NEW AND IMPORTANT CHANGES IN THE TNM SYSTEM

New and Important Changes in the TNM Staging System for

Breast Cancer
Gabriel N. Hortobagyi, MD, Stephen B. Edge, MD, and Armando Giuliano, MD

OVERVIEW
Expanded understanding of biologic factors that modulate the clinical course of malignant disease have led to the gradual
integra on of biomarkers into staging classifica ons. The American Joint Commi ee on Cancer (AJCC) TNM staging system
is universally used and has largely displaced other staging classifica ons for most, although not all, cancers. Many of the
chapters of the eighth edi on of the AJCC TNM staging system integrated biomarkers with anatomic defini ons. The Breast
Chapter added estrogen receptor (ER) and progesterone receptor (PR) expression, HER2 expression, and/or amplifica on
and histologic grade to the anatomic assessment of tumor size, regional lymph node involvement, and distant metastases
(known as TNM). While preserving an anatomic staging system for con nuity and for regions where modern biomarkers
are not always available, the eighth edi on emphasizes the increased prognos c precision of the clinical prognos c stage
groups and the pathologic prognos c stage groups. The clinical prognos c stage groups are applicable to all pa ents with
primary breast cancer before any treatment has been implemented, but require a clinical and imaging evalua on as well
as a biopsy with grade and available ER, PR, and HER2 results; the pathologic prognos c stage groups are applicable to all
pa ents treated with complete surgical excision as first treatment and also require a complete pathology report, grade,
and ER, PR, and HER2. Applying the pathologic prognos c stage groups to a large database of pa ents staged by basic TNM
groupings changed the stage grouping of almost 40% of pa ents. Grouping by pathologic prognos c stage groups led to a
be er prognos c distribu on of the group and more precise individual prognos ca on.

S taging classifica ons were developed to be er under-

stand the clinical behavior of specific malignancies, de-
termine prognosis, and enable physicians and their pa ents
to compare outcomes of similar groups of pa ents. In the
early part of the 20th century, mul ple staging classifica-
ons were developed by experienced clinicians as well as
professional organiza ons. Star ng in 1943, and for the next
10 years, Pierre Denoix, a French surgeon, devised a staging
system based on the dimensions of the primary tumor, the
presence and extent of regional lymph node metastases,
and the presence and absence of distant metastases.1 The
system was adopted by the Union Interna onale Contre le
Cancer (UICC) in 1968, and in 1977, the AJCC published its
first staging system based on the TNM concept.2 Since 1977,
the AJCC has developed a detailed and extensive staging sys-
tem that covers the spectrum of human malignancies and
has updated it seven mes, as new informa on about clini-
cal behavior and management strategies has become avail-
able. Such updates have been the result of mul disciplinary
delibera ons by teams of experts, including surgeons, med-
ical and radia on oncologists, pathologists, radiologists, and
tumor registry experts, including representa ves of the
AJCC and UICC. In 1987, with the publica on of the fourth
edi on of the UICC TNM Atlas, the UICC and AJCC TNM clas-
sifica ons were unified. The eighth edi on of this staging
system, which became effec ve January 1, 2018, is the most
drama c departure from previous staging classifica on.3 In
several tumor types, the new staging system has incorpo-
rated biomarkers that modify the anatomic TNM classifica-
on. Although such changes were incorporated into staging
of a few cancers, including prostate, in the past couple of
edi ons, the paucity of appropriately performed analyses
of large clinical databases and relevantly forma ed results
in the peer-reviewed literature precluded the incorpora on
of most biomarkers into the staging system. In the mean-
me, the prac ce of oncology has moved forward, many
biomarkers have been incorporated into the process of se-
lec ng therapies, especially systemic therapies, and many
cancers have been subdivided into specific subtypes based
on these biomarkers.
Specifically in the case of breast cancer, it is now consid-
ered to be a conglomerate of at least four specific molecular

From the Department of Breast Medical Oncology, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX; Departments of Surgical
Oncology and Cancer Preven on and Control, Roswell Park Comprehensive Cancer Center, Buffalo, NY; Department of Surgery, Division of Surgical Oncology, John Wayne Cancer
Ins tute, Santa Monica, CA.

Disclosures of poten al conflicts of interest provided by the authors are available with the online ar cle at asco.org/edbook.

Corresponding author: Gabriel N. Hortobagyi, MD, The University of Texas MD Anderson Cancer Center, Unit 1354, PO Box 301439, Houston, TX 77230; email: ghortoba@
mdanderson.org.

© 2018 American Society of Clinical Oncology

asco.org/edbook | 2018 ASCO EDUCATIONAL BOOK  457

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
HORTOBAGYI, EDGE, AND GIULIANO
subtypes based on gene expression profiling: luminal A, lu-
minal B, HER2-enriched, and basal breast cancer.4 Because
gene expression profiling is not available to most prac cing
physicians, this molecular classifica on has been adapted to
clinical prac ce on the basis of frequently used biomarker
assays: ER, PR, HER2, and some measure of prolifera on,
usually the Ki-67 assay. As demonstrated by the evolu on of
classifica ons in hematologic malignancies (leukemias and
lymphomas), it is quite likely several more molecular and
clinical subtypes will be iden fied and validated in the breast
cancer field. In addi on, mul gene assays have been devel-
oped, using a variety of technologies, to provide prognos c
informa on for pa ents with early breast cancer.5,6 These
assays have also made major inroads into clinical prac ce
and have been validated by retrospec ve and prospec ve
analyses and, in some cases, by prospec ve controlled tri-
als.7,8 Thus, it became apparent that the AJCC staging system
needed to incorporate state-of-the-art biology and prognos-
c assays or be considered obsolete.
METHODS
In 2013, expert panels were created for each of the ma-
jor tumor types and charged with developing the best and
most effec ve staging system based on validated prognos-
c markers and technology that was widely available. The
panels were also instructed to keep the anatomic TNM
classifica on as a basic level of staging, so that areas of the
world where modern biomarker analyses were not being
performed rou nely, or at all, could s ll use the system. The
biomarker addi ons were envisioned as a second er of
prognos ca on, with genomic assays, if relevant, represent-
ing a third layer. The AJCC staff and members of the panels
for the seventh edi on had kept track of items that users of
the staging system had iden fied as in need of clarifica on
or upda ng. A list of such items was made available to each
panel. In addi on, each panel collected peer-reviewed pub-
PRACTICAL APPLICATIONS
• Biologic a ributes of tumor cells modulate clinical
course and therapeu c outcomes.
• Staging systems divide pa ent cohorts into dis nct
prognos c categories and allow more precise
comparison of pa ent cohorts, clinical trial results, and
therapeu c outcomes.
• ER and PR expressions by breast cancer cells iden fy
different types of breast cancer, with dis nct clinical
behavior and prognosis.
• Overexpression and amplifica on of HER2 is associated
with adverse prognosis, but it can be overcome with
specific an -HER2 targeted treatments.
• The eighth edi on AJCC staging system assumes all
pa ents receive state-of-the-art local, regional, and
systemic therapies.
• The use of pathologic prognos c stage groups is the
recommended staging system for North America, since it
provides more precise individual prognos ca on.
458 2018 ASCO EDUCATIONAL BOOK | asco.org/edbook
Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
lica ons that could form the basis for modifica ons of the
staging system. AJCC staff performed a review of the litera-
ture, and individual panelists contributed such publica ons
as they thought were relevant to the delibera ons of the
panel. This paper focuses on the changes that were imple-
mented in the breast staging system. However, considerable
changes have also been introduced into the staging systems
of many other tumor types.
The major issue under discussion was the integra on of
biomarkers into the staging system to improve prognos -
ca on. Because ER, PR, HER2, and Ki-67 are in widespread
use, these were the major focus of the discussion. Histologic
and nuclear grade were also discussed in great detail. The
panel approached organiza ons that controlled large, con-
temporary databases that included informa on about the
biomarkers under considera on and requested analyses to
determine the incremental benefit of adding each of the bio-
markers, as well as a group of biomarkers, to the basic TNM
staging system. These organiza ons included the Na onal
Cancer Database (NCDB), the Na onal Comprehensive Can-
cer Network, the Early Breast Cancer Trialists’ Collabora ve
Group, the California Cancer Registry, the Na onal Clinical
Trials Network of the Na onal Cancer Ins tute, and large
comprehensive cancer centers. Several of these groups un-
dertook detailed analysis of their databases and provided
invaluable informa on to the panel. The principal source
of informa on was a massive analysis based on 334,243
women diagnosed between 2010 and 2012 and included in
the NCDB. Others are s ll in the process of analyzing their
data, and, as the results of such analyses become available,
they will be considered in future panel discussions.
The breast panel met by monthly conference calls and had
a 2-day face-to-face mee ng. All issues raised by previous
users and major topics brought up by panel members were
addressed and resolved by consensus. The resul ng docu-
ment was placed on a whiteboard under password protec-
ons, and all panel members reviewed and made repeated
changes. The final document was approved by all panel
members and the AJCC execu ve leadership. Some illus-
tra ons from the seventh edi on were retained, whereas
other new ones were added.
CHANGES TO THE AJCC/UICC STAGING
SYSTEM INCLUDED IN THE EIGHTH EDITION
A number of changes reflected the need to clarify previ-
ously included defini ons and approaches to specific stag-
ing circumstances. Emerging scien fic evidence suggested
that lobular carcinoma in situ is a benign en ty, so it was
removed from the list of malignant tumors considered by
the Breast Chapter. Standard procedures for defining the di-
mensions of the primary tumor were addressed: (1) rounding
the size of very small tumors was discouraged, (2) T size in the
presence of mul ple tumor foci was clarified, and (3) a clear
defini on of satellite tumor nodules in the skin was included.
Clarifica ons to the N category were also added: (1) mea-
surement of nodal metastases was clearly defined; and (2)
cNX was further defined.

The designa on pM0 was determined to be invalid, whereas
cM1 and pM1 were reaffirmed.
The postneoadjuvant systemic therapy classifica on was
further elaborated: (1) determina on of ypT size was clar-
ified to exclude surrounding fibrosis, (2) determina on of
the dimensions of residual nodal metastases was restated,
and (3) the defini on of pathologic complete remission was
revisited, and clarifica on was made of pathologic complete
remission in the presence of M1 disease.
The major modifica ons were based on the integra on
of four biomarkers into the staging system: ER, PR, HER2,
and grade (No ngham Grading System). The panel deter-
mined that whenever possible, all invasive cancers should
have determina on of histologic grade and assays to mea-
sure the expression of ER, PR, and HER2, following broadly
accepted guidelines, such as those publicized by the ASCO/
College of American Pathologists collabora on.9,10 A second
er of prognos ca on, using informa on about grade, ER,
PR, and HER2, was added to the anatomic information of
TN and M: the resulting, combined categories were called
prognostic stage groups. Clinical prognostic stage groups
were developed for use in all patients with primary breast
cancer when all information about anatomic extent and
biomarkers is available. The clinical prognostic groups are
based on baseline assessment, before any therapeutic in-
tervention has been initiated. It uses clinical assessment
of dimensions, based on physical examination and imag-
ing (mammography, ultrasound, and/or MRI). The clinical
prognostic stage groups can also be applied to patients
who are initially treated with neoadjuvant systemic ther-
apy, chemotherapy, or endocrine therapy; when com-
pared with the baseline clinical prognostic stage, this will
provide an accurate assessment of clinical response prior
to surgical therapy.
For pa ents who undergo defini ve surgical resec on as
their ini al treatment modality, the new system includes
pathologic prognos c staging. This is based on pathologic
evalua on of extent of disease (T and N), added to bio-
marker informa on. The panel considered this to be the
most accurate and precise staging system to be used when-
ever all of the informa on was available (Table 1).11 The
pathologic prognos c staging should not be used for pa-
ents who received neoadjuvant systemic therapy before
surgery. Rather, the postneoadjuvant (ypT and ypN) staging
system should be used for these pa ents. Because there are
no large enough databases of pa ents who have received
neoadjuvant systemic therapy and have complete informa-
on about ypTNM and biomarkers, no prognos c staging
system has been developed for this popula on. The panel
will con nue to assess the availability of appropriate data-
bases and relevant analyses to develop such a prognos c
staging system in the future.
Although the tables associated with these changes in
the AJCC staging system are large and complex, it should be
relatively easy to find an individual patient’s prognostic
stage group by star ng from the le of the table and grad-
ually proceeding to the right, using the pa ent’s specific
Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
NEW AND IMPORTANT CHANGES IN THE TNM SYSTEM
staging informa on. In addi on, the AJCC is in the process of
developing a staging calculator that can be used as a stand-
alone applica on or incorporated into other electronic record
and registry systems to further facilitate the determina on
of an individual’s prognos c stage and, therefore, prognosis.
The addi on of biomarkers to the TNM staging system
has a major effect on prognos ca on. It is es mated that
upwards of 40% of pa ents classified by anatomic TNM
groupings will change at least one stage grouping when the
prognos c stage groups are used.
INCORPORATION OF MULTIGENE PANELS
INTO THE STAGING SYSTEM
Over the past couple of decades, and based on the com-
ple on of the Human Genome Project, several groups have
developed mul gene panels that have been shown to pro-
vide more accurate individual prognos c informa on.5,6,12-15
These panels have been used in the determina on of prog-
nosis in exis ng tumor collec ons, mostly tumor banks, al-
though some used pa ent samples from prospec ve clinical
trials. These commercially available panels can reproducibly
iden fy pa ents with be er and worse prognosis a er ini-
al treatment with cura ve intent. Most panels were de-
veloped for hormone receptor–posi ve, HER2-nega ve
tumors, although MammaPrint was developed in an unse-
lected group of pa ents with breast cancer. Most of the clin-
ical valida on has taken place in pa ent groups with lymph
node–nega ve breast cancer, although informa on based
on lymph node–posi ve breast cancer is star ng to appear
in the peer-reviewed literature.
MammaPrint was recently tested in a prospec ve clinical
trial (MINDACT) in which risk of recurrence was assessed
by MammaPrint and by clinic-pathologic methods (Adju-
vant! Online).8 Pa ents with hormone receptor–posi ve,
HER2-nega ve, node-nega ve or -posi ve, low-risk tumors
by both methods were assigned to adjuvant endocrine ther-
apy only; pa ents with high-risk tumors by both methods
were assigned to chemotherapy, whereas pa ents with dis-
crepant results in the two methods were randomly assigned
to use one method or the other to determine whether they
should get chemotherapy or endocrine therapy. At the first
analysis, MammaPrint iden fied a group of pa ents with
excellent prognosis who were quite unlikely to benefit from
chemotherapy.
The 21-gene assay (Oncotype DX) was the prognos c as-
say for the TAILORx clinical trial.7 Patients with hormone
receptor–posi ve, HER2-nega ve, node-nega ve breast can-
cer and a recurrence score (RS) lower than 11 were assigned
to endocrine therapy alone, whereas those with an RS 25
or higher were assigned to chemotherapy followed by en-
docrine therapy. Pa ents with an intermediate RS (11–24)
were randomly assigned to endocrine therapy alone or en-
docrine therapy plus chemotherapy. Only the low-risk group
has been reported: these pa ents had an outstanding result
(98.6% disease-free survival at 6.9 years) with endocrine
therapy only. The rest of the trial has not been reported.
Confirmatory reports from three large databases (Surveillance,
asco.org/edbook | 2018 ASCO EDUCATIONAL BOOK  459
HORTOBAGYI, EDGE, AND GIULIANO

TABLE 1. Pathologic Prognos c Stage Groups

Then the Pathologic

Prognos c Stage
When TNM Is… And Grade Is… And HER2 Status Is… And ER Status Is… And PR Status Is… Group Is…
Tis(c or p)N0M0, Any Any Any Any 0
*
T1 N0M0, Posi ve IA
T0N1miM0,  Posi ve
Nega ve IA
T1*N1miM0 Posi ve
Posi ve IA
Nega ve
Nega ve IA
1
Posi ve IA
Posi ve
Nega ve IA
Nega ve
Posi ve IA
Nega ve
Nega ve IA
Posi ve IA
Posi ve
Nega ve IA
Posi ve
Posi ve IA
Nega ve
Nega ve IA
2
Posi ve IA
Posi ve
Nega ve IA
Nega ve
Posi ve IA
Nega ve
Nega ve IB
Posi ve IA
Posi ve
Nega ve IA
Posi ve
Posi ve IA
Nega ve
Nega ve IA
3
Posi ve IA
Posi ve
Nega ve IA
Nega ve
Posi ve IA
Nega ve
Nega ve IB
Con nued

460 2018 ASCO EDUCATIONAL BOOK | asco.org/edbook

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
 NEW AND IMPORTANT CHANGES IN THE TNM SYSTEM

TABLE 1. Pathologic Prognos c Stage Groups (Cont'd)

Then the Pathologic

Prognos c Stage
When TNM Is… And Grade Is… And HER2 Status Is… And ER Status Is… And PR Status Is… Group Is…
T0N1**M0, T1N1**M0, Posi ve IA
T2N0M0 Posi ve
Nega ve IB
Posi ve
Posi ve IB
Nega ve
Nega ve IIA
1
Posi ve IA
Posi ve
Nega ve IB
Nega ve
Posi ve IB
Nega ve
Nega ve IIA
Posi ve IA
Posi ve
Nega ve IB
Posi ve
Posi ve IB
Nega ve
Nega ve IIA
2
Posi ve IA
Posi ve
Nega ve IIA
Nega ve
Posi ve IIA
Nega ve
Nega ve IIA
Posi ve IA
Posi ve
Nega ve IIA
Posi ve
Posi ve IIA
Nega ve
Nega ve IIA
3
Posi ve IB
Posi ve
Nega ve IIA
Nega ve
Posi ve IIA
Nega ve
Nega ve IIA
Con nued

asco.org/edbook | 2018 ASCO EDUCATIONAL BOOK  461

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
HORTOBAGYI, EDGE, AND GIULIANO

TABLE 1. Pathologic Prognos c Stage Groups (Cont'd)

Then the Pathologic

Prognos c Stage
When TNM Is… And Grade Is… And HER2 Status Is… And ER Status Is… And PR Status Is… Group Is…
T2N1M0, T3N0M0 Posi ve IA
Posi ve
Nega ve IIB
Posi ve
Posi ve IIB
Nega ve
Nega ve IIB
1
Posi ve IA
Posi ve
Nega ve IIB
Nega ve
Posi ve IIB
Nega ve
Nega ve IIB
Posi ve IB
Posi ve
Nega ve IIB
Posi ve
Posi ve IIB
Nega ve
Nega ve IIB
2
Posi ve IB
Posi ve
Nega ve IIB
Nega ve
Posi ve IIB
Nega ve
Nega ve IIB
Posi ve IB
Posi ve
Nega ve IIB
Posi ve
Posi ve IIB
Nega ve
Nega ve IIB
3
Posi ve IIA
Posi ve
Nega ve IIB
Nega ve
Posi ve IIB
Nega ve
Nega ve IIIA
Con nued

462 2018 ASCO EDUCATIONAL BOOK | asco.org/edbook

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
 NEW AND IMPORTANT CHANGES IN THE TNM SYSTEM

TABLE 1. Pathologic Prognos c Stage Groups (Cont'd)

Then the Pathologic

Prognos c Stage
When TNM Is… And Grade Is… And HER2 Status Is… And ER Status Is… And PR Status Is… Group Is…
T0N2M0, Posi ve IB
T1N2M0, T2N2M0, Posi ve
Nega ve IIIA
T3N1M0, T3N2M0 Posi ve
Posi ve IIIA
Nega ve
Nega ve IIIA
1
Posi ve IB
Posi ve
Nega ve IIIA
Nega ve
Posi ve IIIA
Nega ve
Nega ve IIIA
Posi ve IB
Posi ve
Nega ve IIIA
Posi ve
Posi ve IIIA
Nega ve
Nega ve IIIA
2
Posi ve IB
Posi ve
Nega ve IIIA
Nega ve
Posi ve IIIA
Nega ve
Nega ve IIIB
Posi ve IIA
Posi ve
Nega ve IIIA
Posi ve
Posi ve IIIA
Nega ve
Nega ve IIIA
3
Posi ve IIB
Posi ve
Nega ve IIIA
Nega ve
Posi ve IIIA
Nega ve
Nega ve IIIC
Con nued

asco.org/edbook | 2018 ASCO EDUCATIONAL BOOK  463

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
HORTOBAGYI, EDGE, AND GIULIANO

TABLE 1. Pathologic Prognos c Stage Groups (Cont'd)

Then the Pathologic

Prognos c Stage
When TNM Is… And Grade Is… And HER2 Status Is… And ER Status Is… And PR Status Is… Group Is…
T4N0M0, Posi ve IIIA
T4N1M0, Posi ve
Nega ve IIIB
T4N2M0, T(any)N3M0 Posi ve
Posi ve IIIB
Nega ve
Nega ve IIIB
1
Posi ve IIIA
Posi ve
Nega ve IIIB
Nega ve
Posi ve IIIB
Nega ve
Nega ve IIIB
Posi ve IIIA
Posi ve
Nega ve IIIB
Posi ve
Posi ve IIIB
Nega ve
Nega ve IIIB
2
Posi ve IIIA
Posi ve
Nega ve IIIB
Nega ve
Posi ve IIIB
Nega ve
Nega ve IIIC
Posi ve IIIB
Posi ve
Nega ve IIIB
Posi ve
Posi ve IIIB
Nega ve
Nega ve IIIB
3
Posi ve IIIB
Posi ve
Nega ve IIIC
Nega ve
Posi ve IIIC
Nega ve
Nega ve IIIC
T(any)N(any)M1 Any Any Any Any IV

*Includes T1mi.
**Does not include N1mi.
For cases in which HER2 is determined to be equivocal by in situ hybridiza on (ISH; fluorescence ISH or chromogenic ISH) tes ng under the 2013 ASCO/College of American Pathologists HER2 tes ng guide-
lines, the HER2-nega ve category should be used for staging in the pathologic prognos c stage group table. The prognos c value of these prognos c stage groups is based on popula ons of persons with
breast cancer who have been offered and mostly treated with appropriate endocrine and/or systemic chemotherapy (including an -HER2 therapy).
Reprinted with permission from Amin et al.3
Abbrevia ons: TNM, tumor node metastasis; ER, estrogen receptor; PR, progesterone receptor.

Epidemiology, and End Results [SEER],16 West German
Study Group PlanB,17 and Clalit18) indicate the highly re-
producible nature of the assay and provide reassurance
that pa ents with very low RS can safely forego adjuvant
chemotherapy.
These two assays have been used in tens of thousands or
hundreds of thousands of pa ents and provide the largest
proof of concept that they provide reproducible and reliable
individual prognos ca on. Other assays (PAM50,12 Breast
Cancer Index,13 IHC4,14 and EndoPredict15) also iden fy low-
and high-risk popula ons and could presumably contribute
to the biomarker-based prognos c staging system. Based
on these results, and on the preponderance of evidence in
the peer-reviewed literature, as well as a comprehensive re-
view of the relevant literature by a panel of the ASCO,19,20
the AJCC Breast Panel recommended that, when available
and indica ng a low-risk category, in pa ents with hormone
receptor–posi ve, HER2-nega ve invasive breast cancer
with a T1-2 primary tumor and nega ve lymph nodes, these
pa ents should be considered to have a stage IA breast can-
cer, regardless of the size of the tumor. For the current edi-
on of the AJCC Breast Cancer Staging System, the expert
panel applied this change to stage IA only for pa ents with a
low Oncotype DX score. However, the AJCC is commi ed to
more rapid re-evalua on and upda ng, and it is likely with
such updates the panel will include other genomic profiles
for this downstaging.
USE OF PROGNOSTIC SCORES
Another approach, perhaps somewhat simpler, to deter-
mine individual prognosis is based on informa on obtained
from mul variable analysis of a large database housed at

464 2018 ASCO EDUCATIONAL BOOK | asco.org/edbook

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
 NEW AND IMPORTANT CHANGES IN THE TNM SYSTEM

TABLE 2. Univariate and Mul variate Analysis of Prognos c Factors in Rela on to Disease-Specific Survival

Univariate Analysis Mul variate Analysis 2

5-Year DSS
Prognos c Factor (%) HR p HR p Assigned Points
Pathologic stage (seventh edi on)
 I 99.1 Referent Referent 0
 IIA 98.0 2.8 .002 2.3 .01 1
 IIB 95.6 4.8 < .0001 4.0 < .0001 2
 IIIA 95.4 6.8 < .0001 7.2 < .0001 3
 IIIC 79.5 26.6 < .0001 19.9 < .0001 4
Nuclear grade
 I 99.8 Referent Referent 0
 II 98.9 5.0 .1 4.0 .2 0
 III 95.3 25.0 .001 13.1 .01 1
ER status
 Posi ve 98.8 Referent Referent 0
 Nega ve 92.9 4.9 < .0001 2.5 .001 1
PgR status
 Posi ve 98.8 Referent Referent
 Nega ve 95.2 4.0 < .0001 NS
HER2 status
 Posi ve 97.5 Referent Referent 0
 Nega ve 98.0 0.8 .5 2.2 .04 1

Abbrevia on: DSS, disease-specific survival; ER, estrogen receptor; PgR, progesterone receptor.
Reprinted with permission from Amin et al.3

The University of Texas MD Anderson Cancer Center in
Houston, TX, which is an NCI-Designated Cancer Center. The
ins tu on developed an online, prospec ve database of all
pa ents with breast cancer treated or assessed at the in-
s tu on since 1997. The database provides detailed infor-
ma on about pa ent demographics, tumor characteris cs,
treatment modali es used, and outcomes. Informa on in
the database is updated about once a year. Inves gators
iden fied 3,728 pa ents treated with defini ve surgery and
appropriate adjuvant treatments between 1997 and 2006
and with complete biomarker informa on. Univariate and
multivariate analyses were performed to identify factors
associated with disease-specific survival (DSS). Variables asso-
ciated with outcomes were assigned points, and a prognos-
c model based on the sum of points was developed (Table
2).11 A model that included pathologic stage, ER, and grade
provided a highly precise prognos c system.21 Developed on
the MD Anderson Cancer Center database (Table 3),11 the
TABLE 3. Determina on of Risk Score/Profile
Factor 0 Points
Grade Grade 1/2
ER status ER-posi ve
HER2 status HER2-posi ve
Abbrevia on: ER, estrogen receptor.
prognos c score was validated on a cohort of pa ents of the
SEER database and subsequently on the California Cancer
Registry database (Table 4).22
CLINICAL RELEVANCE OF CHANGES TO THE
AJCC STAGING SYSTEM
A few clinical examples might illustrate the impact of these
changes on clinical prac ce.
• A 58-year-old schoolteacher developed a lump in the
right breast. By physical examination, it measured
3.5 × 4.0 cm. By imaging, the lesion measured 3.2 ×
3.6 cm. There was no palpable axillary lymphadenop-
athy. A percutaneous needle biopsy showed a grade 2
invasive ductal carcinoma (IDC), ER+, PR−, HER2−. The
pa ent underwent breast-conserving surgery, which
confirmed an IDC, measuring 3.0 × 3.5 cm. Sen nel
lymph node biopsy was nega ve. An Oncotype DX assay
was requested and showed an RS of 9. Using the basic
TNM staging system, the anatomic stage for this primary
breast cancer would be IIA (pT2N0M0). Applying the
new clinical prognos c stage groupings would also lead
1 Point to a stage IIA. The pathologic prognos c stage would be
Grade 3 IIA, and, with the result of the Oncotype DX RS (genomic
modifier), her tumor would be downstaged to IA.
ER-nega ve
• A 63-year-old homemaker developed a lump in the
HER2-nega ve
le breast. By physical examina on, it measured 7.5 ×
6.0 cm. By imaging, the lesion measured 8.2 × 6.6 cm.

asco.org/edbook | 2018 ASCO EDUCATIONAL BOOK  465

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
HORTOBAGYI, EDGE, AND GIULIANO

TABLE 4. Overall and Disease-Specific Survival Determined by Risk Score/Profile

Stage (Seventh No. of

Edi on) Risk Profile Pa ents 5-Year DSS (%) 95% CI (%) 5-Year OS (%) 95% CI (%)
I (IA and IB) 0 36 100 97.0 80.4–99.6
1 1,173 99.4 98.7–99.7 96.7 95.4–97.0
2 274 98.8 96.4–99.6 94.6 91.0–96.8
3 119 96.6 91.1–98.7 93.8 87.5–97.0
IIA 0 31 100 96.8 79.2–99.5
1 634 99.4 97.5–99.8 97.1 94.7–98.4
2 236 97.5 93.2–99.1 94.1 88.7–97.0
3 98 91.0 81.8–95.7 88.2 78.5–93.8
IIB 0 11 100 100
1 309 96.9 92.6–98.8 94.6 89.6–97.2
2 107 92.9 83.6–97.1 89.3 80.1–94.4
3 40 91.5 75.6–97.2 91.5 75.6–97.2
IIIA 0 3 100 100
1 134 98.3 88.2–99.8 91.5 82.6–96.0
2 50 92.2 77.2–97.5 90.3 75.7–96.3
3 7 68.6 21.3–91.2 68.6 21.3–91.2
IIIC 0 0
1 39 92.2 72.1–98.0 84.4 63.7–93.9
2 16 80.8 51.4–93.4 80.8 51.4–93.4
3 10 33.3 6.3–64.6 33.3 6.3–64.6

Abbrevia ons: DSS, disease-specific survival; OS, overall survival.

There was one palpable axillary node measuring 1.5 × (pT1N1M0), but her clinical prognos c stage would be
1.5 cm. Biopsy showed a grade 1 IDC, ER+, PR+, HER2+. IA, and her pathologic prognos c stage would be IA.
Breast-conserving surgery confirmed an IDC measuring This would again have major implica ons on the selec-
8.0 × 6.5 cm. Sen nel lymph node biopsy was posi ve. on of op mal adjuvant therapy.

The anatomic stage of this tumor would be IIIA (pT3N1M0), CONCLUSION

the clinical prognos c stage IIA, and the pathologic prog-
nostic stage IB. Note, however, that the change in prog-
nosis is based on the assump on that if it is appropriate
in her personal circumstance, she is offered and receives
systemic therapy based on the T, N, and biomarker status
of her cancer.
• A 72-year-old execu ve was found to have a mam-
mographic abnormality in the right breast. The lesion
was not detectable by physical examina on. By imag-
ing, the lesion measured 1.1 × 0.8 cm. There was no
palpable axillary node. Biopsy showed a grade 3 IDC,
ER−, PR−, HER2−. Breast-conserving surgery confirmed
an IDC, measuring 1.0 × 0.7 cm. Sen nel lymph node
biopsy was posi ve (0.4 cm). Axillary lymph node
dissec on not performed. Oncotype DX RS was not
performed (not indicated for triple-nega ve breast
cancer).
Anatomic stage: IIA (pT1N1M0); clinical prognos c stage:
IB; and pathologic prognos c stage: IIA.
• If this 72-year-old execu ve had been diagnosed with
the exact same cancer but different biomarker profile,
ER+, PR+, HER2+, her anatomic stage would s ll be IIA
The development of these prognos c models was a ma-
jor step forward for the AJCC staging system. However, it
is understood that this system will require ongoing revision
as informa on on biomarkers, genomic profiles, and treat-
ment evolves. The AJCC is commi ed to making such re-
visions more o en than the historical 6- to 8-year revision
cycle. The NCDB represents a very large group of pa ents, but,
when distributed into 120 possible stage groups, the num-
bers decrease drama cally. In addi on, the median follow-
up of this large cohort was only 41.7 months. Although
this provides a reliable preliminary analysis, longer follow-
up will be needed, in view of the rather protracted nature
of some breast cancer subtypes. This is par cularly true
for hormone receptor–posi ve breast cancers, for which
a 10-year follow-up is barely acceptable, and recurrences
con nue to occur for more than 20 years.23 Unfortunately,
the collec on of informa on about hormone receptors and
HER2 did not start in earnest un l 2010 in the SEER data-
base and the NCDB. Larger databases of pa ents treated
with state-of-the-art therapies, longer follow-up mes, and
confirmatory analyses from other large oncology databases
will help refine the modifica ons implemented in the eighth

466 2018 ASCO EDUCATIONAL BOOK | asco.org/edbook

Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.

edi on of the AJCC staging system. Something similar must
be said about the mul gene panels and their role in stag-
ing. Although the databases of pa ents whose tumors have
been tested with these assays is growing, the denominator
is usually much smaller when, in addi on to the results of
genomic assays, complete clinic-pathologic informa on,
biomarkers, and appropriate follow-up with outcomes is
References
1. Denoix PF. Importance of a classifica on common to the various forms
of cancer. Acta Radiol. 1954;116(suppl):122-129.
2. American Joint Commi ee for Cancer Staging and End Results
Repor ng. Manual for Staging of Cancer. Chicago, IL: American Joint
Commi ee; 1977.
3. Amin MB, Edge S, Greene F, et al (eds). AJCC Cancer Staging Manual,
8th Ed. Chicago, IL: Springer; 2017.
4. Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of human
breast tumours. Nature. 2000;406:747–752.
5. van de Vijver MJ, He YD, van’t Veer LJ, et al. A gene-expression
signature as a predictor of survival in breast cancer. N Engl J Med.
2002;347:1999–2009.
6. Paik S, Shak S, Tang G, et al. A mul gene assay to predict recurrence
of tamoxifen-treated, node-nega ve breast cancer. N Engl J Med.
2004;351:2817–2826.
7. Sparano JA, Gray RJ, Makower DF, et al. Prospec ve valida on of a 21-
gene expression assay in breast cancer. N Engl J Med. 2015;373:2005–
2014.
8. Cardoso F, van’t Veer LJ, Bogaerts J, et al; MINDACT Inves gators. 70-
Gene signature as an aid to treatment decisions in early-stage breast
cancer. N Engl J Med. 2016;375:717–729.
9. Wolff AC, Hammond MEH, Hicks DG, et al; College of American
Pathologists. Recommenda ons for human epidermal growth factor
receptor 2 tes ng in breast cancer: American Society of Clinical
Oncology/College of American Pathologists clinical prac ce guideline
update. J Clin Oncol. 2013;31:3997–4013.
10. Hammond MEH, Hayes DF, Dowse M, et al. American Society of Clinical
Oncology/College of American Pathologists guideline recommenda-
ons for immunohistochemical tes ng of estrogen and progesterone
receptors in breast cancer. J Clin Oncol. 2010;28:2784–2795.
11. American Joint Commi ee on Cancer. Updated breast chapter for 8th
edi on. https://cancerstaging.org/references-tools/deskreferences/
Pages/Breast-Cancer-Staging.aspx. Accessed February 8, 2018.
12. Nielsen TO, Parker JS, Leung S, et al. A comparison of PAM50 intrinsic
subtyping with immunohistochemistry and clinical prognos c factors
in tamoxifen-treated estrogen receptor-posi ve breast cancer. Clin
Cancer Res. 2010;16:5222–5232.
13. Jerevall PL, Ma XJ, Li H, et al. Prognos c u lity of HOXB13:IL17BR and
molecular grade index in early-stage breast cancer pa ents from the
Stockholm trial. Br J Cancer. 2011;104:1762–1769.
Downloaded from ascopubs.org by 101.255.127.33 on December 9, 2021 from 101.255.127.033
Copyright © 2021 American Society of Clinical Oncology. All rights reserved.
NEW AND IMPORTANT CHANGES IN THE TNM SYSTEM
sought. The development of more sophis cated mul gene
panels in triple-nega ve and HER2-enriched popula ons
would be a welcome addi on in the future. We have taken
the first steps in biology-driven staging and prognos ca on
in breast cancer. As our knowledge base expands and the
use of personalized cancer therapy grows, our staging sys-
tems will con nue to improve.
14. Sgroi DC, Sestak I, Cuzick J, et al. Predic on of late distant recurrence
in pa ents with oestrogen-receptor-posi ve breast cancer: a
prospec ve comparison of the breast-cancer index (BCI) assay, 21-
gene recurrence score, and IHC4 in the TransATAC study popula on.
Lancet Oncol. 2013;14:1067–1076.
15. Fitzal F, Filipits M, Rudas M, et al. The genomic expression test
EndoPredict is a prognos c tool for iden fying risk of local recurrence
in postmenopausal endocrine receptor-posi ve, her2neu-nega ve
breast cancer pa ents randomised within the prospec ve ABCSG 8
trial. Br J Cancer. 2015;112:1405–1410.
16. Petkov VI, Miller DP, Howlader N, et al. Breast-cancer-specific mortality
in pa ents treated based on the 21-gene assay: a SEER popula on-
based study. NPJ Breast Cancer. 2016;2:16017.
17. Nitz U, Gluz O, Christgen M, et al. Reducing chemotherapy use in
clinically high-risk, genomically low-risk pN0 and pN1 early breast
cancer pa ents: five-year data from the prospec ve, randomised
phase 3 West German Study Group (WSG) PlanB trial. Breast Cancer
Res Treat. 2017;165:573–583.
18. Stemmer SM, Steiner M, Rizel S, et al. Clinical outcomes in pa ents
with node-nega ve breast cancer treated based on the recurrence
score results: evidence from a large prospec vely designed registry.
NPJ Breast Cancer. 2017;3:33.
19. Harris LN, Ismaila N, McShane LM, et al; American Society of Clinical
Oncology. Use of biomarkers to guide decisions on adjuvant systemic
therapy for women with early-stage invasive breast cancer: American
Society of Clinical Oncology Clinical Prac ce Guideline. J Clin Oncol.
2016;34:1134–1150.
20. Krop I, Ismaila N, Andre F, et al. Use of biomarkers to guide decisions
on adjuvant systemic therapy for women with early-stage invasive
breast cancer: American Society of Clinical Oncology Clinical Prac ce
Guideline Focused Update. J Clin Oncol. 2017;35:2838–2847.
21. Chavez-MacGregor M, Mi endorf EA, Clarke CA, et al. Incorpora ng
tumor characteris cs to the American Joint Commi ee on Cancer
Breast Cancer Staging System. Oncologist. 2017;22:1292–1300.
22. Weiss A, Chavez-MacGregor M, Lichtensztajn DY, et al. Valida on
study of the American Joint Commi ee on Cancer Eighth Edi on
prognos c stage compared with the anatomic stage in breast cancer.
JAMA Oncol. Epub 2017 Dec 7.
23. Pan H, Gray R, Braybrooke J, et al; EBCTCG. 20-Year risks of breast-
cancer recurrence a er stopping endocrine therapy at 5 years. N Engl
J Med. 2017;377:1836-1846.
asco.org/edbook | 2018 ASCO EDUCATIONAL BOOK  467