Molecular Diagnostics Consultation / Special Article

Molecular Pathology of Breast Cancer

What a Pathologist Needs to Know
Kimberly H. Allison, MD

Key Words: Molecular diagnostics; Breast; HER2 FISH testing; Oncotype DX

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

DOI: 10.1309/AJCPIV9IQ1MRQMOO
CME/SAM

Upon completion of this activity you will be able to: The ASCP is accredited by the Accreditation Council for Continuing
• identify the pathologic features that typically correlate with the 4 Medical Education to provide continuing medical education for physicians.
major molecularly defined subtypes of breast cancers. The ASCP designates this journal-based CME activity for a maximum of 1
• list the uses and limitations of clinically utilized molecular tests, AMA PRA Category 1 Credit ™ per article. Physicians should claim only the
including Oncotype DX, MammaPrint, and HER2 fluorescence in situ credit commensurate with the extent of their participation in the activity.
hybridization in breast cancer. This activity qualifies as an American Board of Pathology Maintenance of
• provide an algorithm for applying molecular testing in the histologic Certification Part II Self-Assessment Module.
workup of breast cancer in guiding appropriate breast cancer subtype The authors of this article and the planning committee members and staff
diagnosis and treatment. have no relevant financial relationships with commercial interests to disclose.
Questions appear on p 904. Exam is located at www.ascp.org/ajcpcme.

Abstract When it comes to breast cancer diagnostics, patholo-

Pathologists are now more than ever “diagnostic gists have become the translators of many layers of biologic
oncologists” and serve a critical role as clinical information including morphologic features, protein and gene
consultants on the biology of disease. In the last decade expression, and genetic information. Molecular techniques
and a half, molecular information has transformed our have both changed our understanding of the basic biology of
thinking about the biologic diversity of breast cancers breast cancer and provided the foundation for new methods
and redirected the way clinical treatment decisions are of “personalized” prognostic and predictive testing. While
made. A basic understanding of the current molecular traditional staging on the basis of tumor size and lymph node
classification of breast cancers and the biologic status remains the cornerstone of outcome indicators, it has
pathways from precursors to invasive disease is key become clear that not all breast cancers presenting at the same
to both informing diagnostic practice and serving as stage have the same underlying biology or clinical behavior.1,2
clinical consultants. In addition, both single-marker While morphologic assessment of factors such as Nottingham
and panel-based molecular tests are currently being grade and histologic subtype offer prognostic information
utilized in breast cancer tissue to predict the benefit about how aggressively a cancer may behave clinically, the
of specific therapies such as HER2-targeted biologic biology of each cancer can additionally be resolved by com-
therapy and chemotherapy. Familiarity with the bining morphologic factors with results of ancillary tissue-
current issues involving these molecular tests as well based testing.
as the pathologist’s role in ensuring appropriate tissue As pathologists and as “diagnostic oncologists,” we are
handling, tissue selection, and results interpretation the experts in tissue-based testing and will be looked to for
and correlation are paramount to providing optimal both clinical guidance as well as leadership in test quality and
patient care. innovation. Therefore, it is critical to have a current under-
standing of both the benefits and limitations of molecular
testing in the clinical setting. This review will cover (1) how
molecular testing has changed our current understanding of
the biology of breast cancer and how this knowledge can be
used to inform clinical diagnosis; (2) how molecular testing is
currently used to inform clinical practice and treatment deci-
sions in breast cancer; and (3) how pathologists can serve as
molecular testing consultants and help guide the clinical deci-
sions related to them.

770 Am J Clin Pathol 2012;138:770-780 © American Society for Clinical Pathology

770 DOI: 10.1309/AJCPIV9IQ1MRQMOO
 Molecular Diagnostics Consultation / Special Article

Molecular Testing as a Tool for expression levels of hormone receptors, higher Nottingham
Understanding the Biology of Breast Cancer grade, and higher proliferative rates; and can be HER2 posi-
The explosion of molecular information in the past tive.16 There is clinical interest in distinguishing the luminal
decade and a half has led to a better understanding of the B cancers from luminal A cancers because they may be a
biologic diversity of breast cancers as well as clues to the subset of ER-positive cancers that derive benefit from more
different etiologic pathways to breast cancer development. aggressive therapy.17 However, it is also acknowledged that
As pathologists, we are expected to serve our clinical teams the differences between these 2 groups are largely based on
as consultants on the biologic understanding of breast cancer differences in proliferation-related genes and, rather than rep-
subtypes and their pathogenesis. resenting distinct subtypes of breast cancer, are more likely 2
ends of a spectrum of ER-positive disease.3,18,19
The basal-like subtype appears to overlap substantially

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

with the population of breast cancers that are “triple negative”
Molecular/Intrinsic Subtypes
(ER/progesterone receptor (PR) and HER2 negative) and high
It is useful to establish groupings or subtypes of breast grade but also are associated with characteristic histologic
cancers that have both biologic and clinical relevance. With features such as solid-pushing borders, geographic areas of
more than 80% of breast cancers classified histologically as necrosis, and dense lymphocytic infiltrates.20-25 Cancers with
“invasive ductal carcinoma, not otherwise specified,” there is a basal-like molecular profile can be identified with high
clearly more biologic diversity among breast cancers than our specificity (but only about 75% sensitivity) by the following
histologic classification scheme alone implies. Using addi- immunohistochemical profile: ER and PR negative, HER2
tional factors such as Nottingham grade, hormone receptor negative, and cytokeratin (CK) 5/6 or epidermal growth fac-
status, and HER2 status, breast cancers can be categorized by tor receptor (EGFR) positive (although other variations of
both biology and therapeutic opportunities. this panel have also been used).14,22,24,26,27 However, clini-
Gene expression profiling has reconfirmed that the cally validated thresholds for CK5/6 or EGFR staining are
major drivers of breast cancer biology are hormone recep- still lacking. In addition, there is currently no difference in
tor–related genes, HER2-related genes, and have highlighted the standard treatment of a triple-negative and a basal-like
the importance of proliferation-related genes.3-8 Seminal stud- breast cancer; therefore definitive classification of a basal-
ies evaluating the gene expression profiles of breast cancers like cancer has not become standard clinical practice for drug
have segregated them into “intrinsic subtypes” or “molecular therapy–related decisions.28,29
subtypes” based on the relatedness of their gene expression Clinically, basal-like cancers are noted to occur more
patterns using a hierarchical clustering.9-11 These data support often in younger patients and African American women and
4 main molecular/intrinsic breast cancer subtypes that have are associated with a worse prognosis.30,31 Interestingly, the
prognostic relevance to survival. These subtypes have been vast majority of BRCA1-associated breast cancers appear to
termed luminal A, luminal B, HER2-related, and basal-like (a have a basal-like molecular profile, suggesting a common
fifth “normal breast–like” category has not been reproducibly pathway of carcinogenesis in these patients.32-35 However, a
defined). ❚Figure 1❚ summarizes the characteristics of these basal-like or triple-negative profile by itself does not necessar-
molecularly defined main subtypes. These 4 subtypes, origi- ily predict BRCA1 mutation status (ie, many of these cancers
nally defined by gene-expressing profiling, were also recently are in BRCA1-negative patients).36,37
demonstrated as valid groupings using different platforms, The clinical usefulness of molecular subtype testing of
including genomic DNA copy number arrays, DNA methyla- individual breast cancers is still unclear, but assays to clas-
tion, exome sequencing, microRNA sequencing, and reverse- sify clinical cases into subtypes have recently become com-
phase protein assays.12 At the most basic level, the luminal mercially available. The PAM-50 Breast Cancer Intrinsic
subtypes share expression of estrogen receptor (ER)–related Classifier assay (NanoString, ARUP Laboratories, Salt Lake
genes and have better overall survival than the HER2-related City, UT) is a reverse transcription polymerase chain reac-
and the basal-like subtypes, which are typically (but not uni- tion (RT-PCR) assay that uses the gene product levels of 50
formly) ER negative.10,13 genes thought to be inherent to the molecular subtype.38,39
Studies correlating the molecular subtypes of breast can- It can be performed on formalin-fixed, paraffin-embedded
cer with the more traditional, slide-based pathologic features tissue and categorizes individual cancers into the molecular
of the same cancers have identified correlations between the subtype to which they are most similar. The MammaPrint
2.14,15 Luminal cancers express hormone receptors and are test (Agendia, Irvine, CA), which currently requires fresh
lower grade while the HER2 subtypes overexpress HER2 tissue specimens because of its microarray-based platform,
gene products and are higher grade. Luminal B cancers have also includes an option to test for molecular subtype (Blue-
a worse prognosis than luminal A cancers; often have lower Print) in its panel of tests.40 However, these assays have

© American Society for Clinical Pathology Am J Clin Pathol 2012;138:770-780 771

771 DOI: 10.1309/AJCPIV9IQ1MRQMOO 771
Allison / Molecular Pathology of Breast Cancer

Molecular Luminal HER2 Basal

Subtype (A and B)

Genetic ↑ Luminal CKs and ↑ HER2-related ↑ Basal CKs

profile ER-related genes genes
(A>B)
B↑ in proliferation-
related genes

Histologic

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

correlates

A B High-grade, High-grade, sheet-

Lower- Higher- +/– apocrine like, necrosis,
grade grade features inflammation
ER+ ER+ *See exceptions

Surrogate
markers ER/PR–
HER2–

A B HER2+,
Strong Weaker +/– ER/PR CK5/6+/–
ER+, ER+, EGFR+/–
PR+/–, PR+/–,
HER2–, HER2+/–,
low Ki67 ↑ Ki67

Prognosis Good Intermediate Worse Worse

Response to Lower Intermediate Higher Higher

chemotherapy

Targeted Hormone therapies HER2-targeted Currently

therapies therapies investigational

❚Figure 1❚ Summary of the features of the basic molecular/intrinsic breast cancer subtypes. The characteristics of these
categories have been generalized for simplification. * Notable exceptions to the typically high-grade histologic features
associated with the basal molecular subtype are lower-grade cancers of the following special types: adenoid cystic carcinomas,
low-grade metaplastic carcinomas, apocrine carcinomas, and invasive carcinomas associated with microglandular adenosis.
CK, cytokeratin; EGFR, epidermal growth factor; ER, estrogen receptor; PR, progesterone receptor.

been criticized as single-sample predictors of molecular the additional subtypes can be variable.25,26 The impact of
subtypes because they were developed on the basis of hierar- intermixed normal tissue has also been suggested as a source
chical clustering rather than as a predictive test on individual of interference with gene expression profiling used as a pre-
samples.41 In fact, the results of molecular classification dictive single-sample test.25,36,37,47,48
testing on a case-by-case basis is highly dependent on the Although molecular/intrinsic subtypes have empha-
platform and data evaluation techniques used, with only sized the importance of the biology driving different breast
moderate agreement among them.42-46 While the classifica- cancers, it remains to be seen whether molecular assays
tion of the basal-like subtype appears to be the most reliable, for subtyping will prove to be reproducible, clinically

772 Am J Clin Pathol 2012;138:770-780 © American Society for Clinical Pathology

772 DOI: 10.1309/AJCPIV9IQ1MRQMOO
 Molecular Diagnostics Consultation / Special Article

useful, and practical. However, as discussed in the follow- This concept recognizes that different subpopulations that
ing sections, knowledge of these categories and how they have acquired additional genetic mutations compete with each
are defined can inform diagnosis in clinical practice. other under selective pressure as cancers grow, progress, and
metastasize.79 An extensive review of this topic is beyond the
scope of this article, but the concept of clonal evolution can
be incorporated into the current classification of molecular
Molecular Etiology of Breast Cancer
subtypes in breast cancer by recognizing that some cancer
Molecular evidence has shed light on the different path- subtypes are more rapidly evolving because of their inherently
ways leading to the development of invasive breast cancers. high genetic instability (ie, the basal-like subtype). It follows
Studies looking at the patterns of gene copy number changes that with disease progression, any cancer’s clinically relevant
and mutations present have identified certain genetic altera- phenotype may change. Changes in the hormone receptor

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

tions (such as deletions of 16q and gains of 1q) in ER-positive or HER2 status with metastatic progression are estimated to
cancers that appear to be rare in ER-negative cancers.49-54 change approximately 10% to 40% of the time.80,81 These
ER-negative cancers tend to harbor more severe genetic aber- findings may have various alternative explanations, includ-
rations such as p53 mutations, HER2 amplifications, BRCA1 ing the way in which non–breast tissue is handled. However,
dysfunction, and high genomic instability. This evidence because of the effect on therapeutic options for a patient with
suggests that ER-positive and ER-negative cancers have 2 dis- metastatic disease (especially in the setting of disease previ-
tinct pathways of development that do not commonly overlap. ously negative for these markers), the current National Com-
In addition, contrary to initial data suggesting that pro- prehensive Cancer Network (NCCN) recommendations are
gression from low- to high-grade invasive cancers was rare, it to retest metastatic disease for hormone receptors and HER2
appears that the ER-positive subset of cancers may progress status.82 The implications for reporting HER2 heterogeneity
from low to high grade.50,55 Approximately 50% of grade 3 by means of fluorescence in situ hybridization (FISH) testing
ER-positive cancers have the same 16q and 1q alterations as are also relevant to this topic and will be discussed.
grade 1 ER-positive cancers, suggesting a common pathway The aforementioned combination of concepts recog-
to development.50 Grade 3 ER-positive cancers tend to have nizes the common molecular pathways in breast cancer
accumulated additional genetic changes resulting in higher development and progression as well as the inherent plas-
proliferation rates and additional genomic instability. ❚Figure ticity of cancer that makes biology reasonably predictable
2❚ shows an overview of these proposed pathways in breast when looking at large groups but less predictable on an
cancer development. individual case basis.
Recent evidence also suggests that in situ and invasive
ER-positive ductal and lobular cancers are nearly identical
from a genetic standpoint, with the same 16q and 1q altera-
ER-driven growth
tions but the additional loss of E-cadherin expression in the 16q deletions ER+ ER+
1q gains low-grade low-grade
lobular phenotype. In fact, the molecular similarities among
in situ invasive
flat epithelial atypia, atypical ductal hyperplasia, and lobular carcinomas carcinomas
in situ neoplasias all point to a very similar low-grade path- Additional mutations,
way of neoplastic progression and explain their frequent asso- genomic instability
ciation with each other on histology.51,56-62 In contrast, basal- Normal ER+ ER+
high-grade high-grade
like breast cancers are frequently noted to lack a recognizable proliferative in situ invasive
breast
in situ component. Although initially thought to arise from the carcinomas carcinomas

basal/myoepithelial cell of the breast, recent evidence refutes

this theory but sheds little light on other possible precursors.63
ER– ER–
A high-grade, triple-negative form of ductal carcinoma in situ high-grade high-grade
(DCIS) has been recognized, and microglandular adenosis ER-independent in situ invasive
growth carcinomas carcinomas
has also been linked as a possible precursor, but both are p53 loss
much less prevalent than invasive triple-negative/basal-like HER2 amplification
BRCA1 loss
breast cancers.64-70 Extensive high-grade DCIS (often of the Genomic instability
so-called “comedo type”) is more frequently associated with
HER2-positive invasive disease. ❚Figure 2❚ A model of the molecular pathogenesis of breast
Lastly, recent evidence highlighting the molecular and cancers. The more common pathway is shown at the top,
genetic heterogeneity of cancers (including breast) has brought with increasingly less common pathways shown below.
to the forefront the concept of clonal evolution in cancer.71-78 ER, estrogen receptor.

© American Society for Clinical Pathology Am J Clin Pathol 2012;138:770-780 773

773 DOI: 10.1309/AJCPIV9IQ1MRQMOO 773
Allison / Molecular Pathology of Breast Cancer

Correlating Molecular Information With complexity has major implications for the development of
Morphology to Inform Diagnostic Practice tests to identify candidates for therapy and may result in blur-
Recognition of the histologic features that correlate ring the lines between cancer types as we identify common
with molecular subtypes and what we know about etiologic targets in cancers arising from different organs.
pathways can assist in clinical diagnostic categorization.
Characteristics classically associated with these subtypes
should serve as a check and, if they are discordant, should
Molecular Testing in Clinical Practice
prompt reevaluation of the details of the diagnosis, the valid-
and Treatment Decisions
ity of prognostic/predictive markers on the case, as well as
consideration of recognized exceptions to the classic features
of the intrinsic subtypes. Well-characterized exceptions to the Single-Marker Molecular Testing for Prediction

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

descriptions of subtypes listed in Figure 1 include low-grade of Response to Therapy: HER2
triple-negative cancers with unique histologic subtypes such Traditionally, ancillary tissue-based testing has used a
as adenoid cystic carcinoma, apocrine carcinoma, microglan- single marker at a time. It is currently standard practice to
dular adenosis–associated cancers, and lower-grade meta- test breast cancer tissue for ER and PR using IHC.89 These
plastic carcinoma variants (fibromatosis-like, squamous, hormone receptor IHC markers correlate with benefit from
and adenosquamous variants), all of which have basal-like endocrine therapy and are indicators of overall prognosis.
molecular profiles but, unlike most cancers with this profile, HER2 testing is also standard practice because it predicts
are classically low grade.83-86 Using this approach, recogni- response to HER2-targeted therapies as well as a likelihood of
tion of discordant features (such as low grade but ER nega- response to certain chemotherapeutic regimens and prognoses.
tive) can aid in reconsideration of histologic type, grade, and HER2 testing is performed most commonly with either IHC or
the results of ancillary tests. molecular methods such as FISH. Additional in situ molecular
Knowledge of the relatedness on the molecular underpin- methods have been approved and used, such as chromogenic
nings of precursor lesions can also aid in recognizing a con- in situ hybridization and silver staining in situ hybridization.
stellation of findings that make biologic sense in a given case. Because of the clinical impact of this test and reports of high
For example, the presence of low-grade precursor lesions such discordance rates among laboratories, the details of HER2
as flat epithelial atypia, atypical ductal hyperplasia, or atypical testing and reporting (and now also ER/PR testing) became the
lobular hyperplasia/lobular carcinoma in situ may prompt a focus of specific guidelines put forth by the College of Ameri-
more careful examination for low-grade invasive lesions with can Pathologists (CAP) and the American Society of Clinical
subtle histologic findings such as small invasive lobular or Oncology (ASCO) in 2007.90 These guidelines make recom-
tubular carcinomas. In contrast, a case with extensive high- mendations on appropriate tissue handling, test methodology,
grade, ER-negative DCIS may harbor small foci of invasive validation, scoring, and reporting, and adherence to certain
HER2-positive disease (which can often be highlighted by aspects are CAP required. This regulation is likely to continue
HER2 or pancytokeratin immunohistochemistry [IHC]). to expand to any laboratory test that is used to offer a patient a
specific therapeutic option (whether molecular or not).
Despite guidelines, there are still grey zones in HER2
testing. HER2 FISH results in the equivocal range (mean
The Future of Molecularly Defined Subtypes
HER2/CEP17 ratio of 1.8-2.2 or 4-6 absolute HER2 signals/
of Breast Cancer: A Continuing Evolution
cell), which occur in an estimated 5% of cases, still have
The major gene expression subtypes serve as useful unclear treatment implications, but clinicians will often seek
categories, but given the current therapeutic options, when pathologist consultation in this setting.91 CAP has issued a
genomic data are incorporated into the picture, these catego- clarification statement that all cases with a ratio of 2.0 or
ries can become even more complex; this suggests that the more were considered eligible for HER2-targeted therapy in
molecular classification of breast cancers is still evolving. several of the initial trials conducted and that this threshold
Curtis and colleagues recently analyzed both the genomic and should still be used to guide clinical management.92 But
transcriptomic profiles of more than 2,000 fresh frozen breast repeat testing by a second method (such as IHC) or on an
cancer samples and found not 4 but 10 biologically distinct additional tissue specimen is also recommended in the CAP/
subgroups with correlations to outcome.87 In addition, data ASCO guidelines for cases in the equivocal range due to
looking at the genomic alterations in breast cancers indicate proximity to the 2.0 threshold.
that although approximately half of all driver mutations are In addition, in recognition of possible genetic heteroge-
found in more than 10% of breast cancers, the other half are neity for HER2 gene amplification on FISH testing, a CAP
relatively unique to each cancer (present in <10%).88 This expert panel issued additional recommendations for reporting

774 Am J Clin Pathol 2012;138:770-780 © American Society for Clinical Pathology

774 DOI: 10.1309/AJCPIV9IQ1MRQMOO
 Molecular Diagnostics Consultation / Special Article

HER2 heterogeneity found on FISH.90 However, the frequen- Both have been well validated retrospectively using tissue
cy of heterogeneity for HER2 gene amplification by their pro- specimens from clinical trials, but their clinical usefulness
posed criteria (5%-50% of individual cells amplified) appears is still being more rigorously evaluated in prospective stud-
to be much higher than early limited evidence suggested (as ies.103-112 MammaPrint segregates all breast cancers into
many as 20%-30% meet the proposed criteria, most of which low- and high-risk profiles using a microarray-based gene
have only 5%-15% amplified cells), and these criteria also expression profile focused on 70 genes.113,114 Oncotype DX
result in many nonamplified cases being classified as hetero- was developed for use in node-negative, ER-positive cancers
geneous. Therefore this proposal has been questioned and not (although it is also used in the node-positive setting). RT-PCR
widely adopted.93-99 Interestingly, the evidence suggests that levels of 16 cancer-related gene products are used to calculate
FISH equivocal cases often display significant percentages a recurrence score (RS) that is reported as a continuous vari-
of amplified cells (30%-50%), and therefore this may be an able with stratification into low-, intermediate-, and high-risk

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

appropriate setting in which to report the presence of hetero- categories. Because of its microarray platform, MammaPrint
geneity and consider treatment.93,99 However, evidence is still requires fresh tissue, which has somewhat restricted its clini-
lacking regarding which thresholds are clinically meaningful. cal usefulness. Oncotype DX can be used on formalin-fixed
paraffin-embedded tissue, allowing the test to be run on
Molecular Signatures for Prediction and Prognostication archived tissue blocks at any point in clinical decision mak-
In contrast to the single-marker approach, panel-based ing.115 In addition, Oncotype DX has validated its RS to
molecular testing has been used more recently to develop both overall outcome as well as prediction of chemotherapy
tests that predict both prognosis and response to therapy.100 benefit.109 Therefore, the current NCCN guidelines advise
Although hormone receptor and HER2 tests have been clini- oncologists to withhold chemotherapy in ER-positive, lymph
cally validated to predict which patients will respond to hor- node–negative patients with a low Oncotype DX RS and offer
mone and HER2-targeted therapies, there has been great inter- it if the result is in the high-risk category. But an intermediate
est in tests that will predict response/need for chemotherapy as risk result is still considered a treatment gray zone. Results
well. Various calculators or algorithms, such as ADJUVANT! from the Trial Assigning Individualized Options for Treat-
Online (www.adjuvantonline.com), that determine chemo- ment (TAILORx) hope to further stratify this group.110
therapy benefit based on clinical and pathologic features are The cost of these assays is relatively high (approximately
available.101 There is clear benefit to treating groups with mul- $3,500 for Oncotype DX and $4,000 for MammaPrint) com-
tiple high-risk features (ie, younger age, lymph node positive, pared with traditional pathology testing, and the benefit over
high grade, ER negative) with chemotherapy. However, there traditional clinicopathologic predictive factors may be limited
is less clear benefit to treating patients with traditionally low- to cancers with intermediate features. The RS calculation used
risk features (ie, lymph node negative, ER positive). Therefore, in Oncoytpe DX testing is heavily weighted for proliferation-
gene expression profiles have been developed as a way to tease related markers but also includes markers used routinely in
out differences in prognosis and chemotherapy benefit in this diagnoses such as ER, PR, and HER2. Although evidence
group. Although there are many signatures in various phases of shows that tests such as Oncoytpe DX more accurately pre-
development, Oncotype DX (Genomic Health, Redwood City, dict outcome and spare more women from chemotherapy
CA) and MammaPrint are currently the most common clinical- than algorithms such as ADJUVANT! Online, other evolving
ly used tests. ❚Table 1❚ summarizes the features of these tests. evidence suggests that different algorithms may be used to
Both Oncotype DX and MammaPrint offer prognostic predict RS and outcomes that take into account more of the
signatures in patients with breast cancer but with slightly dif- pathology-based factors such as levels of ER, PR, and HER2,
ferent target populations and different testing platforms.102,103 grade, and proliferation index markers such as Ki67.106,116-122

❚Table 1❚
Techniques to Determine the Benefit of Chemotherapy in Patients With Breast Cancer

Technique Type of Test Validated in Results Reported as Cost Factors Used

ADJUVANT! Online calculator ER+ or ER– Percent benefit in overall survival Free Age, comorbidities, size, grade,
  Online     lymph node status, ER status
Oncotype DX FFPE tissue, RT-PCR ER+ only Recurrence score: stratified into low-, $3,500 mRNA levels of 16 cancer-
  intermediate-, and high-risk groups   related genes
MammaPrint Fresh tissue, microarray LN –, ER+ or ER– Low or high risk $4,000 70-gene expression signature

ER, estrogen receptor; FFPE, formalin-fixed paraffin-embedded; LN, lymph node; RT-PCR, reverse transcription polymerase chain reaction.

© American Society for Clinical Pathology Am J Clin Pathol 2012;138:770-780 775

775 DOI: 10.1309/AJCPIV9IQ1MRQMOO 775
Allison / Molecular Pathology of Breast Cancer

Combining molecular prognostic and predictive sig- Ensure control Selection of

natures with traditional studies can add further data to the of preanalytic tissue for Results
variables molecular reviewed
treatment-related decision-making process. However, it is • Ischemic time testing
important to remember that molecular testing is not necessar- • Fixation time
ily a new “gold standard.” One benefit of more traditional in Reevaluate
Correlation
situ tests such as IHC is the visual confirmation that scoring is • Initial histology/ of results
clinical information Discordant with histologic
only performed on the invasive cancer. With molecular-based • Results/technique results and clinical
multiplexed tests, intermixed inflammatory cells, carcinoma of molecular test features
• Effect of preanalytic
in situ, and normal tissue may influence results.47,48,118 In variables
addition, many of these proprietary tests are only performed • Consider retesting
on additional
in a single clinical laboratory without the benefit of external material Concordant

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

• Discuss with results reported
confirmation of results on a routine basis. Pathologists are clinician
responsible for selecting material to be sent to these commer-
cial laboratories for molecular testing and should be cognizant ❚Figure 3❚ A model for the pathologist’s role in molecular
of unusual results as well as which material is most appropri- testing of breast cancers.
ate to send (recommendations for which are listed herein).

6 and a maximum of 48 hours in neutral-buffered formalin

How Pathologists Can Help Ensure Molecular
for HER2 FISH testing.89,90 Although these time points have
Testing Accuracy
not been extensively validated, they are a starting point for
Molecular and nonmolecular tissue-based testing can be standardization, and adherence to these recommendations
effected via many variables. Pathologists, as experts in both should be targeted with documentation when reporting cases
tissue morphology and ancillary testing, must strive to ensure that are outliers because of potential effects on molecular and
the accuracy of molecular-based tests performed on tissue nonmolecular testing.
samples used either in their own laboratory or sent to another
laboratory. This may occur by ensuring appropriate tissue Selection of Blocks for Testing
handling before an assay is performed or blocks are selected For laboratories performing HER2 FISH testing, it is
for send-out tests, or may involve review and correlation of required to circle the area with the invasive carcinoma so
results with clinicians. These recommendations are summa- that technologists know in which areas to count signals. To
rized in ❚Figure 3❚. appropriately screen for amplified populations or subpopula-
tions, as much of the invasive carcinoma should be circled as
Control of Preanalytic Variables: Ischemic Time and possible, with attention being paid to the exclusion of areas of
Fixation carcinoma in situ. This can be done on the H&E-stained slide,
Most molecular tests are susceptible to preanalytic vari- or if performed with concurrent HER2 IHC the IHC-stained
ables such as ischemic time because of potential degradation slide can be circled. The latter can be especially useful when
of liable targets.123 The CAP/ASCO HER2 and ER/PR test- there are distinctly different levels of HER2 expression on
ing guidelines currently recommend that ischemic time be IHC in different areas of the slide. When a distinctly clus-
as minimal as possible and preferably less than 1 hour.89,90 tered subpopulation appears to have higher HER2 protein
Ensuring that this happens requires coordination between the expression and these findings match clustered areas of HER2
clinical team acquiring the specimen (typically radiology or gene amplification on FISH, these areas can be counted and
surgery) and pathology and is a conversation that pathology reported as a separate HER2-amplified subpopulation.124,125
often needs to initiate. The often immediate fixation of breast Pathologists will have to guide technologists regarding where
needle core biopsies can make them a preferable sample for to count each population in these cases and review findings to
some tests compared with surgically acquired specimens that ensure appropriate scoring.
may have more prolonged ischemic time. However, limited Requests for block selection for additional testing, such
material may be present in core biopsy samples, so stan- as Oncotype DX, are frequently made to pathology. The block
dard procedures are recommended to ensure that surgically with invasive cancer of the worst grade and phenotype should
acquired samples also receive appropriate handling. be offered unless otherwise guided by the requesting clinician.
For molecular testing that can be performed on formalin- Ideal block selection would avoid substantial amounts of car-
fixed tissue, the type of fixative and time in fixative are sig- cinoma in situ, biopsy site changes, and inflammation because
nificant variables. CAP/ASCO recommends a minimum of these may influence results.

776 Am J Clin Pathol 2012;138:770-780 © American Society for Clinical Pathology

776 DOI: 10.1309/AJCPIV9IQ1MRQMOO
 Molecular Diagnostics Consultation / Special Article

Review and Correlation of Results With Histologic 6. Sotiriou C, Wirapati P, Loi S, et al. Gene expression profiling
Findings in breast cancer: understanding the molecular basis of
histologic grade to improve prognosis. J Natl Cancer Inst.
Finally, the results of molecular tests should be correlated 2006;98:262-272.
with the original histologic findings. For example, an unex- 7. Weigelt B, Baehner FL, Reis-Filho JS. The contribution
pected result might be a high Oncotype DX score or a HER2- of gene expression profiling to breast cancer classification,
prognostication and prediction: a retrospective of the last
positive FISH result in a low-grade invasive cancer. Re-review decade. J Pathol. 2010;220:263-280.
of the material sent for additional testing to look for potential 8. Weigelt B, Reis-Filho JS. Histological and molecular types
confounders such as inflammation, carcinoma in situ, etc, as of breast cancer: is there a unifying taxonomy? Nat Rev Clin
well as a reevaluation of the initial histologic findings, can be Oncol. 2009;6:718-730.
useful in identifying causes for the unexpected result. 9. Perou CM. Molecular portraits of human breast tumours.
Nature. 2000;406:747-752.

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

10. Sorlie T, Perou CM, Tibshirani R, et al. Gene expression
patterns of breast carcinomas distinguish tumor subclasses
with clinical implications. Proc Natl Acad Sci U S A.
Conclusions 2001;98:10869-10874.
11. Sorlie T, Tibshirani R, Parker J, et al. Repeated observation
Molecular testing in breast cancer is still in evolution. of breast tumor subtypes in independent gene expression data
With the increasing use of more sophisticated techniques sets. Proc Natl Acad Sci U S A. 2003;100:8418-8423.
such as deep sequencing, we may continue to discover more 12. The Cancer Genome Atlas Network. Comprehensive
high-throughput ways of harnessing large amounts of data to molecular portraits of human breast tumours. Nature.
2012;490:61-70.
predict outcomes and treatment responses in breast cancer.
13. van’t Veer LJ. Gene expression profiling predicts clinical
However, moving a test from the bench to the bedside is a outcome of breast cancer. Nature. 2002;415:530-536.
long process that requires robust validation. For a new assay 14. Nielsen TO, Hsu FD, Jensen K, et al. Immunohistochemical
to become standard practice, in addition to the test validation and clinical characterization of the basal-like subtype of
required, it must be clinically useful, practical, and cost-effec- invasive breast carcinoma. Clin Cancer Res. 2004;10:5367-
5374.
tive. Because of this challenge, some tests, despite using very
15. Tang P, Skinner KA, Hicks DG. Molecular classification of
sophisticated molecular techniques, may remain most useful breast carcinomas by immunohistochemical analysis: are we
in a research setting and not prove to be clinically essential. ready? Diagn Mol Pathol. 2009;18:125-132.
However, there is little doubt that molecular testing is chang- 16. Cheang MC, Chia SK, Voduc D, et al. Ki67 index, HER2
status, and prognosis of patients with luminal B breast cancer.
ing, and will continue to change, the way we diagnose and J Natl Cancer Inst. 2009;101:736-750.
treat breast cancer. 17. Desmedt C, Haibe-Kains B, Wirapati P, et al. Biological
processes associated with breast cancer clinical outcome
From the Department of Pathology, University of Washington depend on the molecular subtypes. Clin Cancer Res.
Medical Center, Seattle, WA. 2008;14:5158-5165.
18. Reis-Filho JS, Weigelt B, Fumagalli D, et al. Molecular
Address reprint requests to Dr Allison: Dept of Pathology, profiling: moving away from tumor philately. Sci Transl Med.
University of Washington Medical Center, 1959 NE Pacific St, 2010;2:47ps3.
Box 356100, Seattle, WA 98195; kallison@uw.edu. 19. Wirapati P, Sotiriou C, Kunkel S, et al. Meta-analysis of
gene expression profiles in breast cancer: toward a unified
understanding of breast cancer subtyping and prognosis
signatures. Breast Cancer Res. 2008;10:R65.
References
20. Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast
1. Foulkes WD, Reis-Filho JS, Narod SA. Tumor size and cancer. N Engl J Med. 2010; 363:1938-1948.
survival in breast cancer: a reappraisal. Nat Rev Clin Oncol. 21. Dent R, Trudeau M, Pritchard KI, et al. Triple-negative
2010;7:348-353. breast cancer: clinical features and patterns of recurrence.
2. Foulkes WD, Grainge MJ, Rakha EA, et al. Tumor size is an Clin Cancer Res. 2007;13(15 Pt 1):4429-4434.
unreliable predictor of prognosis in basal-like breast cancers 22. Fulford LG, Easton DF, Reis-Filho JS, et al. Specific
and does not correlate closely with lymph node status. Breast morphological features predictive for the basal phenotype in
Cancer Res Treat. 2009;117:199-204. grade 3 invasive ductal carcinoma of breast. Histopathology.
3. Reis-Filho JS, Pusztai L. Gene expression profiling in breast 2006;49:22-34.
cancer: classification, prognostication, and prediction. Lancet. 23. Jacquemier J, Padovani L, Rabayrol L, et al. Typical
2011;378:1812-1823. medullary breast carcinomas have a basal/myoepithelial
4. Hu Z, Fan C, Oh DS, et al. The molecular portraits of breast phenotype. J Pathol. 2005;207:260-268.
tumors are conserved across microarray platforms. BMC 24. Livasy CA, Karaca G, Nanda R, et al. Phenotypic evaluation
Genomics. 2006;7:96. of the basal-like subtype of invasive breast carcinoma. Mod
5. Sotiriou C, Neo SY, McShane LM, et al. Breast cancer Pathol. 2006;19:264-271.
classification and prognosis based on gene expression profiles 25. Rakha EA, El-Sayed ME, Reis-Filho J, et al. Patho-biological
from a population-based study. Proc Natl Acad Sci U S A. aspects of basal-like breast cancer. Breast Cancer Res Treat.
2003;100:10393-10398. 2009;113:411-422.

© American Society for Clinical Pathology Am J Clin Pathol 2012;138:770-780 777

777 DOI: 10.1309/AJCPIV9IQ1MRQMOO 777
Allison / Molecular Pathology of Breast Cancer

26. Cheang MC, Voduc D, Bajdik C, et al. Basal-like breast 44. Lusa L, McShane LM, Reid JF, et al. Challenges in projecting
cancer defined by five biomarkers has superior prognostic clustering results across gene expression-profiling datasets.
value than triple-negative phenotype. Clin Cancer Res. J Natl Cancer Inst. 2007;99:1715-1723.
2008;14:1368-1376. 45. Mackay A, Weigelt B, Grigoriadis A, et al. Microarray-based
27. Rakha EA, Elsheikh SE, Aleskandarany MA, et al. Triple- class discovery for molecular classification of breast cancer:
negative breast cancer: distinguishing between basal and analysis of interobserver agreement. J Natl Cancer Inst.
nonbasal subtypes. Clin Cancer Res. 2009;15:2302-2310. 2011;103:662-673.
28. Badve S, Dabbs DJ, Schnitt SJ, et al. Basal-like and triple- 46. Fan C, Oh DS, Wessels L, et al. Concordance among gene-
negative breast cancers: a critical review with an emphasis on expression-based predictors for breast cancer. N Engl J Med.
the implications for pathologists and oncologists. Mod Pathol. 2006;355:560-569.
2011;24:157-167.
47. Elloumi F, Hu Z, Li Y, et al. Systematic bias in genomic
29. Carey LA, Dees EC, Sawyer L, et al. The triple negative classification due to contaminating non-neoplastic tissue in
paradox: primary tumor chemosensitivity of breast cancer breast tumor samples. BMC Medical Genomics. 2011;4:54.

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

subtypes. Clin Cancer Res. 2007;13:2329-2334.
48. Cleator SJ, Powles TJ, Dexter T, et al. The effect of the
30. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer stromal component of breast tumours on prediction of
subtypes, and survival in the Carolina Breast Cancer Study. clinical outcome using gene expression microarray analysis.
JAMA. 2006;295(21):2492-2502. Breast Cancer Res. 2006;8:R32.
31. Morris GJ, Naidu S, Topham AK, et al. Differences in 49. Natrajan R, Lambros MB, Rodriguez-Pinilla SM, et al. Tiling
breast carcinoma characteristics in newly diagnosed African- path genomic profiling of grade 3 invasive ductal breast
American and Caucasian patients: a single-institution cancers. Clin Cancer Res. 2009;15:2711-2722.
compilation compared with the National Cancer Institute’s
Surveillance, Epidemiology, and End Results database. 50. Natrajan R, Weigelt B, Mackay A, et al. An integrative
Cancer. 2007;110:876-884. genomic and transcriptomic analysis reveals molecular
pathways and networks regulated by copy number aberrations
32. Foulkes WD, Stefansson IM, Chappuis PO, et al. Germline in basal-like, HER2 and luminal cancers. Breast Cancer Res
BRCA1 mutations and a basal epithelial phenotype in breast Treat. 2010;121:575-589.
cancer. J Natl Cancer Inst. 2003;95:1482-1485.
51. Bombonati A, Sgroi DC. The molecular pathology of breast
33. Joosse SA, Brandwijk KI, Mulder L, et al. Genomic signature
cancer progression. J Pathol. 2011;223:307-317.
of BRCA1 deficiency in sporadic basal-like breast tumors.
Genes Chromosomes Cancer. 2011;50:71-81. 52. Andre F, Job B, Dessen P, et al. Molecular characterization
of breast cancer with high-resolution oligonucleotide
34. Turner NC, Reis-Filho JS. Basal-like breast cancer and the
comparative genomic hybridization array. Clin Cancer Res.
BRCA1 phenotype. Oncogene. 2006;25:5846-5853.
2009;15:441-451.
35. Turner NC, Reis-Filho JS, Russell AM, et al. BRCA1
dysfunction in sporadic basal-like breast cancer. Oncogene. 53. Chin SF, Teschendorff AE, Marioni JC, et al. High-
2007;26:2126-2132. resolution aCGH and expression profiling identifies a novel
genomic subtype of ER negative breast cancer. Genome Biol.
36. Collins LC, Martyniak A, Kandel MJ, et al. Basal cytokeratin 2007;8:R215.
and epidermal growth factor receptor expression are not
predictive of BRCA1 mutation status in women with triple- 54. Jonsson G, Staaf J, Vallon-Christersson J, et al. Genomic
negative breast cancers. Am J Surg Pathol. 2009;33:1093- subtypes of breast cancer identified by array-comparative
1097. genomic hybridization display distinct molecular and clinical
characteristics. Breast Cancer Res. 2010;12:R42.
37. Lakhani SR, Reis-Filho JS, Fulford L, et al. Prediction
of BRCA1 status in patients with breast cancer using 55. Lopez-Garcia MA, Geyer FC, Lacroix-Triki M, et al.
estrogen receptor and basal phenotype. Clin Cancer Res. Breast cancer precursors revisited: molecular features and
2005;11:5175-5180. progression pathways. Histopathology. 2010;57:171-192.
38. Chia SK, Bramwell VH, Tu D, et al. A 50-gene intrinsic 56. Oyama T, Maluf H, Koerner F. Atypical cystic lobules: an
subtype classifier for prognosis and prediction of benefit from early stage in the formation of low-grade ductal carcinoma in
adjuvant tamoxifen. Clin Cancer Res. 2012;18:4465-4472. situ. Virchows Arch. 1999;435:413-421.
39. Nielsen TO, Parker JS, Leung S, et al. A comparison of 57. Nishizaki T, Chew K, Chu L, et al. Genetic alterations in
PAM50 intrinsic subtyping with immunohistochemistry lobular breast cancer by comparative genomic hybridization.
and clinical prognostic factors in tamoxifen-treated Int J Cancer. 1997;74:513.
estrogen receptor-positive breast cancer. Clin Cancer Res. 58. Weigelt B, Kreike B, Geyer FC, et al. The molecular
2010;16:5222-5232. underpinning of lobular histological growth pattern: a
40. Nguyen B, Cusumano PG, Deck K, et al. Comparison of genome-wide transcriptomic analysis of invasive lobular
molecular subtyping with BluePrint, MammaPrint, and carcinomas and grade- and molecular subtype-matched
TargetPrint to local clinical subtyping in breast cancer invasive ductal carcinomas of no special type. J Pathol.
patients. Ann Surg Oncol. 2012;19:3257-3263. 2010;220:45-57.
41. Pusztai L, Mazouni C, Anderson K, et al. Molecular 59. Simpson PT, Reis-Filho JS, Gale T, et al. Molecular
classification of breast cancer: limitations and potential. evolution of breast cancer. J Pathol. 2005;205:248-254.
Oncologist. 2006;11:868-877. 60. Reis-Filho JS, Lakhani SR. The diagnosis and management of
42. Weigelt B, Mackay A, A’Hern R, et al. Breast cancer pre-invasive breast disease: genetic alterations in pre-invasive
molecular profiling with single sample predictors: a lesions. Breast Cancer Res. 2003;5:313-319.
retrospective analysis. Lancet Oncol. 2010;11:339-349. 61. Dabbs DJ, Carter G, Fudge M, et al. Molecular alterations in
43. Kapp AV, Tibshirani R. Are clusters found in one dataset columnar cell lesions of the breast. Mod Pathol. 2006;19:344-
present in another dataset? Biostatistics. 2007;8:9-31. 349.

778 Am J Clin Pathol 2012;138:770-780 © American Society for Clinical Pathology

778 DOI: 10.1309/AJCPIV9IQ1MRQMOO
 Molecular Diagnostics Consultation / Special Article

62. Lee S, Medina D, Tsimelzon A, et al. Alterations of 83. Acs G, Simpson JF, Bleiweiss IJ, et al. Microglandular
gene expression in the development of early hyperplastic adenosis with transition into adenoid cystic carcinoma
precursors of breast cancer. Am J Pathol. 2007;171:252-262. of the breast. Am J Surg Pathol. 2003;27:1052-1060.
63. Molyneux G, Geyer FC, Magnay FA, et al. BRCA1 basal-like 84. Salarieh A, Sneige N. Breast carcinoma arising in
breast cancers originate from luminal epithelial progenitors microglandular adenosis: a review of the literature. Arch
and not from basal stem cells. Cell Stem Cell. 2010;7:403-417. Pathol Lab Med. 2007;131:1397-1399.
64. Bryan BB, Schnitt SJ, Collins LC. Ductal carcinoma in situ 85. Shin SJ, Simpson PT, Da Silva L, et al. Molecular evidence
with basal-like phenotype: a possible precursor to invasive for progression of microglandular adenosis (MGA) to
basal-like breast cancer. Mod Pathol. 2006;19:617-621. invasive carcinoma. Am J Surg Pathol. 2009;33:496-504.
65. Paredes J, Lopes N, Milanezi F, et al. P-cadherin and 86. Wetterskog D, Lopez-Garcia MA, Lambros MB, et al.
cytokeratin 5: useful adjunct markers to distinguish basal-like Adenoid cystic carcinomas constitute a genomically distinct
ductal carcinomas in situ. Virchows Arch. 2007;450:73-80. subgroup of triple-negative and basal-like breast cancers.
J Pathol. 2012;226:84-96.
66. Tamimi RM, Baer HJ, Marotti J, et al. Comparison of

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

molecular phenotypes of ductal carcinoma in situ and 87. Curtis C, Shah SP, Chin SF, et al. The genomic and
invasive breast cancer. Breast Cancer Res. 2008;10:R67. transcriptomic architecture of 2,000 breast tumours reveals
novel subgroups. Nature. 2012;486:346-352.
67. Vincent-Salomon A, Lucchesi C, Gruel N, et al. Integrated
genomic and transcriptomic analysis of ductal carcinoma in 88. Stephens PJ, Tarpey PS, Davies H, et al. The landscape
situ of the breast. Clin Cancer Res. 2008;14:1956-1965. of cancer genes and mutational processes in breast cancer.
Nature. 2012;486:400-404.
68. Dabbs DJ, Chivukula M, Carter G, et al. Basal phenotype of
ductal carcinoma in situ: recognition and immunohistologic 89. Hammond ME, Hayes DF, Dowsett M, et al.
profile. Mod Pathol. 2006;19:1506-1511. American Society of Clinical Oncology/College of
American Pathologists guideline recommendations for
69. Meijnen P, Peterse JL, Antonini N, et al. immunohistochemical testing of estrogen and progesterone
Immunohistochemical categorisation of ductal carcinoma in receptors in breast cancer. Arch Pathol Lab Med.
situ of the breast. Br J Cancer. 2008;98:137-142. 2010;134:907-922.
70. Hannemann J, Velds A, Halfwerk JB, et al. Classification of 90. Wolff AC, Hammond ME, Schwartz JN, et al. American
ductal carcinoma in situ by gene expression profiling. Breast Society of Clinical Oncology/College of American
Cancer Res. 2006;8:R61. Pathologists guideline recommendations for human epidermal
71. Nik-Zainal S, Van Loo P, Wedge DC, et al. The life history growth factor receptor 2 testing in breast cancer. Arch Pathol
of 21 breast cancers. Cell. 2012;149:994-1007. Lab Med. 2007;131:18-43.
72. Geyer FC, Weigelt B, Natrajan R, et al. Molecular analysis 91. Grimm EE, Schmidt RA, Swanson PE, et al. Achieving
reveals a genetic basis for the phenotypic diversity of 95% cross-methodological concordance in HER2 testing:
metaplastic breast carcinomas. J Pathol. 2010;220:562-573. causes and implications of discordant cases. Am J Clin Pathol.
73. Ding L, Ellis MJ, Li S, et al. Genome remodelling in a 2010;134:284-292.
basal-like breast cancer metastasis and xenograft. Nature. 92. Hammond ME, Hayes DF, Wolff AC. Clinical notice for
2010;464:999-1005. American Society of Clinical Oncology-College of American
74. Ding L, Ley TJ, Larson DE, et al. Clonal evolution in Pathologists guideline recommendations on ER/PgR and
relapsed acute myeloid leukaemia revealed by whole-genome HER2 testing in breast cancer. J Clin Oncol. 2011;29:e458.
sequencing. Nature. 2012;481:506-510. 93. Allison KH, Dintzis SM, Schmidt RA. Frequency of HER2
75. Schmitt MW, Prindle MJ, Loeb LA. Implications of genetic heterogeneity by fluorescence in situ hybridization according
heterogeneity in cancer. Ann N Y Acad Sci. 2012;1267:110-116. to CAP expert panel recommendations: time for a new
look at how to report heterogeneity. Am J Clin Pathol.
76. Navin N, Kendall J, Troge J, et al. Tumour evolution inferred 2011;136:864-871.
by single-cell sequencing. Nature 2011;472:90-94.
94. Yang YL, Fan Y, Lang RG, et al. Genetic heterogeneity
77. Shah SP, Morin RD, Khattra J, et al. Mutational evolution of HER2 in breast cancer: impact on HER2 testing and its
in a lobular breast tumour profiled at single nucleotide clinicopathologic significance. Breast Cancer Res Treat.
resolution. Nature. 2009;461:809-813. 2012;134:1095-1102.
78. Burrell RA, Juul N, Johnston SR, et al. Targeting 95. Chang MC, Malowany JI, Mazurkiewicz J, et al. ‘Genetic
chromosomal instability and tumour heterogeneity in HER2- heterogeneity’ in HER2/neu testing by fluorescence in
positive breast cancer. J Cell Biochem. 2010;111:782-790. situ hybridization: a study of 2,522 cases. Mod Pathol.
79. Greaves M, Maley CC. Clonal evolution in cancer. Nature. 2012;25:683-688.
2012;481:306-313. 96. Ohlschlegel C, Zahel K, Kradolfer D, et al. HER2
80. Amir E, Miller N, Geddie W, et al. Prospective study genetic heterogeneity in breast carcinoma. J Clin Pathol.
evaluating the impact of tissue confirmation of metastatic 2011;64:1112-1116.
disease in patients with breast cancer. J Clin Oncol. 97. Yang J, Luo H, Li Y, et al. Intratumoral heterogeneity
2012;30:587-592. determines discordant results of diagnostic tests for human
81. Lindstrom LS, Karlsson E, Wilking UM, et al. Clinically used epidermal growth factor receptor (HER) 2 in gastric cancer
breast cancer markers such as estrogen receptor, progesterone specimens. Cell Biochem Biophys. 2012;62:221-228.
receptor, and human epidermal growth factor receptor 2 98. Bernasconi B, Chiaravalli AM, Finzi G, et al. Genetic
are unstable throughout tumor progression. J Clin Oncol. heterogeneity in HER2 testing may influence therapy
2012;30:2601-2608. eligibility. Breast Cancer Res Treat. 2012;133:161-168.
82. National Comprehensive Cancer Network. NCCN Clinical 99. Bartlett AI, Starcyznski J, Robson T, et al. Heterogeneous
Practice Guidelines in Oncology: Breast Cancer Version HER2 gene amplification: impact on patient outcome
3.2012. Available at www.nccn.org/professionals/physician_ and a clinically relevant definition. Am J Clin Pathol.
gls/recently_updated.asp. Accessed October 16, 2012. 2011;136:266-274.

© American Society for Clinical Pathology Am J Clin Pathol 2012;138:770-780 779

779 DOI: 10.1309/AJCPIV9IQ1MRQMOO 779
Allison / Molecular Pathology of Breast Cancer

100. Ross JS, Hatzis C, Symmans WF, et al. Commercialized 114. Mook S, Schmidt MK, Weigelt B, et al. The 70-gene
multigene predictors of clinical outcome for breast cancer. prognosis signature predicts early metastasis in breast
Oncologist. 2008;13:477-493. cancer patients between 55 and 70 years of age. Ann Oncol.
101. Mook S, Schmidt MK, Rutgers EJ, et al. Calibration and 2010;21:717-722.
discriminatory accuracy of prognosis calculation for breast 115. Paik S, Kim CY, Song YK, et al. Technology insight:
cancer with the online Adjuvant! program: a hospital-based application of molecular techniques to formalin-fixed
retrospective cohort study. Lancet Oncol. 2009;10:1070-1076. paraffin-embedded tissues from breast cancer. Nat Clin Pract
102. Albain KS. Prognostic and predictive value of the 21-gene Oncol. 2005;2:246-254.
recurrence score assay in postmenopausal women with 116. Allison KH, Kandalaft PL, Sitlani CM, et al. Routine
node-positive, oestrogen-receptor-positive breast cancer on pathologic parameters can predict Oncotype DX recurrence
chemotherapy: a retrospective analysis of a randomised trial. scores in subsets of ER positive patients: who does not always
Lancet Oncol. 2010;11:55-65. need testing? Breast Cancer Res Treat. 2012;131:413-424.
103. Paik S, ShakS, Tang G, et al. A multigene assay to predict 117. Barton S, Zabaglo L, A’Hern R, et al. Assessment of the

Downloaded from https://academic.oup.com/ajcp/article-abstract/138/6/770/1765928 by guest on 23 January 2019

recurrence of tamoxifen-treated, node-negative breast cancer. contribution of the IHC4+C score to decision making
N Engl J Med. 2004;351:2817-2826. in clinical practice in early breast cancer. Br J Cancer.
104. Habel LA. A population-based study of tumor gene 2012;106:1760-1765.
expression and risk of breast cancer death among lymph 118. Yamamoto-Ibusuki M, Yamamoto Y, Yamamoto S, et
node-negative patients. Breast Cancer Res. 2006;8:R25. al. Comparison of prognostic values between combined
105. Esteva FJ, Sahin AA, Cristofanilli M, et al. Prognostic role of immunohistochemical score of estrogen receptor,
a multigene reverse transcriptase-PCR assay in patients with progesterone receptor, human epidermal growth factor
node-negative breast cancer not receiving adjuvant systemic receptor 2, Ki-67 and the corresponding gene expression
therapy. Clin Cancer Res. 2005;11:3315-3319. score in breast cancer [published online ahead of print August
24, 2012]. Mod Pathol.
106. Cuzick J, Dowsett M, Pineda S, et al. Prognostic value
of a combined estrogen receptor, progesterone receptor, 119. Harris L, Fritsche H, Mennel R, et al. American Society
Ki-67, and human epidermal growth factor receptor 2 of Clinical Oncology 2007 update of recommendations for
immunohistochemical score and comparison with the the use of tumor markers in breast cancer. J Clin Oncol.
Genomic Health recurrence score in early breast cancer. 2007;25:5287-5312.
J Clin Oncol. 2011;29:4273-4278. 120. Flanagan MB, Dabbs DJ, Brufsky AM, et al. Histopathologic
107. Gianni L, Zambetti M, Clark K, et al. Gene expression variables predict Oncotype DX recurrence score. Mod Pathol.
profiles in paraffin-embedded core biopsy tissue predict 2008;21:1255-1261.
response to chemotherapy in women with locally advanced 121. Kelly CM, Krishnamurthy S, Bianchini G, et al. Utility of
breast cancer. J Clin Oncol. 2005;23:7265-7277. oncotype DX risk estimates in clinically intermediate risk
108. Mook S, Van’t Veer LJ, Rutgers EJ, et al. Individualization hormone receptor-positive, HER2-normal, grade II, lymph
of therapy using MammaPrint: from development to the node-negative breast cancers. Cancer. 2010;116:5161-5167.
MINDACT Trial. Cancer Genomics Proteomics. 2007;4:147- 122. Gwin K, Pinto M, Tavassoli FA. Complementary value of
155. the Ki-67 proliferation index to the oncotype DX recurrence
109. Paik S. Gene expression and benefit of chemotherapy in score. Int J Surg Pathol. 2009;17:303-310.
women with node-negative, estrogen receptor-positive breast 123. Masuda N, Ohnishi T, Kawamoto S, et al. Analysis of
cancer. J Clin Oncol. 2006;24:3726-3734. chemical modification of RNA from formalin-fixed samples
110. Sparano JA. TAILORx: Trial Assigning Individualized and optimization of molecular biology applications for such
Options for Treatment (Rx). Clin Breast Cancer. 2006;7:347- samples. Nucleic Acids Res. 1999;27:4436-4443.
350. 124. Vance GH, Barry TS, Bloom KJ, et al. Genetic heterogeneity
111. Straver ME. The 70-gene signature as a response predictor for in HER2 testing in breast cancer: panel summary and
neoadjuvant chemotherapy in breast cancer. Breast Cancer guidelines. Arch Pathol Lab Med. 2009;133:611-612.
Res Treat. 2010;119:551-558. 125. Starczynski J, Atkey N, Connelly Y, et al. HER2 gene
112. Wittner BS. Analysis of the MammaPrint breast cancer assay amplification in breast cancer: a rogues’ gallery of challenging
in a predominantly postmenopausal cohort. Clin Cancer Res. diagnostic cases: UKNEQAS interpretation guidelines and
2008;14:2988-2993. research recommendations. Am J Clin Pathol. 2012;137:595-
605.
113. Mook S, Schmidt MK, Viale G, et al. The 70-gene prognosis-
signature predicts disease outcome in breast cancer patients
with 1-3 positive lymph nodes in an independent validation
study. Breast Cancer Res Treat. 2009;116:295-302.

780 Am J Clin Pathol 2012;138:770-780 © American Society for Clinical Pathology

780 DOI: 10.1309/AJCPIV9IQ1MRQMOO