers) and the
3valuation eaaing cause of female cancer death worldwide {221}. Breast
sy or FNA ==ncer accounts for 11.6°/o of cancers in both sexes combined,
-aking it the second most common cancer overall {221,2293,
--74}. Incidence rates of lBC have been increasing in most

cw-and middle-income countries in recent decades. In high-

rcome countries such as the USA, Canada, the United King-
com. France, and Australia, incidence rates decreased in the
aariy 2000s, which was partly attributable to decreased use
I postmenopausal hormone treatment after publication of the
rfu'/3men's Health Initiative trial linking postmenopausal hormone
Lse to increased breast cancer risk {222,485,1804}. However,
•th ageing populations, the global burden increased overall. In
_I18, 2.1 million new cases of breast cancer and 627 000 deaths
•ere estimated worldwide {221}. Incidence also varies by more
ran 10-fold geographically. The groups at highest risk are the
aFuent populations of Australia, Europe, and North America,
•rere 8-9°/o of women are diagnosed with an lBC before the
age of 75 years {631}. Geographical variations, time trends, and
s=jdies of populations migrating from low- to high-risk areas
snow that the risk in migrant populations approaches that of
re host country within one or two generations; this suggests
an important role for environmental factors in the etiology of this
:ancer. Studies of variations in breast cancer subtypes in the
LSA have shown that the frequencies of the clinical/biomarker
sLDtypes vary by population characteristics. In screened popu-
Eons, hormone receptor-positive cancer is the most common
World Health
Organization a.btype, with HE82-positive cancers constituting 10-15°/o and
a`lonal Agency for
ch on Cancer Z018
a-negative/HEB2-negative cancers accounting for 13-17°/o; in
.-screened populations the frequency is different, with higher
Fig,2,74 Invasive breast carcinoma. Digital mammogram showing a typical small
spiculated breast cancer with associated calcifications.
proportions of EB-negative/HEB2-negative cancers (20-40°/o)
and HEP2-positive cancers (15-25°/o) {1604,287}.
Etiology
The origin of breast cancer is multifactorial. Most studies point
to hormones, diet, reproductive factors, and genetics as general
risk factors. From descriptive epidemiological data it has clearly
emerged that breast cancer is a diagnosis of affluent societies
that have acquired the so-called Western lifestyle, characterized
by a high-calorie diet rich in animal fat and proteins, combined
with a lack of physical exercise and obesity, older age at first
childbirth, lower parity, and decreased duration of lactation. The
disease occurs more frequently among women who have an
early menarche, remain nulliparous, or have few children and an
older age at first delivery {1486,1736}. There is also overwhelm-
ing evidence from epidemiological studies that exogenous sex
steroids (estrogens and progestogens) play an important role
in the development of breast carcinomas {908}. Breast cancer
incidence rates increase more steeply with age before meno-
pause (by ~8°/a per year) than after (~2°/o per year) {380}. when

Eplthellal tumours of the breas: 83

ovarian synthesis of estrogen and progesterone ceases and
ovarian production of androgens gradually diminishes.
The consumption of alcohol has been consistently asso-
ciated with a moderate increase in the risk of breast cancer,
particularly hormone receptor-positive cancers {1964,2013}.
Current evidence suggests no causal relationship between
active smoking and breast cancer {1254}; however, one large
meta-analysis has reported a link between active smoking and
increased breast cancer risk for women who initiate smoking
before first childbirth, and it suggested that smoking might play
a role in breast cancer initiation {704}. The relationship between
physical activity and risk of breast cancer has been assessed
by the International Agency for Pesearch on Cancer (lABC),
which concluded that higher levels of activity are associated
with a reduction in risk {2135}.
More than most other human neoplasms, breast cancer
shows familial clustering. Two high-penetrance genes have
been identified (Bf?CA7 and Bf?CA2) that greatly increase the
risk of developing breast cancer. Additional polymorphisms and
genes have recently been identified (primarily via genome-wide
association studies), which are of medium or low penetrance
and confer lower risks. The evidence suggests a polygenic
or.ig.in for th.is disease (see Diagnostic molecular pathology
below).
Specific etiological risk factors are suggested to contribute
differently to the risks of developing the various biomarker
clinical subtypes of lBC {2314,325,572}. Germline Bf?CA7 muta-
tions are associated with risk for triple-negative (EB-negative
PB-negative, and HEB2-negative) cancers, whereas germlin=
Bf7CA2 mutations are associated with hormone receptor-pos -
tive breast cancers {403,754}. Early menarche, late menopause
postmenopausal hormone replacement therapy, nulliparit-..
and older age at first childbirth are associated with increases
in the risk of hormone receptor-positive breast cancer {66,12€
1105}. Parity may be associated with an increased risk of triplE-
negative tumours {325,126}. Longer duration of breastfeec-
ing is associated with a decrease in the risk of all subtypes
A positive association between body mass index and breas-.
cancer risk among postmenopausal women is stronger for hor-
mone receptor-positive cancers. A weak inverse associatio-
has been suggested among premenopausal women, in whom
higher body mass index is possibly associated with decreased,

Molecular subtype
Mammary epithelial cell Genetic CScs and
of bulk tumour
hierarchy (cell of origin) eve nts d iffe renti atio n
(clinical)

Mesenchymal
tumour cells

Mesenchymal

®
CS

/Ma,SC
-8 €ay
-',

®CD
Myoepithelial Common luminal
progenitor progenitor

Luminal
progenitor
(ER+/ER-)

'
J

a,I,II
J

Myoepithelial Alveolar Ductal cells

[::;/8H+E2;yozyo-87%]
cells cells (ER-) (ER+/ER-)

Fig. 2.75 A schematic diagram of the presentation of pathogenesis of breast carcinoma deduced from the combination of a model of mammary epithelial cell hierarchy as cell o;
origin, relationships between the hierarchy and molecular subtype of bulky tumour based on gene expression profiling, genetic events that are frequent or characteristic in eac+
subtype, and a model of varying proportions of cancer stem cells (CSCs) in mesenchymal versus epithelial status, as well as differential blocks in the differentiation hierarch,t
seen in normal mammary development in the various subtypes of breast carcinoma {1674,574A,2181A,237A,1223A,2204A}. Masc, mammary stem cell; TNBC, triple-negative
breast cancer.

84 Epithelial tumours of the breast

ygenlc
hology'
ltribute
marker/
7 muta-
gative,
ermline
r-posi-
Pause,
liparity,
3reases
66,126,
)f triple-
lstfeed-
btypes.
I breast
for hor- I+2.76 Breast carcinoma. A Gross appearance. I Invasive breast carcinoma of no special type (NST). Note the irregular stellate outline and the central scar.
ociation
-sK of hormone receptor-positive cancers but increased risk of
n whom
creased Fple-negative cancers {126,2314}. There is some evidence that
acohol, smoking, and physical activity are associated with risk
]f hormone receptor-positive cancers {571}. However, studies
r specific ethnic populations reveal that there may be popu-
ajon/ethnicity-based differences in risk factors for developing
ssecific biomarker/clinical subtypes of breast cancer (1178}.
Pthogenesis
--e pathogenesis of breast cancer follows several pathways.
`Iultiple linear models of breast cancer initiation, transformation,
and progression have been described that are based largely
an hormone receptor status and morphology. The EB-positive
-odel recognizes flat epithelial atypia, atypical ductal hyper-
aasia, and EP-positive ductal carcinoma in situ (DCIS) as the
Ton-obligate precursors of invasive and metastatic breast car-
=noma. The EB-negative model recognizes EB-negative DCIS
=nd microglandular adenosis as precursors of EB-negative
ancers (192}.
At the cell-of-origin level, two leading models accounting
i3r breast carcinogenesis have been described: the sporadic
=onal evolution model and the cancer stem cell model. At the
-olecular level, there is evidence that strongly suggests that
=reast cancer evolves along two divergent molecular pathways
I progression, mainly related to hormone receptors. Molecular
Eta have also demonstrated that EP-positive and EB-negative
]east cancers are fundamentally distinct diseases, and that
•tthin EB-positive cancers, histological grade and proliferation
ae strongly associated with the extent, complexity, and type of
g3netic aberrations {1237,68}. The first pathway (the EB-positive
cathway) is characterized by gains of lq, loss of 16q, infrequent
=rTiplification of 17ql2, and a gene expression signature that is
=redominantly populated with genes associated with the EP-
=ositive phenotype. These lesions express hormone receptors,
ask HE82 overexpression and expression of basal markers,
Fd have relatively simple diploid or near-diploid karyotypes.
r general, this pathway consists of neoplastic lesions pre-
irchy as cell of sominantly of the low- to intermediate-grade phenotype, in
:eristic in each
acdition to a small subset (~9°/a) of morphologically defined
ition hierarchy
lgh-grade tumours. The second pathway, designated the EB-
triple-negat.ive
-egative pathway, is most commonly characterized by loss of
13q, gain of chromosomal region llql3, amplification of 17ql2,
and a gene expression signature populated by genes associ-
ated with cellular proliferation and cell-cycle processes. This
pathway consists predominantly of morphologically defined
intermediate-and high-grade tumours {1237}. P/K3CA muta-
tions commonly occur in both pathways, and 7P53 mutations
are frequent in the EP-negative pathway {277}. Published data
also suggest that the progression of a low-grade tumour to a
high-grade tumour may preferentially occur in breast cancers
of the luminal phenotype {1470}. The EB-negative breast can-
cers include both HEP2-positive and HE82-negative groups.
In these groups (all of which are mostly high-grade, genetically
unstable, and mostly aneuploid), rp53 mutations are common.
In the EB-negative, HEB2-positive group, P/K3CA mutations are
also very frequent, in addition to 17ql2 amplification. Important
elements of EP-negative, HEP2-negative breast cancer biology
include high proliferative activity, an increased immunological
infiltrate, a basal-like and a mesenchymal phenotype, and defi-
ciency in homologous recombination.
Macroscopic appearance
Most lBCs can be visualized or palpated as a grossly evident
mass, with an irregular, stellate outline or nodular configuration.
The tumour edge is usually moderately or poorly defined and
lacks sharp circumscription. Classically, lBCs are firm or even
hard on palpation, and they may have a gritty feel when cut
with a knife (although some types of DCIS, such as the comedo
type, can have a similar gritty texture and cannot always be eas-
ily distinguished grossly from invasion). However, some lBCs,
including neoadjuvant-treated cases, may be grossly inap-
parent and require careful correlation with imaging at the time
of gross examination and tissue sampling. Gross evaluation
should include review Of imaging findings (jf performed), so that
the number of lesions, expected size of lesions, and expected
clips/markers and their location can inform appropriate tissue
sampling {40}. Padiography of larger specimens can be helpful
in identifying clips, calcifications, and other clues to guide mac-
roscopic examination {1033,132,1643}. Ideally, surgical speci-
mens should be differentially inked and serially sectioned into
approximately 0.5 cm thick slices to ensure adequate fixation
and detection of smaller invasive cancers {1965,1156}. Tissue
Epithellal tumours of the breast 85
sampling should be performed such that the T stage of the can-
cer can be accurately established {1965,1156,573}. Sampling
between closely related gross lesions is also important so that
the pathologist can document whether they are truly multiple
or one larger cancer, for staging purposes. A cassette map of
larger lesions is often required when the extent or size is not
macroscopically obvious {768,1867,434}. This is especially
true after neoadjuvant therapy, because residual cancers can
become softer and more difficult to palpate, and only a fibrous
scarred area may be present (except in cases of absent or mini-
mal response) {204,1677,2017}. Another setting that requires
extensive tissue sampling is cases of extensive DCIS, so that
any foci of invasion can be identified {1155}.
Macroscopic assessment of skin ulceration, nipple changes,
and separate skin nodules (if present) are also important for
staging. Distances of grossly apparent cancers from the resec-
tion margins should be recorded and close margins sampled for
microscopic analysis. Lastly, tissue handling parameters such
as the time the specimen was removed by the surgeon, the time
it was placed in fixative, and the amount of t'ime between fixa-
tion and tissue sample processing should be recorded so that
the cold ischaemia time (the amount of time between removal
from the body and placement in fixative) and the time in fixative
before processing are clear, because these parameters can
affect receptor (EB/PB and HE82) testing, grading, and lym-
phovascular invasion assessment {573,783A}.
Histopathology
lBC has a broad spectrum of histological appearances. Four
histological features are described to further characterize the
biology: (1) histological subtype based on tumour architecture
cytonuclear features, and stromal features; (2) Nottinghar
grade (detailed below); (3) the presence or absence of spreac
in the angiolymphatic spaces (only peritumoural lymphovascu-
lar invasion is assessed in breast cancer, and this should be
differentiated from tissue retraction); and (4) an associated ir
situ component. Other features of importance in the classifica-
tion include tumour size, distance to margins, stromal changes
and tumour-infiltrating lymphocytes (TILs). It is also important tc
recognize when the histological features of a breast carcinoma
are unusual or discordant with the hormone receptor or HEPl2
status {37}. For example, a low-grade cancer with either a nega-
tive EP status or a positive HEB2 status would be highly unu-
sual/discordant, and further work-up would be recommendec
to ensure accurate histological typing, grading, and biomarke.
status. In the subsequent sections of this chapter, lBCs are

Nottingham Grading Examples: Nuclear Pleomorphism

Moderate increase in
Normal size + variabilfty = Score of 2 Marked variation = Score of 3
-.,... +_ I `,i+-±ilh,\' . _I. ''.:`-.`--.`;-}':-a-#ri.'J .... i:.3 '.',= '..€',
I.±Effi¥±ij_-i:iii-.Ei---i--_-:iE¥±±±±ljli:-i-i:ii-REiii-ii-ii:i-i:-`i-RE H

•.-i,:.-.,....:...-..-1-.i.,.±-,:.,,...`:;:.;:

Moderate increase in
Small, regular uniform = Score of 1 size + variability = Score of 2 Marked variation = Score of 3

\*:to~!, r,.
", `:f`.:.,P

I,. ffi#
•f -I =ah k`" * ``H3 ±d£3i*±* SF _"-

Flo.2.77 Nottingham grade of invasive breast carcinoma: nuclear pleomorphism scoring.

86 Epithelial tumours of the breast

ters can
lnd lym- Nottingham Grading Examples: Tubule Formation
Majority (>75%) = Score of 1 Moderate (1 0-75%) = Score of 2 Little or none (<10%) = Score of 3
;i?+;=4v..:Tt:if*=:.. ``; ±{.. -',iir#.ngse.-,`=r.-,%, :`-. >= ...,--_: : -I I,., a..wh- . :-: ZFE: I '. •~
`T .=asprr-I ` ir= -,-- t-r- ® `.Lm (A1=:_T\ --tti]-3--< -..-,- ~.ri` -" ---. ` ndrfu -- _. ~~._ _

;es. Four
erize the
litecture,
ttingham
)f spread
hovascu-
hould be
)ciated in
;lassifica-
changes
Dortant tc
arcinoma
or HEB2
)r a nega-
ghly unu-
nmendec
)iomarke.
IBCs are

-Zn Nottingham grade of invasive breast carcinoma; tubule formation scori.ng.

mmiLF zed by their morphological subtypes, although the major- and nucleoli that are either not visible or very inconspicuous.
„mu, ± cases are of no special type (NST). Score 2 nuclei are larger (1.5-2 times the size of benign epithe-
lial cell nuclei), with mild to moderate pleomorphism and visible
thas5rty ical type but small and inconspicuous nucleoli. Score 3 nuclei are even
carcinomas showing a special histological pattern in
an of the tumour are designated as a pure special tumour
Tabl®2.06 Semiqualitative method for assessing histological grade in breast tu-
such as lobular, mucjnous, and tubular carcinomas. mours (585}
s lacking such specific features are designated as jnva-
carcinoma NST, which accounts for the majority of cases, Feature Score
those with mixed patterns. Tubule and gland formation

Majority of tumour (> 75%)

ical grading
characteristics are evaluated: tubule formation as an Moderate degree (10-75o/o)

ion of glandular differentiation, nuclear pleomorphism, Little or none (< 1 oo/a)

"totic count {585,1714}. A numerical scoring system from
Nuclear pleomorphism
3 s used to ensure that each factor js assessed jndepen-
Small, regular, uniform cells
isee Table 2.06).
Tmues and glandular/acini formation are assessed through- Moderate increase in size and variability 2
whole tumour at low magnification; only structures exhib-
Marked variation 3
clear central lumina surrounded by polarized neoplastic
Mitotic count
are counted. Cut-off points of 750/o and loo/o of glandular/
area are used to determine the score (see Fig. 2.78). Dependent on microscope field areaa 14
musar pleomorphism is assessed in the area showing the
Total Score Final grading
degree of pleomorphism in comparison with the regu-
3i the nuclear size and shape of normal epithelial cells Add the scores for gland formation,
nuclear polymorphism, and mitotic count:
wirfent breast tissue (see Fig. 2.77). Increasing irregular-
fff riLclear outlines and the number and size of nucleoli are us Grade 1
additional features in allocating scores for pleomor-
6 or7 Grade 2
Score 1 nuclei are very similar in size to the nuclei of
ere-existing epithelial cells (< 1.5 times the size), and 8or9 Grade3

I sriL`w minimal pleomorphism, an even chromatin pattern, asee Table 1.01 (p. 6).

Epjthelial tumours of the breast 87


..I??t`:.2`i;
Fig. 2.79 Invasive breast carcinoma. Invasion of the angiolymphatic space.
larger (> 2 times the size of benign epithelial cell nuclei), with
vesicular chromatin; they vary markedly in size and shape and
often show prominent nucleoli. The preferred magnification for
nuclear scoring is 40x objective.
The evaluation of mitotic figures requires care and relies on
optimal tissue fixation and good preparation of sections. Observ-
ers must count only definite mitotic figures; hyperchromatic and
pyknotic nuclei should be ignored because they are more likely
to represent apoptosis rather than cells in mitosis. Mitotic counts
Hormone receptor staining interpretation (ER and PR)
Evaluate overall percentage of cancer in sample with nuclear staining and intensity of stain
'.` '.-: ....::. ''`;'y`# I" :fit i...::Yin. : `.J „.>=i| .:T-..t:` I:u`.`S?,.®q?ro?.{.:,S." ,
Interpretation: Positive*
(include % and intensity in report)
*Report as low positive if 1-10% of cells stain
FIg. 2.80 Invasive breast carcinoma. Hormone receptor staining interpretation (EPl and PP).
88 Epithelial tumours of the breast
require standardization to a fixed field area, because field area
varies between microscopes. The total number of mitoses per
10 high-power fields is recorded. The cut-off points for scoring
depend on the field area size, and the microscope used should
be calibrated by measuring the diameter of the high-power field
(40x objective) (see Table 1.01, p. 6). Scoring is performed on
the area exhibiting the highest frequency of mitotic figures (the
hotspot method), typically the peripheral leading edge of the
tumour. If there is heterogeneity, regions exhibiting a higher freL
quency of mitoses should be chosen. Once the hotspot is choc
sen and the first field is considered, subsequent field selection
is by random meander through the chosen area, but only fields
with a representative tumour cell burden should be assessed.
The three values are added together to produce a score Of
3-9, to which grades are assigned as follows: 3-5 points =
grade 1, well differentiated; 6-7 points = grade 2, moderately
differentiated; and 8-9 points = grade 3, poorly differentiated_
For the purpose of quality assurance, it is recommended thai
the individual score components be reported in addition to the
calculated grade. The grading of small tissue samples such
as core needle biopsy (CNB) specimens is possible and often
necessary in the era of neoadjuvant treatment, but it should be
recognized to have limitations, in particular due to the inherem
reduced ability to accurately assess mitotic frequency. This may
lead to underestimation of the true grade in such specimens.
..`,:givrff#ft.
.,
...,. ¢` - `-~.-:-
?fi:'-,.'i,3.ii:.:
I nterpretation : Negative
(note whether result was < 1 °/o or 0%)
:ield area `mmunohistochemistry
.oses per 'BC cells are generaHy positive for low-molecular-weight cyto-
ir scoring reratins (CK7, CK8/18, CK19), EMA, E-cadherin, BCL2, and
;d should SATA3 (however, poorly differentiated forms may lose expres-
)wer f ield son of one or more of these markers). A proportion of lBCs
)rmed on ?.fpically the better-differentiated forms) are also positive for
lures (the 3CDFP-15, mammaglobin, milk fat globule, Iactalbumin, CEA,
ge Of the and 872.3. Approximately 300/o of breast cancers express
igher fre- ae or more basal markers, including high-molecular-weight
ot is cho- =..tokeratins (CK5/6, CK14, CK17) and EGm (HER1), and these
selection ~arkers are more frequently positive in EB-negative tumours.
nly f ields S-00 is expressed in a proportion of breast cancers but also jn
sessed. re myoepithelial cells. Breast cancers are negative for CD34
. score of and typically negative for CK20 and p63 (although exceptions
points = eecur, such as p63 expression in metaplastic and salivary
oderately gand-like carcjnomas). Most breast cancers are negative for
rentiated. rier tissue-specific markers, such as CDX2, PAX8, WT1, TTF1,
|ded that thl845, melan-A, CD20, and CD3, but may be positive for
ion to the I:J138.
)les such
and often tthone receptors: Nuclear EB expression should be evalu-
Should be and in IBC because of its utility in predicting clinical benefit
3 inherent in endocrine therapy and its use in various clinical treatment
This may aprithms for ER-positive versus EF]-negative cancers (see
)imens. Fg. 2.80). in expression tends to vary more than ER expres-
art. and this helps account for the effectiveness of PB to
further stratify EB-positive cases into prognostic categories,
Efrough specific PB thresholds for this purpose are not well
anated (see Prognost's ancy prec};.cfi.on, below) {1081}. The [1g.2.81 AIgorithm tor interpretl.ng EF1882 (HEF12) immunohistochemical staini'ng in
2:10 American Society of Clinical Oncology (ASCO) / College invasive breast carcinoma. ISH, in §itu hybrjdization.

I American Pathologi.sts (CAP) guideline recommendations for

rmunohjstochemical testing of EB and PB in breast cancer
sine that breast cancers with as few as 1°/o of cells with weak
rrmsity of EP staining should be considered positive, because
I evidence of potential benefit from endocrine therapy, and the
2Dl9 update to these guidelines has maintained this threshold
E31.1394,815,816}. However, invasive cancers with 1-loo/o EB
asltivity should be reported as EP low positive, with additional
lmeapretation and reporting recommendations. There are more-
hited data on the benefit of endocrine therapies in this group,
ar they suggest possible benefit from endocrine therapies, so
©nts are considered eligible for this treatment. However, this
group is noted to be heterogeneous, and the biological behav-
ar of EB low positive cancers may be more similar to that of
anegative cancers. This should be considered in decision-
marlngforotheradjuvanttherapyandoveralltreatmentpathway/
±ification. An external control should always be evaluated
I,I,urn appropriately positive and negative tissues), and internal
Grmols should also stain as expected when present. See the
mum recent version of the guidelines for up-to-date information
amit test interpretation (60A}. There can be regional variability
im tormone receptor expression, and the percentage reported
igrnild reflect the percentage of positive cells jn the entire sam-
ue of the invasive cancer tested (not just the area of highest
g=fession). When the intensity is variable, it can be reported
aE a range or an average intensity, with many systems using
or 0%) a ijige of 0,1+, 2+, and 3+. Different scoring systems can be test performance and interpretation recommendations are
uE5ed to combine the intensity and percentage information for
IT~ overall score (e.g. the H-score or Allred score). In addition
urn cercentage and intensity results, an interpretation as EP/
PP-positive or -negative should also be included in the report.
Low-power scanning is often insufficient for the detection of
focal or weak levels of hormone receptor expression; therefore,
in cases that appear negative on low power, a higher-power
scan of the slide js also appropriate, to rule out weak hormone
receptor positjvity.
HEP2: HEB2 js a member of a family of growth factor recep-
tors -also including EGFB (HEB1), EP883 (HEB3), and E8884
(HER4) -that regulate normal cell proliferation, development,
and survival. HEB2 is located on the cell surface at low lev-
els in normal breast epithelium. In 10-20°/a of lBCs, the Ef?882
(HEf?2) gene is amplified, resulting in overexpression of the
HEB2 protein at the cell surface. This protein overexpression
can then result in the promotion of more-aggressive cancer
biology due to increases in cancer proliferation, cell motility,
and angiogenesis. There are a number of HEB2-targeted thera-
pies available today, some of which are used in combination
with the first (and still standard) anti-HER2 biologic therapy,
trastuzumab (Herceptin) {1564A,1627}. HER2 testing is required
on any new lBC-NST, because positive cases can be treated
with HEF}2-targeted therapies jn addition to chemotherapy,
with significant increases in survival. HEP2 protein overexpres-
sion can be assessed using jmmunohistochemistry, or Ef7882
amplification can be identified by jn situ hybridization. Detailed
available in country-specific guidelines, for example those pub-
lished by ASCO/CAP {2285,2284}. Because immunohistcohen-
istry is readily available, with a lower cost, many lat"atories

Epithelial turmoLrs of the c'eas: 89

Tal]le2.07 Standard required biomarkers in invasive breast cancer: purpose, reporting, and scoring criteria {2285,2284}

Test Pleporting
Biomarker & purpose Scoring criteria (ASCO/CAP)
type catego ries

ER
Validated for:
Prediction of benefit from hormone Positive 21% of invasive cancer has nuclear staining of any intensity

therapies if positive

Other uses: lHC

Categorization for overall treatment

pathways < 1°/a or 0°/o of invasive cancer has nuclear staining (follow proper QA and most-recent guidelines to
Characterization as the lHC luminal group Negative
ensure not a false negative result)
if positive
Poor prognostic marker if negative

PR
Validated for:
Primarily prognostic in ER-positive Positive 21°/o of invasive cancer has nuclear staining of any intensity

cancers (not well-validated for prediction

of endocrine therapy benefit)
lHC
Other uses:
Poor prognostic marker if negative (in
< 1 a/a or 0°/o ot invasive cancer has nuclear staining (follow proper QA and most-recent guidelines to
EPl-positive cancers) Negative
ensure not a false negative result)
Further characterization of the lHC group
subtype

ASCO, American Society of Clinical Oncology; CAP, College of American Pathologists; lHC, immunohistochemistry; QA, quality assurance.

use immunohistochemistry as the first test performed; if there

is an equivocal result (2+ staining), the case is automaticalL,\
referred for in situ hybridization (either in-house or at a reference
laboratory). The 2018 update of the ASCO/CAP HER2 testing
HER2-positive guidelines recommend greater emphasis on using immunc-
histochemistry to guide interpretation and testing of unusua
in situ hybridization categories (so-called groups 2, 3, and 4
See Fig. 2.81 (p. 89), Fig. 2.82, and Table 2.07 for the inter-
pretation criteria for HEP2 immunohistochemistry and in sit.
hybridization. The evaluation of stain intensity is also critical t: ffl
HER2-negative
the appropriate determination of HEB2 immunohistochemica
status, with overinterpretation of stain intensity being one of the
most common causes of discordance between HEB2 immir
nohistochemistry (positive) and in situ hybridization (negativet
unusual HER2 lsH result categories requiring additional work-up {766}. The comparison of stain intensity with negative (0-1+
and positive (3+) controls should serve as a reference for inter-
sity measures. A practical approach is to use the magnifica-
tion rule, which takes into account the magnification of the lens
required to assess the staining, as follows: using a lox ocular
if staining is noted with a 2-5x objective, it can be considerec
strong intensity; staining detected using a 10-20x objective is
Review concurrent lHC from the same sample considered weak to moderate; and staining detectable only o-
40x is considered faint / barely perceptible. Interpreters shoulc
be aware of unusual staining patterns (e.g. clustered hetero-
geneity) and discordant results (e.g. a grade 1 cancer with 3-
HEP2 staining) {37}. Laboratories and pathologists assessing
HEP2 immunohistochemistry and/or Ef?882 in situ hybridizatior,
should participate in external quality assurance processes to
ensure the accuracy of results.

Histology of hormone receptor-positive cancers: Hormone

Flo.2.82 Dual-probe EB882 (HEf72) in situ hybridization (lsH) test interpretation.
CEP17, chromosome enumeration probe 17; lHC, immunohistochemistry. *As deter-
mined by concurrent lHC and lsH. Pleport comments recommended; see American
Society of Clinical Oncology (ASCO) / College of American Pathologists (CAP) guide-
lines for details {2284},
receptor-positive breast cancers have a spectrum of morpholo-
gies and grades. They range from well-differentiated cancers
composed of bland cells forming tubular/ductal structures to
poorly differentiated cancers composed of cells with substantial

90 Epithelial tumours of the breast

 "e2.07 Standard required biomarkers in invasive breast cancer: purpose, reporting, and scoring criteria {2285,2284} (continued)

-...- I i .-.....-...---...-..- :h.i.- `

B:` : i
3+: Circumferential membrane staining that is complete, intense, and in > loo/a of tumour cells

Equivocal 2+: Weak to moderate complete membrane staining observed in > loo/o of tumour cells
lHC
1+: Incomplete membrane staining that is faint / barely perceptible and in > too/a of tumour cells
Negative 0: No staining is observed, or incomplete membrane staining that is faint / barely perceptible and in
i 10% of tumour cells

Negative Fg7oBu:25()rm2)/CEP17 ratlo < 2.OAND average Ea882(ma4 copy number <4.o signals perceii

Efl882 (Hffl2)/CEP17 ratio 2 2.0 AND average Ef?882 (Hffl4 copy number < 4.0 signals per cell
ER882 (HER2) (group 2) and concurrent lHC 0,1+, or 2+; or
rdidated for: Negatlvea (:PoBu%23()Ha:E2!/ocnEc::r7ernattj.°H€2o.Oof,N+: :rverage EP882 (Hffl4 Copy number 2 6.0 signals per cell
>ediction of benefit from HER2-targeted

rerapy if positive (administered with
-Otherapy)
fror uses:
:alegorization for overall treatment
anays
3iaracterizatlon as the HER2-enriched
ut subtype (if ER-negative) or luminal 8
.i =Fl-positive)
uE:rker of aggressive biology
Ef]882 (Hffl2)/CEP17 ratio < 2.0 AND average Ef?882 (Hffl4 copy number 2 4.0 but < 6.0 signals
Dual-
per cell (group 4) and concurrent IHC 0,1+, or 2+
probe ---
lsH Ef]882 (Hffi2)/CEP17 ratio 2 2.0 AND average EPI882 (Hffl4 copy number < 4.0 signals per cell
(group 2) and concurrent IHC 3+; or
Positlvea (:Fo%23()Ha:Eac/ocnEc:|r7ernatt',°H€22.: :rN3D+:::rage ffi882 (Hffi4 Copy number 2 6.o signals per ceN
EP882 (Hffl2)/CEP17 ratio < 2.0 AND average Ef3882 (Hffl2) copy number 2 4.0 but < 6.0 signals
per cell (group 4) and concurrent IHC 3+
Ef]882 (Hffl4/CEP17 ratio 2 2.0 AND average Ef?882 (Hffl2) copy number 2 4.0 signals per cell
positive (groupl)

I; if there Average Ef]882 (Hffl2) copy number < 4.0 signals per cell; or
•matically
•eference Negative A,::::: Eggg2(:Egg :::; ::::::2 ::: :::: ::: :i:::i: :::::i: ::: :::::::::|Hu:,,po:!:; or
Single-
12 testing lsH group 5
probe --
immuno- lsHb Average Ef?882 (Hffl2) copy number 2 6.0 signals per cell; or
: unusua Average Ef]882 (HEfl2) copy number 2 4.0 but < 6.0 signals per cell and concurrent IHC 3+; or
Positive
), and 4) Average Efl882 (Hffl2) copy number 2 4.0 but < 6.0 signals per cell and concurrent dual-probe
lsH group 1
the inter-
ld in sit,. flsc€. American Society of Clinical Oncology; CAP, College of American Pathologists; lHC, immunohistochemistry; lsH, in situ hybridization.
C± 3jal-probe groups 2L4, the final lsH results are based on concurrent review of lHC, with recounting of the lsH test by a second reviewer if lHC is 2+ (per the updated 2018
critical tc
JS:a CAP recommendations {2285,2284}).
ichemica
Comment for group 2 negative result: "Evidence is limited on the efficacy of HEPl2-targeted therapy jn the small subset of cases with a HEF2rcEP17 ratio 2 2.0 and an
)ne of the
average Hffl2 copy number of < 4.0 per cell. In the first generation of adjuvant trastuzumab trials, patients in this subgroup who were randomly assigned to the trastuzumab
]2 immu-
arm did not seem to derive an improvement in disease-free or overall survival, but there were too few such cases to draw definitive conclusions, lHC expression of HER2
(negative should be used to complement lsH and define HEPl2 status. If lHC result is not 3+ positive, it is recommended that the specimen be considered HEPl2-negative because of
ve (0-1+ the low HEfl2 copy number by lsH and the lack of protein overexpression."
I for inter- Comment for group 3 negative result: "There are insufficient data on the efficacy of HE82-targeted therapy in cases with a HEfl9CEP17 ratio of < 2.0 in the absence of
lagnifica- protein overexpression because such patients were not eligible for the first generation of adjuvant trastuzumab clinical trials. When concurrent lHC results are negative (0 or
1 +), it is recommended that the specimen be considered HE82-negative."
>f the lens
Comment for group 4 negative result: "lt is uncertain whether patients with an average of 2 4.0 and < 6.0 HEPl2 signals per cell and a HEf32JCEP17 ratio of < 2.0 benefit
)x ocu'a'
from HER2-targeted therapy in the absence of protein overexpression (lHC 3+). If the specimen test result is close to the lsH ratio threshold for positive, there is a higher
3nsidere= likelihood that repeat testing will result in different results by chance alone. Therefore, when lHC results are not 3+ positive, it is recommended that the sample be considered
3jective s HEPl2-negative without additional testing on the same specimen."
e Only 0- fy single-probe lsH: lt is recommended that concomitant lHC review become pan of the interpretation of single-probe lsH results. ASCO/CAP also preferentially recommends the
3rs shoul= IE I; dual-probe rather than single-probe lsH assays.
3d heterc-
er with 3-
assessir: rui=,ear pleomorphism and more sheet-like growth. However, atypia, as well as lobular in situ lesions, these lesions are fre-
3ridizatio: imest breast cancers that are strongly EP-positive and HE82- quently identified in the background of EB-positive lBC-NSTs,
icesses := imegative are in the low-to intermediate-grade spectrum. Of note, and they sometimes account for a large proportion of the initial
iaacers with low percentages of EB-positive cells (and HE82 imaging findings. Cancers associated with germline Bf?CA2
imEgativity) frequently have histological features more similar to mutations are often EB-positive {1099,1057}.
Hormone mcse of high-grade triple-negative cancers {1913}. Because
morpholc- ne Pathway to typical strongly EPl-positive breast cancers Histology of triple-negative cancers: Although triple-negative
d Cancer-I imaudes other non-obligate precursor/risk lesions with high EB (hormone receptor-negative, HEP2-negative) breast cancers
uctures I m3ression, such as DCIS (predominantly of low to intermediate can also have a spectrum of morphologies (including spe-
5ubstant,a iruc:ear grade), atypical ductal hyperplasia, and flat epithelial cial types such as adenoid cystic carcinoma and metapiastic

Eplthelial tumours of the breas: 91


% ofbreastcancers:!| 15-20%
Receptor expression
Histologic grade:
Recurrence risk:
Therapies used:
Fig.2.83 Correlation of breast cancer molecular subtypes with clinicopathological
features.
carcinoma), the majority of these lBCs are high-grade, with
tumour cells displaying high N:C ratios, a solid growth pattern,
frequently pushing borders, and geographical necrosis {266,
1222}. Mitotic counts are typically very high (with proliferation
often > 80°/a). The lesions can have a central scar and/or a large
central acellular zone, and some triple-negative cancers have
a dense lymphocytic stromal infiltrate, which is associated with
better response to treatment. DCIS is less frequently identified
in the background of triple-negative cancers than in hormone
receptor-positive cases {1702,1713}. Cancers associated with
germline Bf?CAT mutations are frequently of this morphology
{1099}. As noted above, cancers with low percentages of EP-
positive cells (and HEP2 negativity) frequently have histologi-
cal features more similar to those of high-grade triple-negative
cancers {1913}.
Histology of HEPl2-positive cancers: These breast cancers
are typically of high histological grade, and they are frequently
associated with background high-grade comedo DCIS. HEB2-
positive cancers usually have very pleomorphic nuclei and
more-abundant eosinophilic cytoplasm than triple-negative
cancers, sometimes giving cells a vaguely apocrine appear-
ance. Proliferation is high, but lower on average than in triple-
negative cancers (in the 20-60°/a range). The typical growth
patterns are infiltrative cords or solid nests of ceHs, but single-
file patterns are occasionally seen.
Stromal response patterns and tumour microenvironment
The stromal component is extremely variable. There may be
a highly cellular fibroblastic proliferation, a scant element of
connective tissue, or marked hyalinization. Foci of elastosis
may also be present in a periductal or perivenous distribution.
Some lBC-NSTs show a fibrotic focus, defined as an area of
exaggerated reactive tumour stroma formation > 1 mm within
the tumour {832,1276,2143}, with or without coagulative necro-
sis {2143,1276}, and these cases have been reported to show
more-aggressive behaviour {832,833}, independent of other
variables {1276,2169}.
The immune infiltrate in tumours is referred to as TILs. TILs
are mononucleated lymphoid cells infiltrating the tumour and
its stroma; they reflect the host immune response against the
92 Epithelial tumours of the breast
tumour cells. The extent of TILs in lBC is gaining importance as
a prognostic marker, with high numbers of TILs associated w+
a better outcome and better response to neoadjuvant therap..
in triple-negative and HE82-positive breast carcinomas (see
Prognos/.s ar}c/ prec/t'of;.on, below) {482,1227,1228,1830}.
For quantifying TILs, it is recommended to follow the interra
consensus scoring recommendations {1831,922}. The steps
are outlined in Fig. 2.84. It is recommended that quantificaticr
is done on H&E-stained tissue sections at a magnification I;
20-40x with a lox ocular in core biopsies or surgical spec~
mens, on the most representative tumour block. TILs should be
scored in the stroma between the areas of carcinoma, and al
mononuclear cells (lymphocytes and plasma cells) should t€
included. Stromal TILs should be scored as a percentage I;
the stromal areas alone - the carcinoma cells should not tE
included in the total assessed surface area. Peritumoural fo -
cular aggregates and tertiary lymphoid structures with germira
centres are indicative of an active immune response, but the..I
should not be included in the stromal TIL assessment. TIL qua--
tification should be reported as a percentage - a mean scoE
based on the available tissue analysed. If the TILs appear to be
heterogeneously distributed, an average should be reporte.=
disregarding hotspots {1831}. The development of comput_1
tional pathology methods is likely to make automated countirvg
an option in the future.
Morphological characteristics after neoadjuvant treatment
The residual carcinoma or tumour bed in the breast and tre
response in the lymph nodes can serve as measures jF
response to neoadjuvant treatment and can provide vai.-
able prognostic information, so these should be evaluated a-a
reported {204,203}. Multidisciplinary collaboration and correii
tion of clinical, imaging, gross, and microscopic findings a?
essential for accurately sampling and histologically quantifyirlgi
response to treatment. The residual cancer burden index is a
clinically validated, standardized reporting system that incorp=-
rates response in the breast and response in the lymph nodes
into a score that can be combined with other emerging pro;-
nostic factors.
Changes in the residual tumour cells, if present, are extremei!ii
variable in degree. In therapy-resistant cancers, no morpholog-
cal alteration may be detected. More commonly, carcinomas
become less cellular and are often present as scattered sin,a"
nests across the tumour bed. The size and cellularity of foci if
the overall residual cancer should be recorded, because hg
extent of residual invasive carcinoma, together with lymph none
status, is a powerful predictor of long-term survival {286}. Ir a
few cases, the remaining cancer cells become bizarre, win
large and irregular nuclei. The cytoplasm of the residual tumcijIT
cells may become vacuolated (in ~40°/o of cases) {1400}. 1
some cases, the only residual cancer is in lymphatic spaces_,
and this finding has been associated with recurrence after ne.=-
adjuvant therapy {320}. The mitotic count is often lower in t-e
residual carcinoma. Nevertheless, histological grade remains a
prognostic factor after neoadjuvant therapy and therefore may
be relevant to report (this is not standard) {218}. After a complee
response, only a loose, oedematous, vascularized fibroelasto:i=
area of connective tissue with chronic inf lammatory cells ard
macrophages may mark the tumour bed. When only small fci
of atypical cells are present, immunohistochemical studies m€
lnce as
ed with
Evaluation of tumour-infiltrating lvmphocvtes (TILs
therapy
as (see Step 1: Define the area for evaluation Step 2: Focus only on stromal llLs Ste|) 3: Dctemlne thetvDe oflnflammatorv lnflltrate
Large areas of central necrosis or Include only mononuclear infiltrate (lymphocytes
internal fibrosis are not included in the and plasma cells).
3 Steps evaluation.
ification
ation of
}\ spec`,-
hou\d bE
1. and al
hould be
ntage cF
d not be
)ural fo"-
germin=,
but they,
TI L qur-
;an scxre
rear to be
Stems 4 & 5: Assess and reDort the Dercentage of the stromal area involved bv TILs
reportefl
Report the average of the stromal area; do not focus on hotspots.
Comput>
countin=
For more detail, visit: www.tilsinbreastcancer.ore

a:£I E`raluation of tumour-infiltrating lymphocytes ITILs) in breast cancer.

•atment
t and tie for distinguishing cancer cells from benign histiocytic throughout the specimen rather than only in the vicinity of the
asures Of at invasive carcinoma from carcinoma in situ. DCIS may tumour bed; care should be taken not to overdiagnose these
ide valL+ t in the absence of resi.dual invasive carcjnoma. This as jn si.tu di.sease. Padi.ati.on can cause the stroma to be dense
Jated artl does not exclude pathological complete response, and and hypocellular. The epithelial cells of ducts and lobules may
d corret} rients have a good prognosis {194,649}. Primary neo- show slightly irregular and hyperchromatic nuclei, and lobules
clings are endocrine therapy is sometimes administered, but it may become sclerotic. Badiation fibroblasts may be seen, and
uantifyino results in pathological complete response. Treated can- bizarre stromal cells may also be present.
ndex is a •trq/ have a central area of fibrous scarring {2074}.
t incorpe mural breast epithelial structures may show atypia, in the Differential diagnosis
ph nodes of enlarged and occasionally pleomorphic nuclei, after lBC should be differentiated from malignant breast in situ lesions
ling prog- therapy. These may be present at some dis- (carcinoma in situ) and benign infiltrative lesions such as scleros-
tom the site of the invasive tumour and can be present ing lesions and microglandular adenosis. Although most lBCs
extreme,i.
)rpholog -
ircinomas
ired smal
of foci o;
)ause the
nph node
286). In a
arre, with
al tumour
(1400). Ir
) Spaces
after neo-
ver in the
remains a
3fore may
complete
oelastotic
cells and
-Z85 Breast carcinoma. A Low-power view of breast carcinoma after neoadjuvant chemotherapy, featuring partial response with areas of fibrosis without residual tumour
small foci mesenting tumour bed. Residual invasive tumour is seen jn part of the tumour bed. 8 Higher-power view shows residual breast carcinoma after neoadjuvant treatment. There
Jdies may s : aecrease of tumour cellularity compared with cellularity in the core biopsy.

Epithelial tumours c; :r€ ='==s: 93

have an infiltrative growth pattern, more-nested patterns can
mimic in situ lesions, such as DCIS. Conversely, DCIS involving
a sclerosing lesion can mimic invasive carcinoma. Benign scle-
rosing lesions such as sclerosing adenosis and radial scar can
mimic the infiltrative pattern of an invasive lesion. Myoepithelial
stains are essential for distinguishing between these lesions,
with invasive lesions lacking myoepithelial cells at their periph-
ery. However, the correct interpretation of myoepithelial staining
is important, because stromal or vascular staining can mimic a
myoepithelial cell layer in some cases. A nuclear myoepithelial
stain (e.g. p63) in addition to a cytoplasmic stain (e.g. calponin
or SMMHC) can be useful to help avoid pitfalls in interpretation.
Bare breast lesions that can mimic primary lBC include
microglandular adenosis (which can mimic a well-differentiated
invasive cancer) and primary breast melanoma and lymphoma
(which can mimic a poorly differentiated breast cancer). In
addition, primary breast sarcomas resemble lBCs, often with
metaplastic features but sometimes also lBC-NST. Diagnostic
difficulties can also be encountered with malignant phyllodes
tumour (especially when the lesion lacks characteristic epithe-
lial areas of phyllodes tumour), primary breast sarcoma NOS,
and primary or secondary breast angiosarcoma (with epithe-
lioid subtypes being a problematic mimic of carcinoma).
Although rare, metastatic carcinoma to the breast should be
considered when unusual growth patterns are present or there
is a previous history or a clinical concern {2104}. Among the
most common metastatic cancers to the breast are lung ade-
nocarcinomas and ovarian, uterine, and renal cell carcinomas
{2104,1128,707}. Clinical history and immunohistochemistry are
essential for making the correct diagnosis. IBC should also be
.,
`'.*-:.;S i: fry
'®
.,.
t# +
C
fig.2.86 Morphological aspects on FNA (Papanicolaou staining). A Invasive carcinoma of no special type (NST). B Invasive lobular carcinoma, classic type. 0 lnvasiv5
mucinous carcinoma. D Tubular carcinoma.
94 Epithelial tumours of the breast
distinguished from primary or metastatic cutaneous carcino-
mas, including squamous cell carcinoma (differential diagnosis
with metaplastic squamous breast carcinoma), skin adnexa
carcinomas, and cutaneous melanoma. Parely, primary breas:
salivary gland-like carcinomas may need to be distinguishec
from metastatic salivary gland carcinomas, in particular if there
is a clinical history of salivary gland carcinoma of similar type.
Cytology
Because it is not possible to distinguish between in situ anc
invasive carcinomas cytologically, CNB has largely replacec
breast FNA, except in settings where access to image-guidec
biopsy is limited {1460}. However, cytological preparations
from FNA samples of suspicious axillary lymph nodes or othe-
sites of suspected metastasis are commonly used to rule ou:
metastatic breast cancer and have reasonable sensitMty anc
high specificity {110}. The analysis of EB/PB and HEB2 on FN4
material should be limited to cases of metastasis, to avoid test-
ing non-cancerous breast tissue {2089,1031,1906,1991}.
Cytological preparations of lBC are often characterized b}
highly cellular, loosely cohesive groups and single atypical cells
with syncytial arrangement and loss of polarity. The cells have
hyperchromatic nuclei; irregular, thickened nuclear membranes
nucleoli; and increased N:C ratios. Cells are often arranged i-
3D clusters or syncytial groups, or occasionally in acinar c..
gland-like patterns. The malignant cells are usually larger tha-
lymphocytes or benign ductal cells, but they can be less obv-
ously enlarged in well-differentiated cancers. The nuclei can b€
eccentrically located, with a plasmacytoid appearance. Bipc+
lar naked nuclei are not a feature. Occasionally, cytoplasmic
rty'C
...* , .I: ..#.p .
..S ,
*
JT±T € ..
E-i
Car=:m-

aJJ-ca,,I
ry ban
TJLJESTq
IT rf reel
arty¥-

sin an
ref-
equon
-a-'rm
; Or CritlE
i rule taLIT -izr ,r,Esive breast carcinoma of no special type (NST). A Core biopsy of a breast tumour reveals an invasive carcinoma NST. B At medium magnification, invasive
m[rIT== ``ST shows irregular malignant tumour cords amid chronic inflammatory cells on core biopsy.
ivity an=
2 on Flu
/oid test-
'iim-cilization, signet-ring cells, or intracellular lumina can be
). Diagnostic molecular pathology
Tized b, gEF Smears of poorly differentiated carcinoma may contain Molecular classification of breast cancer
ical celE eecnorphic, bizarre cells and multinucleated malignant cells. Breast cancer is heterogeneous at the molecular level, with
31ls have different patterns of gene expression leading to differences in
nbranes .-`:re needle biopsy behaviour and prognosis {277}. Over the past few years, there
lnged ir =`B for the initial evaluation of a breast lesion has been used has been considerable effort to characterize and classify breast
lcinar o. grmrsively for years as a non-surgical approach that allows for cancer at the molecular level in order to effectively tailor treat-
ger thar rmare-rapid diagnosis of palpable and non-palpable imaging- ment. However, due to time and cost constraints, in the great
)ss obv'- \anneted findings than excision (open) biopsy. With imaging majority of health care systems, surrogate molecular breast
)i can bE igutfance, CNB is highly sensitive and specific for the diag- cancer classification is still largely based on immunohistochem-
:e. BipcL im- and initial classification of breast cancer (233}. There is ical assessment of biomarkers (EP, PB, HE82, and Ki-67). Nev-
plasmic egcellent correlation between the findings on CNB and those ertheless, the examination of global gene expression patterns
IT open excisional biopsy {2287}, and diagnostic agreement (especially of genes involved in the regulation of cell growth and
r CNB is also high {389}. Sometimes a definitive diagnosis other important aspects of cell behaviour, such as .Invasion) has
I .ivasive cancer is not possible on the limited sampling of a resulted in the identification of intrinsic molecular subtypes of
=JB. If there is not 100°/o certainty in the diagnosis of invasion biological and clinical relevance and of gene signatures predic-
r CNB, an equivocal classification of a lesion as "suspicious", tive of outcome or response to therapy.
Tdeterminate", "cannot rule out invasion", or "uncertain malig-
~cit potential" may be most appropriate, with deferral to a surgi-
Intrinsic subtype classification
= specimen for definitive classification {1707,513}. When the Hierarchical cluster analysis of the genes that vary more
]agnosis of invasion is made on CNB, a preliminary histological between tumours than between repeated samples of the same
a.ade should be reported, and EB/PP and HEP2 testing can tumour (i.e. intrinsic genes) has revealed the existence of four
=e performed if there is sufficient invasive cancer for testing. major breast cancer intrinsic subtypes (luminal A, luminal 8,
3acause of the limited cold ischaemic time and excellent fixa- HE82-enriched, and basal-like), as well as a normal breast-like
rori of CNB samples, as well as the ability to make treatment group {277,1630}. Other rare subtypes, such as the claudin-
=cisions about possible neoadjuvant therapy before surgery, low class, which mostly comprises triple-negative tumours and
_=are needle biopsies are the preferred sample type for these shows a poor prognosis, have also been added. Subclassifi-
r€illary tests. Therefore, all breast core needle biopsies should
cation of the major subtypes has also been attempted, includ-
=e performed per the recommendations for adequate total fixa- ing HEB2-enriched and triple-negative subtypes. In order to
ron time in formalin (a minimum of 6 hours) before processing improve the standardization and reproducibility of the intrinsic
and should not be rush processed {2285,2284,831,1394,815, subtype classification, a quantitative BT-PCFLbased test with
='6). a curated list of 50 genes (the PAM50 gene signature) was
proposed. These genes were selected to classify lBCs into
luminal A, luminal 8, HEB2-enriched, and basal-like subtypes
{1608}. Because high-throughput transcriptomic analysis is
expensive and by no means widespread, a classification based
on the above-mentioned immunohistochemical biomarkers was
further developed, classifying tumours into the five subtypes
shown jn Box 2.01 (p. 96) {745}.

C Invasive

Epithelial tumours of the breast 95

Box2.01 The clinicopathological surrogate definitions of early invasive breast car- subtype has a better prognosis than the basal-like immunosup-
cinoma subtypes adopted by the 13th St. Gallen International Breast Cancer Confer-
pressed subtype). However, despite multiple efforts, there is no:
ence (2013) Expert Panel, based on immunohistochemical measurements of EPl, PPl,
yet an established or clinically verified diagnostic assay for the
ER882 (HER2), and Ki-67 with in situ hybridization confirmation where appropriate
classification of TNBC.
(745)

Luminal A-like Mutation profiles of IBC

• ER: positive
• PPl: positive
• HE82: negative
• Ki-67 proliferation index: low
Luminal B-like (HER2-negative)
• EFl: positive
• HEPl2: negative
• At least one of the following:
o Ki-67 proliferation index: high
o PPl: negative or low
Luminal B-like (HEPl2-positive)
• EPl: positive
• HEPl2: overexpre§sed or amplified
• Ki-67 proliteration index: any
• PPl: any
HEF12-positive (non-luminal)
• HEP2: overexpressed or amplified
• EPl: absent
• PPl: absent
Triple-negative
• ER: absent
• PR: absent
• HER2: negative
lntegrative cluster classificatlon
This classification divides breast cancer into 10 integrative clus-
ter subgroups (lntclusters) based on the integration of genomic
and transcriptomic data {423}. Each subgroup has a distinct
pattern of copy-number aberrations and is associated with dif-
ferent clinical outcomes and response to therapy. Six of the sub-known that inherited mutations in Bf?CA7 or Bf?CA2 (preset
groups (lntclusters 1, 2, 3, 6, 7, and 8) comprise predominantly in ~1-5°/o of breast tumours) result in homologous recombira-
EB-positive tumours and the PAM50 subtypes luminal A and
luminal 8, but with distinct genomic alterations. Intcluster 10 fied using whole-genome sequencing. The signature can also
comprises mostly EB-negative tumours, with high genomic
instability and a worse prognosis. Intcluster 4 comprises mostly with tumours showing BBCA1/2 dysfunction, such as sporaci:
tumours with extensive intratumoural lymphocytic infiltration.
Triple-negative breast cancer molecular subclassification
Triple-negative breast cancer (TNBC) is defined by the absence
of expression of EP, PP, and HE82 by immunohistochemistry,
which results in limited targeted therapeutic options. This is a
heterogeneous group of tumours with different molecular driv-
ers and prognoses {1702,1360}. Using gene expression data,
a classification of TNBC into the following four tumour-specific
subtypes (TNBctype-4) has been developed {1140}: basal-like 1
and basal-like 2 (which differ in immune response), mesenchy-
mal, and luminal AP. These subtypes present distinct survival
patterns and sensitivity to neoadjuvant chemotherapy {1140A}.
Combining BNA and DNA pro filing analyses also produces four
distinct subtypes: luminal AP, mesenchymal, basal-like immu-
nosuppressed, and basal-like immune-activated {256}. Each
subtype presents specific targets for treatment (e.g. AP and
the cell surface mucin EMA [MUC1] in the luminal AB subtype)
and a different prognosis (e.g. the basal-like immune-activated
Breast cancer subtypes present different mutation patterns
which influence response to treatment and prognosis. Lum-
nal A tumours have a high prevalence of P/K3CA mutations
(49%), whereas basal-like tumours have mostly rp53 mutations
(84%). In the setting of early disease, P/K3CA mutations havE
an overall prevalence of 32°/o, and they are associated wit-
older patient age; histologically well-differentiated and smaller
tumours; and EP-positive, HEB2-negative tumour status {2344
Becently, there was approval for using a P/K3CA mutation tes:-
ing method (by BT-Pop) to determine whether patients witr
advanced EB-positive breast cancer are eligible for treatmert
with alpelisib (a P13K inhibitor) {69A}.
Characterization of the genomic drivers of the various molec-
ular subtypes of TNBC has been attempted {122}. The basaL
like 1 subtype was the most genomically unstable, with a hig-
7P53 mutation rate (92°/a) and copy-number deletions in genes
•involved .in DNA repalir (BFICA2, MDM2, PTEN, FiB1. aind TP53 .
Luminal AP tumours were also found to be associated with =
higher mutation burden, with significantly enriched mutations r
PIK3CA (55°Mo), AKT1 (13°Mo), anc] CDH1 (13°Wo).
With the wider availability of next-generation sequencing
breast tumours are increasingly being tested for multipe
mutations and/or other genetic alterations, such as mutations
•in PIK3CA, ER882 (HER2), and ESF]1. Flecer\fty, the iise ct
whole-genome sequencing has led to the identification o=
mutation signatures. These signatures can reflect etiology I
biology, for example, damage due to ultraviolet (UV) radi
tion or defective DNA repair pathways. In breast cancer, it s
tion deficiency, and the characteristic signature can be ider:-
be identified in patients without germline Bf?CA7/2 mutations
basal-like/triple-negative cancers. This has led to the develoc-
ment of tools to detect homologous recombination deficiencg
(e.g. the HPDetect assay). As of recently, patients who ha.a
recurrent or stage 4 HE82-negative cancer and a knowii
BFCA7 or Bf?CA2 germline mutation are now considered ca--
didates for therapies that target this deficiency -namely, poiqu',"
(ADP-ribose) polymerase (PABP) inhibitors {1773A,1773B}.
Using next-generation sequencing, it is also possible to qua-~
tify a tumour's total mutation burden, which is associated vmi
therapeutic response to immunotherapy. Breast cancer dues
not have as high a total mutation burden as melanoma, but tlg
total mutation burden tends to be higher in TNBC (at 1.68 mu-_1
tions/Mb) than in luminal tumours (0.84-1.38 mutations/hJ@iM
{277,246}. Among the TNBCs, the luminal AB subtype see
to have a higher total mutation burden than the mesenchyr
stem-like subtype {122}. Therefore, immunotherapy may
useful in some subsets of breast cancers, especially subs
of TNBC.

96 Epithelial tumours of the breast

munosup- Essential and desirable diagnostic criteria by FNA or CNB. Pathological classification of T and N primarily
here is not isenl/.a/.. an invasive malignant process of breast epithelium relies on the gross and microscopic examination of surgically
Say for the tacking a peripheral myoepithelial cell layer (immunohisto- excised specimens. In most patients, T is based on the size of
chemistry may be required), with or without carcinoma in situ; the invasive carcinoma. If multiple areas of invasion are present,
EP, PB, and HEB2 immunohistochemistry to guide classifica- T classification is based on the largest focus. A small cancer
ton, treatment, and prognosis. is sometimes best evaluated by measuring its size on glass
patterns, slides. Correlation of gross, microscopic, and imaging findings
)sis. Lumi- Staging is often necessary to determine the best T category. Lymph
--e most widely used system for staging breast carcinoma is
mutations nodes should be evaluated in 2 mm slices, to identify all macro-
! mutations the TNM system published by the Union for International Can- metastases (metastases > 2 mm). Metastatic deposits < 2 mm
tions have ¥ Control (UICC) (229} and the American Joint Committee on but > 0.2 mm (or > 200 cells) in a lymph node are known as
)iated with Cancer (AJCC) {61}. This system captures information about the micrometastases and define the case as node-positive (pNlmi),
nd smaller enent of cancer at the primary site (Tumour, T), the regional whereas cases with only isolated tumour cells (< 0.2 mm)
tus {2344}. \FTnph nodes (Node, N), and spread to distant metastatic sites are considered node-negative (pN0i+) for staging purposes.
tation test- htetastasis, M). Special techniques for classification are not M classification is primarily determined by the results of radio-
tients with elc]uired, and comparable information can therefore be col- logical studies, with pathological confirmation after biopsy in
• treatment
eeted across time and in different locations. The T, N, and M some cases. Approximately 48°/o of breast cancers present as
rformation is combined to define five stages (0,I,11 Ill, and lv) stage I disease, 34°/o as stage 11,13°/o as stage Ill, and 5°/o a
)us molec- tat summarize information about the extent of regional disease stage lv {925}. However, the stage distribution is dependent
The basal- Hhimour size, skin or chest-wall invasion, and nodal involvement) on several variables, including race/ethnicity; patient age; EP
Ivith a high and metastasis to distant sites. In individual cases, this informa- status {925,2051,2232}; and the existence of population-based
is in genes rm is important for making decisions concerning the control breast cancer screening, which is associated with earlier stage
mcl TP53). rf k)cal disease, as well as to determine the value of systemic at presentation. There is also a correlation between the compo-
Ited with a raapy. Determining tumour stage is also essential for organ- nents of the stage; for example, only 19% of pTl tumours are
iutations in zng groups of similar patients for comparison in clinical trials, associated with nodal metastasis and only 1°/o are associated
aptdemiological studies, or other types of investigations. with distant metastasis, compared with as many as 40°/o and
quencing, Both clinical staging and pathological staging are used for loo/o, respectively, of pT3 tumours {925}.
r multiple patients with breast cancer. Clinical stage depends on physical Most breast carcinomas first metastasize to regional lymph
mutations eramination and imaging studies, with or without confirmation nodes via the lymphatics draining the tumour. The first draining
the use of
f ication of
etiology Or
UV) radia- Luminal A-like
ancer, it is
Luminal B-like (HER2-and HER+)
i2 (present
ecombina-
•===-=-=:=-=:::==:i::-::--------=- HER2+ (non-luminal)

i be identi- Triple negative

•e can also
) mutations
.s sporadic Luminal A
e develop- Luminal 8
deficiency
HER2-enriched
who have
I a known Basal-like

dered can- Claudin-Iowa

lmely, poly Normal-Iikea
17738).
le to quan-
ciated with
incer does
Tia, but the
1.68 muta-
tations/Mb)
ype seems
;senchymal
y may be
illy subsets

H|Z.88 Classification of breast cancer o{ no special type (NST). BL1, basal-like 1; BL2, basal-like 2; BLIA, basal-like immune-activated; BLIS, basal-like immunosuppressed
_+=, !uminal androgen receptor; TNBC, triple-negative breast cancer. aNot included in the PAM50 signature's classification.

Epithelial tumours of the breast 97

node is known as the sentinel lymph node. There is level 1 patients {2052}. The 10-year breast cancer-specific mortality
evidence that sampling this node for assessment results in rate associated with small, early-stage breast cancer (TNM
accurate staging, and this approach can be used to obviate stage Tla/bNOMO) has been reported to be as low as 49c
the need for axillary dissection and the associated risk of {819}. The 5-year relative survival rate of patients with localizeQ
lymphoedema in patients without metastases {421}. The sen- breast cancer (~60°/o of patients) is > 95°/o, which decreases
tinel lymph node can be identified preoperatively by injection to 85°/o when regional lymph nodes are involved and to aboi,:
of blue dye and a radioactive marker into the tissue around the 25°/o with metastatic disease. Completeness of excision is also
tumour. Axillary dissection is usually unnecessary in patients an important factor affecting local recurrence, but the only
with metastatic foci totalling < 2.0 mm. Immunohistochemistry margin status proven to have a significant effect on local recur-
is not usually required, but it may be used to clarify the nature rence after conservation surgery is a positive margin (0 mm..
of suspicious or atypical cells in some cases (e.g. Iobular tumour at ink) {881,1954}. Therefore, most guidelines endorse
carcinoma or postneoadjuvant cases). Intraoperative molecu- the adequacy of a "no invasive tumour at ink" margin for lBC
lar testing (e.g. one-step nucleic acid amplification) is highly and a 2 mm margin for DCIS (when DCIS is not associated
sensitive but not widespread (315,495,2363,2020,1614,1821},with invasion) {1423,1414}. Lymphovascular invasion (tumour
whereas frozen section or imprint cytology is less sensitive but emboli in blood or lymphatic peritumoural vessels) should be
more commonly performed {1198,638,424,1614,1641}. How- reported when present, because it is associated with risk a;
ever, assessment of nodes by high-resolution ultrasonography local recurrence (for which radiation may be offered even i-
can provide similar benefits without the need for intraoperative the mastectomy setting) and distant recurrence, especially for
assessment {846}. In the neoadjuvant setting, the value of sen- node-negative patients {1710}.
tinel node examination and the clinical significance of residual
isolated tumour cells are unknown, pending the results of Standard required prognostic/predictive markers
ongoing clinical trials. Hormone receptor and HEPl2 status: Hormone receptor (EP
Biological tests, such as receptor status and gene expression and PP) status and HEB2 status of lBC cells are recognized as
pro filing, may complement staging information by estimating indispensable predictive and prognostic factors for lBC ther-
the risk of future metastasis or recurrence or by predicting the apy decision-making by international guidelines such as those
likely response to treatment. In a major departure from tradi- published by the St. Gallen International Breast Cancer Expen
tional anatomical staging, the eighth edition of the AJCC can- Panel {422,421}, the European Society for Medical Oncologr
cer staging manual {61} TNM staging system has introduced (ESMO) {1899}, ASCO {815,2283), and the National Compre-
a breast cancer prognostic stage that incorporates anatomical hensive Cancer Network (NCCN) {759,1469}. Assessment crf
TNM information with tumour-intrinsic biology: namely, histologi- hormone receptor and HEP2 status is obligatory and usually
cal tumour grade {735,62} and predictive biomarkers (EB, PB, reflex for all lBC.
and HEP2 expression), in addition to multigene assays in a sub- Hormone receptors: The primary role of EP status is as a pre-
group of patients. The proposed prognostic stage incorporating dictive marker: patients with cancers expressing 2 1°/o nuclear
these seven variables (T, N, M, EB, PB, HE82, and grade) aims EB are significantly more likely to benefit from endocrine thera-
to improve stratification of patients with respect to outcome. pies than patients with cancers showing < 1°/o EB expression
The increasingly common practice of initiating therapy before (who do not benefit.) Most EB-positive cancers have high leves
definitive surgical treatment (i.e. neoadjuvant or presurgical of EB expression. EB is also prognostic: strongly EB-positive
therapy) requires a combination of information from clinical breast cancers have a better prognosis over the short term
examination, imaging, and pathological examination to deter- (5 years) than EP-negative breast cancers, but this is modified
mine the most likely T, N, and M classifications in addition to by grade and stage, and late recurrences can occur decades
tumour grade and receptor status before treatment. Posttreat- after initial diagnosis. Cancers with low levels of EP (most often
ment yT and yN classifications are determined after definitive defined as 1-10% EP staining) are uncommon but appear
surgery. Determining stage both before and after treatment pro- have a worse prognosis than those with high EB expression le+
vides important prognostic information {286,943}. els. Within the group of cancers that are EB-positive, PB expres-
sion levels (the percentage of stained cells) are considered a
Prognosis and prediction prognostic marker: cases with lower PB expression levels ares
The prognosis of lBC is profoundly influenced by the standard associated with worse outcomes, but patients still receive be--
variables discussed below. Some prognostic markers are also efit from endocrine therapy {1030,1519,1737}.
predictors of therapeutic response, such as EB and HE82 sta- HEPl2: HE82 is both a predictive and prognostic marker. A
tus {1703,925,1710}. The management of breast carcinoma is positive HEB2 test is predictive of survival benefit when HE82-
influenced by both prognostic and predictive characteristics. targeted therapies are added to chemotherapy regimens.
Therefore, a HEP2-positive test makes a patient a candid
Standard clinicopathological prognostic factors for both chemotherapy and HEP2-targeted therapies. HE82
Standard prognostic factors include patient age, disease also a marker of aggressive biology and poor prognosis in the
stage, tumour grade, tumour type, margin status, and lympho- absence of appropriate targeted therapies.
vascular status. Breast cancer in individuals aged < 35 years
at diagnosis is rare (< 5°/o) and potentially more aggressive, Additional prognostic/predictive markers
leading to more chemotherapy. However, one study found that Single markers of proliferation / Ki-67: Breast cancer prol`-
patients who were older at the time of diagnosis experienced eration is an important parameter to consider in the evali,a-
a 17°/o higher disease-specific mortality rate than younger tion of disease aggressiveness and therefore in therapeut.=

98 Epithelial tumours of the breast

c mortality cecision-making, especially concerning chemotherapy. Ki-67 is Of note, lHC4 is an immunohistochemistry-based assay of fou
icer (TNM " universally used or officially recommended, due to a lack of markers, including Ki-67, that was shown to predict residual ris
ow as 4% nernational consensus about scoring and cut-off values {542}, of distant recurrence in patients on adjuvant endocrine therapi
h localized aB well as a potential lack of reproducibility. Nevertheless, it in the ATAC trial as robustly as the 2l-gene BS {426}, althougl
decreases mjld be of clinical value (if available and accurately scored) it has not gained regulatory acceptance. Most often, Ki-67 i;
id to about as a supplement to grade in determining prognosis and poten- used as a preliminary, inexpensive but non-predictive marke
sion is also £al chemotherapy benefit (542}. Because EP-negative, HE82- of proliferation. Correlating a Ki-67 result with other findings
Jt the Only msitive cancers and EB-negative, HEB2-negative cancers are such as grade, EB status, or HE82 status, can also be usef
local recur- "ghly proliferative (Ki-67 proliferation index: 30-100°/o), these to ensure the findings all correlate well with each other and tt
gin (0 mm' cancers are more rapidly growing and are more often treated accurately characterize the biology of the cancer for prognostit
3s endorse " chemotherapy. EP-positive cancers have a wider range purposes.
gin for lBC Of proliferation, with highly proliferative forms correlated with APl status: Becent systematic review and meta-analysis dat{
associated Throre-aggressive behaviour, although most cases have a Ki-67 have shown that AP expression by immunohistochemistr`
on (tumour FTo`iferation index < 20°/o. A threshold of 14°/o or 15°/o has been {2164,217} and high AB inpNA levels by pooled gene analysi;
) should be Proposed for helping to discriminate between cases likely to {217} are associated with improved disease-free survival an(
with risk of correlate with the more aggressive luminal 8 molecular subtype better overall survival in patients with early-stage breast cancel
red even in i'(Hl+67 proliferation index 214°/o or 15°/o) versus the less aggres- However, AB's function in EB-positive and EB-negative breas
;pecially for gve luminal A subtype (Ki-67 proliferation index < 14°/a or 15%) cancers appears to be complex, and data on AB as a predic
E16A,197l}. However, this threshold has not been validated for tor of response to hormone/androgen-targeted therapies art
the purpose of predicting response to chemotherapy. Panel- currently somewhat limited and still under investigation {991
aased gene expression assays that are largely proliferation- 734}. Therefore, AB testing is not currently performed as stan
3ceptor (EB mren, such as the 2l-gene recurrence score (PS; see below), dard practice on all breast cancer cases, although it may bt
)ognized as iEp.'e been validated for this purpose in EB-positive cancers. requested in certain clinical settings.
)r lBC ther-
ch as those
ncer Expert "e2.08 Gene expression signatures used for chemotherapy decision-making in EB-positive, EPI882 (HE82)-negative breast cancer
al Oncology
Mammaprint Oncotype DX Prosigna (PAM50) Endopredict Breast cancer Index Genomic Grade Index
`al Compre-
sessment of umber of
70 2150 11 797
and usuaHy 9ETKrs

t.. DNA microarray PIT-PC B Nanostri ng PT-PCPI PIT-PC PI BT-PC B

; is as a pre-
TEso sample
: 10/o nuclear Frozen/FFPE FFPE FFPE FFPE FFPE FFPE
.+

]cr.ine thera-
} expression
LJrdon Centra|a Central Local Local Ce ntral Central

'e h.igh levels High or low risk High, intermediate, or High, intermediate, or High or low risk,
iEsl results High or low risk High or low risk
' EB-positive + subtype low risk low risk + subtype high or low benefit

3 short term Predicting prognosis Predicting prognosis Predicting prognosis Predicting prognosis Predicting prognosis
s is modif ied and guiding decision- and guiding decision- and guiding decision- and guiding decision- and guiding decision-
Chfal
cur decades rfution
making regarding making regarding making regarding making regarding making regarding

i (most often chemotherapy for chemotherapy for chemotherapy for chemotherapy for chemotherapy for No recommendation
GIEcording to
ut appear to E") women with
ER+"EPl2-EBC,
women with
ER+"ER2- EBC,
women with
EP+/HEPl2-EBC,
women with
EF]+AIER2- EBC,
women with
ER+AIEF]2- EBC,
`pression lev-
LN-or LN+ (1-3) LN-or LN+ (1-3) LN-or LN+ (1i) LN-or LN+ (1-3) LN-
3, PB expres-
3onsidered a tryctjve TAILOPlx (positive)

on levels are
dition MINDACT (positive) and BxPONDEPl OPTIMA (ongoing) None None ASTEPl70 (ongoing)

receive ben-
Hs) (Ongoing)

ftylatory EMA, FDA EMA, FDA EMA, FDA EMA, FDA Not approved Not approved
[ic marker. A rmval
when HEB2- Developed in a

)y reg.iinens heterogeneous patient

Postmenopausal Developed in Initially developed in
Developed in young Developed in patients population(62o/a
. a candidate patients in the training postmenopausal patients treated with
patients (aged < 55 who had received aged i 50 years),
Dies. HE82 is Crigival and development patients who had tamoxifen only and
tamoxifen only in the in order to permit
)gnosis in the dition set years) who had not
received systemic NSABP 8-20 and
Sets received received endocrine then further refined
the measurement of
heterogeneous therapy only in the in heterogeneously
therapy after surgery B-14 trials histological grade
treatment ABCSG-6 and -8 trials treated patients
via gene expression
profile

cancer prolif- 3= 3arly breast cancer; EGTM, European Group on Tumor Markers; EMA, European Medicines Agency; FDA, United States Food and Drug Administration; FFPE, formalin-fix
n the evalua- flrfuembedded; LN, lymph node; NSABP, National Surgical Adjuvant Breast and Bowel Project; PIT-PCPl, reverse transcriptase PCPr
n therapeutic CIEentralization, to allow (or local testing, is ongoing.

Epithellal tumours of the breast I

Assessment of response to neoadjuvant therapy: Achieve- 70-gene s/.gr}afure.. A 70-gene prognostic signature waE
ment of pathological complete response after neoadjuvant ther- developed as a microarray-based test that can be used n
apy is highly prognostic for HE82-positive and triple-negative determine the prognosis of patients with stage I or 11, nooe-
breast cancers {1623}. The residual cancer burden index can negative lBC of tumour size < 5.0 cm {2152}. It was later tes:ed
be used to establish risk classes associated with recurrence in patients with as many as 3 positive lymph nodes {285}. Tns
in cases with residual disease. The parameters to be quanti- dichotomous test classifies tumours as having either a good FT
fied and reported, as well as a calculator, are available online a poor prognosis, which is an independent predictor of distat
{21288,2017}. metastasis {2141}. The prognostic information provided by tie
Gene expression signatures: Gene expression signatures may 70-gene signature largely stems from the expression levels di
be used in clinical practice in some countries (when available EB-related and proliferation-related genes {486}; therefore tl
and/or reimbursed) for chemotherapy decision-making in EB- is useful for EB-positive, HE82-negative breast carcinomas.
positive, HE82-negative breast cancer. Low-risk patients as although its utility is limited or non-existent for the other breas±
defined by a gene expression signature can be spared chemo- carcinoma subtypes. The test can currently be performed in
therapy, whereas high-risk patients should benefit from chemo- formalin-fixed, paraffin-embedded tissue with analytical validT
therapy. Several signature assays are available (see Table 2.08, equivalent to fresh frozen tissue {1845}. The 70-gene signatures
p. 99). First-generation signatures (the 21-gene BS and 70-gene prognostic value is supported by level la evidence {285}, and E
signature) are centrally performed, and their prognostic value use is endorsed by ASCO, ESMO, the EGTM, and the St. Gal[en
is supported by level la evidence {285,1982,1981}. The use Panel {1065,1898,550,740}. Consensus opinion is that patiema
of these assays is endorsed by ASCO, ESMO, the European with a 70-gene signature result associated with low genorii=
Group on Tumor Markers (EGTM), and the St. Gallen Panel in risk can safely be spared the use of adjuvant chemotherapy in
certain clinical situations {824,1898,421}. Consensus opinion the setting of EP-positive, HEB2-negative, node-negative ea"
is that patients with a low genomic risk score result can safely breast cancer (which is considered to have a high clinical rs*
be spared the use of adjuvant chemotherapy in the setting of of relapse according to traditional criteria). The use of this teal
EB-positive, HEP2-negative, node-negative early breast can- in patients with 1-3 positive lymph nodes is more debatabe.
cer (which is considered to have a high clinical risk of relapse High genomic risk as defined by the signature has been shovun
according to traditional criteria). The use of these signatures in to be associated with higher sensitivity to neoadjuvant chemo-
patients with 1-3 positive lymph nodes is more debatable. therapy {1999}. Therefore, the 70-gene signature has also been
27-gene f?S.. The 21-gene BS is a quantitative BT-PCB- used as a selection factor in clinical trials in the neoadjuval
based signature based on the expression of 21 genes (16 can- setting {1812}. Second-generation signatures can be assessed
cer-related and 5 reference genes) that can be applied to PNA using dedicated instruments. They have level lb evidence `I
extracted from formalin-fixed, paraffin-embedded tissue sam- prognostic value in EB-positive, HEP2-negative patients treated
ples {1579}. The score is a mathematical function developed with hormone therapy. High-risk second-generation signatures
to predict the risk of distant relapse at 10 years for patients predict late recurrence, but the clinical value of this information
with EB-positive, HEB2-negative, lymph node-negative breast is low, because patients with a high-risk signature nowada.`S
cancer {1579}. It is a continuous variable (ranging from 0 to receive chemotherapy in the adjuvant setting. Therefore, tie
100) associated with the risk of distant relapse within 10 years, data generated from retrospective trials are not valid in this se=-
as well as an independent prognostic factor for EB-positive, ting, because the trial participants were only treated with hoe
node-negative patients with breast cancer treated with adju- mone therapy.
vant endocrine therapy (i.e. tamoxifen and aromatase inhibi-
tors) {1579,541}. On the basis of the 2l-gene PS, cases were Prognostic scoring systems
initially classified into three categories: low-risk (score: < 18), Prognostic scoring systems such as the Nottingham Prognostc
intermediate-risk (score: 18-31), and high-risk (score: > 31), Index (Npl) {1703} and Adjuvant! Online (www.adjuvantonline.
com) incorporate the clinical parameters of patient age, d;s-
which equate with 10-year relapse rates of 70/a,14°/o, and 3o°/a,
respectively. The 21-gene PS also correlates with benefit from ease stage, EP status, and tumour grade, but not HEP2 status.
chemotherapy in EB-positive, HEB2-negative breast cancers Another prognostic algorithm, Predict (https://breast.predict.nrts.
{1580}. Its use has been prospectively tested in a randomized uk/), does include HEB2 status. These algorithms can guice
trial, supported by level la evidence using new BS thresh- treatment decisions. The Magee equation-based BS is a no--
olds that stratify the benefit of adding chemotherapy on the commercial predictive tool based on tumour pathological chap
basis of a combination of BS and patient age {1982,1981}. acteristics (Scarff-Bloom-Bichardson grade, H-scores for EH'
The 21-gene PS is also endorsed by NCCN, ASCO, ESMO, and PB, HEP2 status, Ki-67 proliferation index, and tumour sizg|
the EGTM, and the St. Gallen Panel for EB-positive, HEP2- that can be used to estimate the 21-gene BS using the Magee
negative, node-negative breast cancer {824,1898,421}. A trial equation {2128A}. The Magee method may be used instead c#
regarding its use in patients with as many as 3 positive lymph the actual 2l-gene BS if the estimated Magee score is cleary
nodes is ongoing {753}. Both the 2l-gene BS and the 70-gene high or low {1026}.
signature have been incorporated in the eighth edition of the
AJCC cancer staging manual for the classification of breast Tumour-infiltrating lymphocytes
cancer {61}. A low genomic risk result downstages selected TIL assessment is gaining importance as a prognostic marker.
EP-positive, HE82-negative, node-negative tumours into the High numbers of TILs are associated with better outcome ardi
same prognostic category as Tla-bNOMO {735}. better response to neoadjuvant therapy in triple-negative and
HE82-positive breast carcinomas (level 1 b evidence) {482,122T.

100 Eplthelial tumours of the breast

 -228,1830}. TILs have a strong prognostic value in improv-
ure was Markers relevant to immune-checkpoint therapy
used to mg estimates of distant recurrence-free survival, disease-free (PDLl testing)
11, node- survival, and overall survival in early-stage TNBCs treated with Clinical trial evidence regarding immune-checkpoint blockade
er tested standard adjuvant/neoadjuvant chemotherapy (level lb evi- therapies in a variety of tumour types (including breast tumours)
85). This cence). This finding is based on an evaluation by pathologists is rapidly evolving. Monoclonal antibodies targeting the PDl/
good Or f ng H&E-stained glass slides at the time of diagnosis (before PDLl pathway or CTLA-4 are thought to function by removing the
)f distant treatment) and in the residual disease after neoadjuvant chemo- inhibition of the antitumour immune response {1598}. Data from
d by the nerapy. The presence and extent of TILs in lBC vary greatly the phase Ill lMpassionl30 clinical trial have shown that immu-
levels of from tumour to tumour. The quantification of TILs is feasible on nohistochemical PDLl expression on > 1°/o of immune cells in
refore, it Th&E-stained tissue sections during the diagnostic procedure metastatic TNBC is predictive of improvements in progression-
3inomas. Fc] follows international recommendations {1831}. The devel- free survival and overall survival when first-line atezolizumab is
3r breast arTient of computational pathology methods is likely to make added to protein-bound paclitaxel (nab-paclitaxel) {1859}. The
)rmed in a.[omated counting an option within the next few years. Clini- use of approved and validated antibodies and their correspond-
11 validity = utility (for treatment allocation) is under investigati.on. TILs ing organ-specific scoring systems is recommended if testing
gnature's sroiuld be considered as a stratification factor in clinical trials is performed. However, the field is rapidly evolving and other
;), and its Fd should be included in studies involving or evaluating prog- biomarkers may emerge that prove to be important for predic-
ncsis. It is recommended to follow the international consensus
;t. Galler tion of response to checkpoint inhibitors.
patients s=enng recommendations for quantifying TILs {1831,922}.
genomic
lerapy ir
tive earl)
nical risk
i this tes:
3batable
3n showr
t chemo-
llso beer
]adjuvan:
issessec
]ence fo.
[s treatec
ignatures
formatior
iowadays
3fore, the
1 this set-
with hor-

rognostic
intonline.
age, dis-
\2 status
3dict.nhs.
an guide
is a non-
ical char-
ts for EF
Tour size:
e Magee
nstead Of
is clearl)

c marker
:ome anc
ative ano
482,1227

Epithelial tumours of the breast 101