J Clin Oncol (2014) Sentinel Lymph Node Biopsy for Patients With Early-Stage Breast Cancer- American Society of Clinical Oncology Clinical Practice Guideline Update

Sentinel Lymph Node Biopsy for Patients With Early-StageBreast Cancer: American Society of Clinical OncologyClinical Practice Guideline UpdateGary H. Lyman, Sarah Temin, Stephen B. Edge, Lisa A. Newman, Roderick R. Turner, Donald L. Weaver,Al B. Benson III, Linda D. Bosserman, Harold J. Burstein, Hiram Cody III, James Hayman, Cheryl L. Perkins,Donald A. Podoloff, and Armando E. Giuliano

Gary H. Lyman, Fred Hutchinson CancerResearch Center, University of Washing-ton, Seattle, WA; Sarah Temin, AmericanSociety of Clinical Oncology, Alexandria,VA; Stephen B. Edge, Baptist CancerCenter, Memphis, TN; Lisa A. Newmanand James Hayman, University of Michi-gan, Ann Arbor, MI; Roderick R. Turner,John Wayne Cancer Institute, Santa Moni-ca; Linda D. Bosserman, Wilshire OncologyMedical Group, Rancho Cucamonga;Armando E. Giuliano, Cedars-Sinai MedicalCenter, Los Angeles, CA; Donald L.Weaver, University of Vermont College ofMedicine and Vermont Cancer Center,Burlington, VT; Al B. Benson III, Northwest-ern University, Chicago, IL; HaroldJ. Burstein, Dana-Farber Cancer Institute,Boston, MA; Hiram Cody III, MemorialSloan Kettering Cancer Center, New York,NY; Cheryl L. Perkins, Patient Representa-tive, Dallas; and Donald A. Podoloff, Univer-sity of Texas MD Anderson Cancer Center,Houston, TX.

Published online ahead of print atwww.jco.org on March 24, 2014.

Clinical Practice Guideline Committeeapproval: November 19, 2013.

Editors note: This American Society ofClinical Oncology Clinical Practice GuidelineUpdate provides recommendations withreview and analysis of the relevant litera-ture for each recommendation. Additionalinformation, which may include datasupplements, slide sets, patient versions,clinical tools, and resources, is available atwww.asco.org/guidelines/breastsnb.

Authors disclosures of potential con-flicts of interest and author contribu-tions are found at the end of thisarticle.

Corresponding author: American Soci-ety of Clinical Oncology, 2318 Mill Rd,Suite 800, Alexandria, VA 22314;e-mail: [email protected].

2014 by American Society of ClinicalOncology

0732-183X/14/3213w-1365w/$20.00

DOI: 10.1200/JCO.2013.54.1177

A B S T R A C T

PurposeTo provide evidence-based recommendations to practicing oncologists, surgeons, and radiationtherapy clinicians to update the 2005 clinical practice guideline on the use of sentinel node biopsy(SNB) for patients with early-stage breast cancer.

MethodsThe American Society of Clinical Oncology convened an Update Committee of experts in medicaloncology, pathology, radiation oncology, surgical oncology, guideline implementation, and advo-cacy. A systematic review of the literature was conducted from February 2004 to January 2013 inMedline. Guideline recommendations were based on the review of the evidence by Up-date Committee.

ResultsThis guideline update reflects changes in practice since the 2005 guideline. Nine randomizedclinical trials (RCTs) met systematic review criteria for clinical questions 1 and 2; 13 cohort studiesinformed clinical question 3.

RecommendationsWomen without sentinel lymph node (SLN) metastases should not receive axillary lymph nodedissection (ALND). Women with one to two metastatic SLNs planning to undergo breast-conservingsurgery with whole-breast radiotherapy should not undergo ALND (in most cases). Women with SLNmetastases who will undergo mastectomy should be offered ALND. These three recommendation arebased on RCTs. Women with operable breast cancer and multicentric tumors, with ductal carcinomain situ (DCIS) who will undergo mastectomy, who previously underwent breast and/or axillary surgery,or who received preoperative/neoadjuvant systemic therapy may be offered SNB. Women who havelarge or locally advanced invasive breast cancer (tumor size T3/T4), inflammatory breast cancer, orDCIS (when breast-conserving surgery is planned) or are pregnant should not undergo SNB. Theserecommendations are based on cohort studies and/or informal consensus. In some cases, updatedevidence was insufficient to update previous recommendations.

J Clin Oncol 32:1365-1383. 2014 by American Society of Clinical Oncology

INTRODUCTION

TheAmerican Society ofClinicalOncology (ASCO)first published evidence-based clinical practiceguidelines on use of sentinel node biopsy (SNB) forpatients with early-stage breast cancer in a guidelinepublished in 2005. At the time of publication, therewas only one published randomized clinical trial(RCT), by Veronesi et al.1 There were no axillaryrecurrences in patients who had tumor-free nodesand did not undergo axillary lymph node dissection(ALND), and the short-termsurvivalwas similar forthose with tumor-free nodes treated with ALND.However, the study was underpowered for overall

survival (OS).ASCOguidelines areupdatedat inter-vals determined by an Update Committee of theoriginal Expert Panel. Since the publication of theoriginal guideline, additional randomized trial re-sults have become available. Practice has changed,but several issues remain unresolved, especiallywithregard to the accuracy of SNB in special circum-stances.A formal update of the systematic review forthis guideline and recommendations was con-ducted. This guideline summarizes the updated lit-erature search; it also reviews and analyzes newdata regarding the recommendations since thesystematic review for the previous update. Since2005, ASCO has updated some of its guideline

JOURNAL OF CLINICAL ONCOLOGY A S C O S P E C I A L A R T I C L E

VOLUME 32 NUMBER 13 MAY 1 2014

2014 by American Society of Clinical Oncology 1365

from 200.76.167.33Information downloaded from jco.ascopubs.org and provided by at INST NACIONAL DE CANCEROLOGIA on June 1, 2014

Copyright 2014 American Society of Clinical Oncology. All rights reserved.

THE BOTTOM LINE

ASCO GUIDELINE UPDATE

Recommendations for Sentinel Lymph Node Biopsy for Patients With Early-Stage Breast Cancer: ASCOClinical Practice Guideline Update

Guideline Question How should the results of sentinel node biopsy (SNB) be used in clinical practice? What is the role of SNB in special circumstances

in clinical practice? What are the potential benefits and harms associated with SNB?

Target Audience Medical oncologists, radiation oncologists, pathologists, surgeons, oncology nurses, patients/caregivers, and guideline implementers.

Methods A comprehensive systematic review of the literature was conducted, and an Update Committee was convened to review the evi-

dence and develop guideline recommendations. The guide for rating recommendations and strength of evidence is provided in theMethodology Supplement.

Recommendations Recommendation 1: Clinicians should not recommend axillary lymph node dissection (ALND) for women with early-stage breast

cancer who do not have nodal metastases. Type: evidence based; benefits outweigh harms. Evidence quality: high. Strength of rec-ommendation: strong.

Recommendation 2.1: Clinicians should not recommend ALND for women with early-stage breast cancer who have one or twosentinel lymph node metastases and will receive breast-conserving surgery (BCS) with conventionally fractionated whole-breastradiotherapy. Type: evidence based; benefits outweigh harms. Evidence quality: high. Strength of recommendation: strong.

Recommendation 2.2: Clinicians may offer ALND for women with early-stage breast cancer with nodal metastases found on SNBwho will receive mastectomy. Type: evidence based; benefits outweigh harms. Evidence quality: low. Strength of recommendation:weak.

Recommendation 3: Clinicians may offer SNB for women who have operable breast cancer who have the following circumstances: 3.1: Multicentric tumors. Type: evidence based; benefits outweigh harms. Evidence quality: intermediate. Strength of recommen-

dation: moderate. 3.2: Ductal carcinoma in situ (DCIS) when mastectomy is performed. Type: informal consensus; benefits outweigh harms. Evi-

dence quality: insufficient. Strength of recommendation: weak. 3.3: Prior breast and/or axillary surgery. Type: evidence based; benefits outweigh harms. Evidence quality: intermediate. Strength

of recommendation: strong. 3.4: Preoperative/neoadjuvant systemic therapy. Type: evidence based; benefits outweigh harms. Evidence quality: intermediate.

Strength of recommendation: moderate. Recommendation 4: There are insufficient data to change the 2005 recommendation that clinicians should not perform SNB for

women who have early-stage breast cancer and are in the following circumstances: 4.1: Large or locally advanced invasive breast cancers (tumor size T3/T4). Type: informal consensus. Evidence quality: insufficient.

Strength of recommendation: weak. 4.2: Inflammatory breast cancer. Type: informal consensus. Evidence quality: insufficient. Strength of recommendation: weak. 4.3: DCIS when breast-conserving surgery is planned. Type: informal consensus. Evidence quality: insufficient. Strength of recom-

mendation: strong. 4.4: Pregnancy. Type: informal consensus. Evidence quality: insufficient. Strength of recommendation: weak.

Qualifying Statements Clinicians may perform SNB for DCIS diagnosed by minimally invasive breast biopsy: one, when mastectomy is planned, because

this precludes subsequent SNB at a second operation; two, when physical examination or imaging shows a mass lesion highly sug-gestive of invasive cancer; or three, the area of DCIS by imaging is large ( 5 cm). SNB may be offered before or after neoadjuvantsystemic therapy (NACT), but the procedure seems less accurate after NACT. This update deleted a recommendation for patientshaving undergone prior nononcologic breast surgery or axillary surgery because of insufficient data to inform a recommendation.

(continued on following page)

Lyman et al

1366 2014 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY



methodology, including the phrasing of recommendations, which isreflected in this document.2 Data Supplement 7 provides both the2005 and 2013 recommendations.

Positive lymph nodes can contain metastases or be tumorinvolved, and the latter terms are more accurate; therefore, thisguideline uses the terms metastasis or metastatic. The term tumorfree is more accurate than the term negative and is therefore used.In addition, ALND is defined as level I or II axillary dissection.

GUIDELINE QUESTIONS

Overarching Clinical QuestionHow should the results of SNB be used in clinical practice, and

what are the potential benefits and harms associated with SNB?

Clinical Question 1Can ALND be avoided in patients who have tumor-free (ie,

negative) findings on SNB?

Clinical Question 2Is ALND necessary for all patients with metastatic findings

on SNB?Clinical Question 2.1. For women with metastatic sentinel

lymph nodes (SLNs) planning to undergo breast-conserving surgery(BCS) with whole-breast radiotherapy?

ClinicalQuestion2.2. Forwomenwithnodalmetastaseswhoareplanning to undergomastectomy?

Clinical Question 3What is the role of SNB in special circumstances in clinical prac-

tice (Data Supplement 8)?

METHODS

The recommendations were developed by an Update Committee (AppendixTable A1, online only) with multidisciplinary representation using a system-atic review of phase III RCTs, some observational studies, and clinical experi-ence as a guide.Most of the Clinical Question 1 and 2 recommendations wereevidencebased andusedpublications found in a literature search from2004 toJanuary 2013.

The Update Committee only considered observational data for specificcircumstances, and not all of the special circumstances were addressed in thisguideline update. In some selected caseswhere evidencewas lacking, but therewas a high level of agreement among Update Committee members, informalconsensus was used (as noted in the Bottom-Line Box).

Articles were selected for inclusion in the systematic review of the evi-dence if they met the following criteria:

Population: women with early-stage breast cancer. For Clinical Questions 1 and 2, fully published or recent meeting

presentations of English-language reports of phase III RCTs or rigor-ously conducted systematic reviews or meta-analyses. Trials with apopulation of women with early breast cancer that compared SNBwith the standard treatment of ALND; this included studies compar-ing SNB alone with SNB plus ALND, for those patients with nega-tive SLNs.

For special circumstances, prospective comparative cohort trials wereaccepted (criteria listed in Data Supplement 8).

Articles were excluded from the systematic review if they were: (1)meetingabstracts not subsequently published in peer-reviewed journals; (2) edito-rials, commentaries, letters, news articles, case reports, or narrative re-views; and (3) published in a language other than English. The guidelinerecommendations were crafted, in part, using Guidelines Into DecisionSupport (GLIDES) methodology.2 Ratings for the type and strength ofrecommendation, evidence, and potential bias are provided in the Meth-odology Supplement (www.asco.org/guidelines/breastsnb).

Detailed information about the methods used to develop this guidelineupdate, regarding the Update Committee composition, guideline develop-ment process, and steps taken in the systematic review and recommendationdevelopment process, is available in further detail in the Methodology andData Supplements at www.asco.org/guidelines/snbbreast.

Guideline DisclaimerTheClinicalPracticeGuidelines andotherguidancepublishedhereinare

provided by ASCO to assist providers in clinical decision making. The infor-mation herein should not be relied on as being complete or accurate, norshould it be considered as inclusive of all proper treatments ormethodsof careor as a statement of the standard of care. With the rapid development ofscientific knowledge, newevidencemay emergebetween the time informationis developed and when it is published or read. The information is not contin-ually updated andmay not reflect the most recent evidence. The informationaddresses only the topics specifically identified therein and is not applicable toother interventions, diseases, or stages of diseases. This information does notmandate any particular course ofmedical care. Furthermore, the informationis not intended to substitute for the independent professional judgment of thetreating provider, because the information does not account for individualvariation among patients. Recommendations reflect high, moderate, or lowconfidence that the recommendation reflects the net effect of a given course ofaction. The use ofwords likemust,must not, should, and should not indicatesthat a course of action is recommended or not recommended for either mostormany patients, but there is latitude for the treating physician to select othercourses of action in individual cases. In all cases, the selected course of actionshould be considered by the treating provider in the context of treating theindividual patient. Use of the information is voluntary. ASCO provides thisinformation on an as-is basis and makes no warranty, express or implied,regarding the information. ASCO specifically disclaims any warranties ofmerchantability or fitness for a particular use or purpose. ASCO assumes no

THE BOTTOM LINE (CONTINUED)

Additional Resources

More information, including a Data Supplement with additional evidence tables, a Methodology Supplement with information about

evidence quality and strength of recommendations, slide sets, and clinical tools and resources, is available at www.asco.org/guidelines/

breastsnb. Patient information is available at www.cancer.net.

SNB in Early-Stage Breast Cancer Update

www.jco.org 2014 by American Society of Clinical Oncology 1367



responsibility for any injury or damage topersons or property arising out of orrelated to any use of this information or for any errors or omissions.

Guideline and Relationships With CompaniesThe Update Committee was assembled in accordance with the ASCO

Conflicts of InterestManagement Procedures for Clinical Practice Guidelines(ie, Procedures, summarized at http://www.asco.org/rwc). Members of thecommittee completed theASCOdisclosure form,which requires disclosure offinancial and other interests that are relevant to the subject matter of theguideline, including relationshipswith commercial entities that are reasonablylikely to experience direct regulatory or commercial impact as a result ofpromulgation of the guideline. Categories for disclosure include employmentrelationships, consulting arrangements, stock ownership, honoraria, researchfunding, and expert testimony. In accordancewith the Procedures, themajor-ity of the members of the committee did not disclose any such relationships.

RESULTS

Assummarized inTable1andDataSupplement1(Tables1Aand2A),a total of nine RCTs were deemed eligible for inclusion in the system-atic review of the evidence for Clinical Questions 1 and 2; 13 cohortstudies were deemed eligible for Clinical Question 3, as presented inData Supplements 1 and 2 (Tables 1B, 2B, and 3).No phase II studies,meta-analyses, or other systematic reviews were found that met theASCOsystematic reviewcriteria for this guideline (DataSupplement8addresses special circumstances). These studies comprise the eviden-tiary basis of the guideline recommendations. The identified trialswere published between 2004 and 2013. The randomized trials com-pared similar interventions.Theprimaryoutcome for fourof the trialsfor Clinical Question 1 was therapeutic efficacy,3,6,9,13 as it was in twoof the trials for Clinical Question 2.4,12 Morbidities and quality of life(QOL) were the primary outcomes for the three other studies,5,7,8,11

although theywere framed inavarietyofways, suchas recurrence-freesurvival, event-free survival (EFS), all-causemortality, and so on. Thecohort studies for Clinical Question 3 reported a mix of efficacy andadverse effectrelated outcomes (Data Supplement 2). Characteristicsof the study participants are listed in Data Supplement 1 (Tables 2Aand 2B).

Study QualityAssummarized inTable2, studyqualitywas formally assessed for

the nine RCTs identified. Design aspects related to individual studyquality were assessed by one reviewer, with factors such as blinding,allocation concealment, placebo control, intention to treat, fundingsources, and soon, generally indicatinga lowto intermediatepotentialrisk of bias formost of the identified evidence. Follow-up times variedamong studies, lowering the comparability of the results. The Meth-odology Supplement provides for definitions of ratings for overallpotential risk of bias.

Outcomes and Adverse EventsDataonkeyoutcomesof interest andkeyadverse events are listed

in Table 1 (RCTs) and in Data Supplement 2 (cohort studies).

OVERARCHING CLINICAL QUESTION

How should the results of SNB be used in clinical practice?NineRCTswere found to inform this recommendation.Noneof

the trials showed any difference inmortality among participants who

underwent ALND or SNB for either those with lymph nodemetasta-ses or tumor-free (negative) SLNs. One trial showed statistically sig-nificant longer disease-free survival (DFS)12 with SNB, and one trialshowed a statistically significant noninferiority outcome in deaths inthose with metastatic nodes who did not undergo ALND.4 ClinicalQuestion 2 is divided into two subquestions for women with SLNmetastases compared with tumor-free nodes.

CLINICAL QUESTION 1

Can ALND be avoided in patients who have tumor-free (negative)SLNs?

RECOMMENDATION 1Clinicians should not recommendALND forwomenwith early-

stage breast cancer who do not have nodalmetastases. Type: evidencebased; benefits outweigh harms. Evidence quality: strong. Strength ofrecommendation: high.

Literature Review and AnalysisSeven RCTs have been published since the previous version of

this guideline15 that prospectively investigated whether ALND can beavoided inpatientswhohave tumor-freefindingsonSNB.These trialsinclude: NSABP (National Surgical Adjuvant Breast and Bowel Proj-ect) B32,13,14,16,17 ALMANAC (Axillary Lymphatic Mapping AgainstNodal Axillary Clearance),5,18,19 Sentinella/GIVOM (Gruppo Inter-disciplinare Veneto di Oncologia Mammaria),6,20,21 Canavese et al,3

RACS (Royal Australasian College of Surgeons)/SNAC (SentinelNode Versus Axillary Clearance),7,8,22,23 NCT00970983,9,10 and theCambridge/East Anglia Study Group.11 NSABP B32 produced fourpublications that met the systematic review selection criteria. Thesepublications included results onOS, recurrence, adverse events, tech-nical success, andperformance.TheALMANACstudypublished fourarticles, which reported morbidities, QOL, recurrence, and perfor-mance. Three reports from Sentinella/GIVOM published results onmorbidities, recurrence, OS/mortality, DFS, performance, adverseevents, and QOL. Canavese et al published one article that reportedrecurrence, OS/mortality, EFS, morbidities, and false-negative rates(FNRs). The RACS/SNAC study was published in four reports on aninterim analysis of the first 150 participants; results includedmorbid-ities, QOL, and FNRs. For NCT00970983, there were two articles,reporting recurrence, mortality/OS, DFS, and performance. TheCambridge/East Anglia Study Group published one article reportingon morbidities and QOL. (The IBCSG [International Breast CancerStudy Group] 23-01 is discussed under Clinical Question 2 becausemost of the patients had one to three micrometastases.)

Clinical OutcomesThis section summarizes the efficacy results from the trials pub-

lishedsince the systematic reviewfor the2005guideline for this clinicalquestion. Five trials reported results on clinical/efficacy out-comes.3,5,6,9,13 In four of those five trials, survival outcomes were theprimary end points.3,6,9,13 Of these four trials, NSABP B32 had thelargest number of participants (N 3,989). The second largest, Sen-tinella/GIVOM,had697participants, but the authors of this study feltit was too small to draw conclusions.

Lyman et al




Table

1.R

CTs

:O

utco

mes

Stu

dyC

ompa

rison

s

No.

ofP

atie

nts

Eva

luat

edor

Ran

dom

lyA

ssig

ned

Med

ian

Follo

w-U

pR

ecur

renc

eO

S/M

orta

lity

QO

Lan

dA

dver

seE

vent

sFN

R,

NP

V,

and

PP

VO

vera

llA

ccur

acy

Can

aves

eet

al3

(200

9)S

NB

ALN

D11

0of

124

5.5

year

s5-

year

EFS

(incl

udin

gre

curr

ence

),94

.5%

;95

%C

I,90

.9to

98.1

;P

.72

Ann

ualr

ate

ofev

ents

:16

.2pe

r1,

000

OS

,97

.2%

;95

%C

I,95

.4%

to98

.9%

NR

FNR

(ALN

D),

6.84

%93

%

SN

B

ALN

D(if

SLN

posi

tive)

115

of12

45-

year

EFS

(incl

udin

gre

curr

ence

),89

.8%

;95

%C

I,86

.9to

92.7

;P

.72

OS

,97

.2%

;95

%C

I,95

.4%

to98

.9%

NP

V,

91.1

%

Ann

ualr

ate

ofev

ents

,18

.6pe

r1,

000

P

.07

PP

V,

71.1

%N

oax

illar

yR

Rof

SN

Bv

SN

B

ALN

D,

0.87

;95

%C

I,0.

38to

2.01

;P

.741

Giu

liano

etal

4

(201

1;A

CO

SO

GZ0

011)

SN

Bal

one

436

of44

66.

3ye

ars

5-ye

arD

FS,

83.9

%;

95%

CI,

80.2

%to

87.9

%5-

year

deat

hs,

42of

426

QO

LN

RH

ighe

rfo

rA

LND

than

SLN

D-a

lone

grou

p(7

0%v

25%

;P

.001

)Ly

mph

edem

ain

ALN

Dgr

oup

sign

ifica

ntly

mor

eco

mm

onby

subj

ectiv

ere

port

(P

.001

)an

dby

obje

ctiv

eas

sess

men

tof

arm

circ

umfe

renc

e

NR

NR

5-ye

arLR

,1.

6%;

95%

CI,

0.7%

to3.

3%5-

year

LRR

-fre

esu

rviv

al,

96.7

%;

95%

CI,

94.7

%to

98.6

%;P

.28

5-ye

arO

S,

92.5

%;

95%

CI,

90.0

%to

95.1

%;P

.008

SN

B

ALN

D42

0of

445

5-ye

arD

FS,

82.2

%;

95%

CI,

78.3

%to

86.3

%5-

year

deat

hs,

52of

445

5-ye

arLR

,3.

1%;

95%

CI,

1.7%

to5.

2%5-

year

OS

,91

.5%

;95

%C

I,89

.1%

to94

.5%

5-ye

arLR

R-f

ree

surv

ival

,95

.7%

;95

%C

I,93

.6%

to97

.9%

;P

.28

HR

,0.

79;

90%

CI,

0.56

to0.

10

DFS

:H

R,

0.82

;95

%C

I,0.

58to

1.17

H

R,

0.88

;95

%C

I,0.

62to

1.25

H

R,

0.87

;90

%C

I,0.

62to

1.23

M

anse

let

al5

(200

6;A

LMA

NA

Ctr

ial)

SN

Bon

ly49

5of

478

12m

onth

sA

xilla

ry,

1A

t12

mon

ths,

seve

nde

aths

(tw

obr

east

canc

ersp

ecifi

c)

SN

Bgr

oup:

over

allp

atie

nt-r

ecor

ded

QO

Lst

atis

tical

lysi

gnifi

cant

lybe

tter

(P

.003

)FN

R,

6.7%

;19

of28

297

.6%

;78

2of

803

ALN

Don

ly47

6of

496

Axi

llary

,4

At

12m

onth

s,se

ven

deat

hs(t

wo

brea

stca

ncer

spec

ific)

SN

Bgr

oup:

over

allp

atie

nt-r

ecor

ded

QO

L;TO

I:1

mon

thaf

ter

surg

ery,

P

.001

;3,

6,an

d12

mon

ths

afte

rsu

rger

y,P

.001

NP

V,

NR

Diff

eren

ce,

2.7%

;95

%C

I,

1.5%

to7.

8%S

NB

grou

p:FA

CT-

B

4:ch

ange

insc

ore

from

base

line

to1

(P

.001

),3

(P

.001

),6

(P

.003

),12

(P

.002

),an

d18

mon

ths

(P

.003

)A

rmfu

nctio

ning

dete

riora

tion

grea

ter

inA

LND

than

inS

NB

grou

pat

each

stud

ytim

epo

int

(P

.001

)M

oder

ate

orse

vere

lym

phed

ema

repo

rted

mor

eof

ten

bypa

tient

sin

ALN

Dth

anin

SN

Bgr

oup

at1,

3,6,

and

12m

onth

saf

ter

surg

ery

(eg,

5%v

13%

at12

mon

ths;

P

.001

)R

R,

0.37

(95%

CI,

0.23

to0.

60)

at12

mon

ths

for

lym

phed

ema

for

SN

Bco

mpa

red

with

ALN

Dgr

oup

PP

V,

NR

(con

tinue

don

follo

win

gpa

ge)





Table

1.R

CTs

:O

utco

mes

(con

tinue

d)

Stu

dyC

ompa

rison

s

No.

ofP

atie

nts

Eva

luat

edor

Ran

dom

lyA

ssig

ned

Med

ian

Follo

w-U

pR

ecur

renc

eO

S/M

orta

lity

QO

Lan

dA

dver

seE

vent

sFN

R,

NP

V,

and

PP

VO

vera

llA

ccur

acy

SN

Bgr

oup:

mea

nch

ange

inar

mvo

lum

eat

12m

onth

s,1.

028;

95%

CI,

1.01

6to

1.03

9A

LND

grou

p:m

ean

chan

gein

arm

volu

me

at12

mon

ths,

1.02

8;95

%C

I,1.

016

to1.

039

Lym

phed

ema:

SN

Bv

ALN

D:

RR

,0.

37;

95%

CI,

0.23

to0.

60;

abso

lute

rate

,5%

v13

%S

enso

rylo

ss:

SN

Bv

ALN

D:

RR

,0.

37;

95%

CI,

0.27

to0.

50;

abso

lute

rate

,11

%v

31%

;P

.001

Hig

her

rate

ofse

nsor

yde

ficit

inA

LND

vS

NB

grou

p(b

yse

lf-as

sess

men

t,at

12m

onth

s),

31%

v11

%;

perc

ent

ofpa

tient

sw

hoha

d

one

arm

prob

lem

1m

onth

post

surg

ery

grea

ter

inS

NB

vA

LND

grou

pR

R,

0.37

;95

%C

I,0.

27to

0.50

;se

nsor

yde

ficit

at12

mon

ths

infa

vor

ofS

NB

grou

pIC

BN

nerv

edi

vide

din

22.6

%(8

8of

390)

ofA

LND

v5.

3%(2

1of

400)

ofS

NB

grou

p(P

.001

)SN

Bgr

oup:

drai

nus

age,

leng

thof

hosp

itals

tay,

and

time

tore

sum

ptio

nof

norm

alda

y-to

-day

activ

ities

afte

rsur

gery

wer

est

atis

tical

lysi

gnifi

cant

lylo

wer

(P

.001

)SN

Bgr

oup:

axilla

ryop

erat

ive

time

redu

ced

(P

.055

)Za

vagn

oet

al6

(200

8;S

entin

ella

/GIV

OM

)SN

B

ALN

D34

556

mon

ths

5-ye

arD

FS,

87.6

%;

95%

CI,

83.3

to90

.95-

year

OS

,95

.5%

;95

%C

I,92

.2to

97.5

SF-

36an

dP

GW

B:

nosi

gnifi

cant

diff

eren

ces

betw

een

grou

psaf

ter

24m

onth

sFN

R,

16.7

%(1

8of

108)

;95

%C

I,10

.2to

25.1

94.4

%(3

05of

323)

;95%

CI,

91.3

to96

.7D

ista

nt,

11of

345

LRR

,16

of34

5C

ontr

alat

eral

,2

of34

5

Dea

thre

sulti

ngfr

omno

n-B

Cca

uses

,11

of34

5

Lym

phed

ema

(SN

B/A

LND

):m

ean

OR

,0.

48;

95%

CI,

0.3

to0.

8;P

.01

PP

V,

NR

NP

V,

92.3

%(2

15of

233)

;95

%C

I,88

.1to

95.4

SN

B

ALN

D(if

SLN

posi

tive)

352

5-ye

arD

FS,

89.9

%;

95%

CI,

85.3

to93

.1;P

.769

5-ye

arO

S,

94.8

%;

95%

CI,

91.6

to96

.8

Num

bnes

s(S

NB

/ALN

D):

mea

nO

R,

0.51

;95

%C

I,0.

4to

0.7;

P

.001

Dis

tant

,16

of35

2LR

R,

3of

352

Con

tral

ater

al,

3of

352

Dea

thre

sulti

ngfr

omno

n-B

Cca

uses

,6

of35

2

Sho

ulde

rpa

in(S

NB

/ALN

D):

mea

nO

R,

0.74

;95

%C

I,0.

5to

1.1;

P

.11

Res

tric

tions

ofsh

ould

erm

obili

ty(S

NB

/ALN

D):

mea

nO

R,

0.55

;95

%C

I,0.

3to

0.9;

P

.016

Sm

ithet

al7

(200

9),

Ung

etal

8

(200

4;R

AC

S/

SN

AC

)

ALN

D31

9N

RN

RN

RN

RFN

R,

5%(in

terim

)N

RS

NB

ALN

D(if

SLN

posi

tive)

324

Arm

volu

me:

Clin

icia

n:4.

4%in

crea

sein

ALN

D;

2.9%

incr

ease

inS

NB

Pat

ient

:9.

5%in

crea

sein

ALN

D;

4.1%

incr

ease

inS

NB

Axi

llary

clea

ranc

e:C

linic

ian

ratin

g:ro

utin

eax

illar

ycl

eara

nce:

mea

n,4.

4;S

E,

0.4;

SLN

-bas

edm

anag

emen

t:m

ean,

2.9;

SE

,0.

4;m

ean

diff

eren

cebe

twee

ngr

oups

:m

ean,

1.5;

SE

,0.

6;M

WW

,P

.001

Pat

ient

repo

rt:

rout

ine

axill

ary

clea

ranc

e:m

ean,

4.4;

SE

,0.

4;S

LN-b

ased

man

agem

ent:

mea

n,2.

9;S

E,

0.4)

;M

ean

diff

eren

cebe

twee

ngr

oups

:m

ean,

1.5;

SE

,0.

6;M

WW

,P

.001

NP

V,

98%

(inte

rim)

PP

V,

95%

(inte

rim)

(con

tinue

don

follo

win

gpa

ge)

Lyman et al




Table

1.R

CTs

:O

utco

mes

(con

tinue

d)

Stu

dyC

ompa

rison

s

No.

ofP

atie

nts

Eva

luat

edor

Ran

dom

lyA

ssig

ned

Med

ian

Follo

w-U

pR

ecur

renc

eO

S/M

orta

lity

QO

Lan

dA

dver

seE

vent

sFN

R,

NP

V,

and

PP

VO

vera

llA

ccur

acy

Ver

ones

iet

al9,

10

(201

0;N

CT0

0970

983)

SN

B

ALN

D25

710

2 mon

ths

10-y

ear

EFS

,88

.8%

;95

%C

I,84

.6to

92.6

OS

,89

.7%

;95

%C

I,85

.5to

93.8

;P

.15

NR

FNR

,8%

;95

%C

I,3

to15

5-ye

arD

FS,

89.9

%;

95%

CI,

85.3

to93

.1;P

.769

Dea

ths

(BC

),14

of25

7D

ista

nt,

20of

257

LRR

,4

of25

7S

uper

clav

ian,

2of

257

Con

tral

ater

al,

10of

257

Dea

ths

(non

-BC

),9

of25

7

SN

B

ALN

D(if

SLN

posi

tive)

259

10-y

ear

EFS

,89

.9%

;95

%C

I,85

.9to

83.9

OS

,93

.5%

;95

%C

I,90

.3to

96.8

;P

.15

5-ye

arD

FS,

89.9

%;

95%

CI,

85.3

to93

.1;P

.769

Dea

ths

(BC

),11

of25

9D

ista

nt,

17of

259

LRR

,4

of25

9A

xilla

ry,

2of

259

Con

tral

ater

al,

9of

259

Dea

ths

(non

-BC

),4

of25

9

Pur

usho

tham

etal

11

(200

5;C

ambr

idge

/E

ast

Ang

liaS

tudy

Gro

up)

SN

Bal

one

(SLN

nega

tive)

8612

mon

ths

NR

NR

SF-

36:

mea

ndi

ffer

ence

,3.

7;95

%C

I,1.

2to

6.1;

P

.001

NR

NR

SN

B

ALN

D(c

ontr

ol)

155;

144

of15

5un

derw

ent

follo

w-u

p

InS

NB

ALN

D(if

SLN

posi

tive)

grou

p:P

hysi

calf

unct

ioni

ngsc

ore,

3.2;

95%

CI,

1.1

to5.

4;P

.003

(hig

her)

Vita

lity

scor

e,3.

7;95

%C

I,1.

1to

6.2;

P

.004

(hig

her)

SN

B

ALN

D(if

SLN

posi

tive)

143;

57of

143

com

pris

edco

ntro

lgr

oup;

139

of14

3un

derw

ent

follo

w-u

p

GS

Isc

ale:

som

atiz

atio

nlo

wer

(P

.001

)M

AC

scal

e:N

SD

epre

ssio

nan

dan

xiet

y:N

S

Lym

phed

ema:

mea

nO

R,

0.30

;95

%C

I,0.

18to

0.68

;P

.004

Diff

eren

cein

arm

volu

me

chan

ge:

mea

n,35

.4m

L;S

E,

12.2

;P

.004

Neu

rolo

gic/

sens

ory

defic

its:

over

allf

ewer

sens

ory

findi

ngs

with

SN

B

ALN

D;P

.001

Sho

ulde

rim

pairm

ent:

redu

ctio

nin

mob

ility

alw

ays

low

eron

aver

age

inst

udy

grou

p(le

ssim

pairm

ent)

;si

gnifi

cant

only

for

flexi

on(P

.04)

(con

tinue

don

follo

win

gpa

ge)





Table

1.R

CTs

:O

utco

mes

(con

tinue

d)

Stu

dyC

ompa

rison

s

No.

ofP

atie

nts

Eva

luat

edor

Ran

dom

lyA

ssig

ned

Med

ian

Follo

w-U

pR

ecur

renc

eO

S/M

orta

lity

QO

Lan

dA

dver

seE

vent

sFN

R,

NP

V,

and

PP

VO

vera

llA

ccur

acy

Gal

imbe

rtie

tal

12

(201

3;IB

CS

G23

-01)

ALN

Don

ly(a

fter

SN

B)

464

5.0

year

sD

FS,

84.4

%;

95%

CI,

80.7

to88

.1;

P

.004

25-

year

OS

,97

.5%

;95

%C

I,96

.0to

99.2

QO

LN

RLy

mph

edem

a:A

LND

,13

%(5

9of

464)

;no

ALN

D,

15%

(15

of46

7);P

.001

Sen

sory

neur

opat

hy:

ALN

D,

18%

(82

of46

3);

noA

LND

,12

%(5

5of

467)

;P

.012

Mot

orne

urop

athy

:A

LND

,8%

(37

of46

4);

noA

LND

,3%

(13

of46

7);P

.001

Infe

ctio

ns:

ALN

D,

1of

464

Pat

ient

sw

ithre

port

edlo

ng-t

erm

surg

ical

even

ts(g

rade

3-4)

incl

uded

one

sens

ory

neur

opat

hy(g

rade

3),

thre

ely

mph

oede

ma

(tw

ogr

ade

3an

don

egr

ade

4),

and

thre

em

otor

neur

opat

hy(g

rade

3),

alli

ngr

oup

unde

rgoi

ngA

LND

,an

don

egr

ade

3m

otor

neur

opat

hyin

grou

pw

ithou

tA

LND

;on

ese

rious

adve

rse

even

tw

asre

port

ed:

post

oper

ativ

ein

fect

ion

inax

illa

ingr

oup

with

ALN

D

NR

NR

LR,

10of

464

LRR

,11

of46

4R

egio

nal,

1of

464

Dis

tant

,34

of46

4A

xilla

ry,

1of

464

All-

caus

em

orta

lity/

deat

hs,

4%(1

9of

464)

No

ALN

D(a

fter

SN

B)

467

DFS

,87

.8%

;95

%C

I,84

.4to

91.2

;P

.004

25-

year

OS

,97

.5%

;95

%C

I,95

.8to

99.1

LRR

,13

of46

7LR

,8

of46

7R

egio

nal,

5of

467

Dis

tant

,25

of46

7A

xilla

ry,

4of

467

All-

caus

em

orta

lity/

deat

hs,

4%(1

7of

467)

HR

,0.

78;

95%

CI,

0.55

to1.

11;

P

.004

2(n

oA

LND

vA

LND

)H

R,

0.89

;90

%C

I,0.

52to

1.54

;P

.73

(con

tinue

don

follo

win

gpa

ge)

Lyman et al




Table

1.R

CTs

:O

utco

mes

(con

tinue

d)

Stu

dyC

ompa

rison

s

No.

ofP

atie

nts

Eva

luat

edor

Ran

dom

lyA

ssig

ned

Med

ian

Follo

w-U

pR

ecur

renc

eO

S/M

orta

lity

QO

Lan

dA

dver

seE

vent

sFN

R,

NP

V,

and

PP

VO

vera

llA

ccur

acy

Kra

get

al13

,14

(201

0;N

SA

BP

B32

;in

terim

repo

rt,

n

150)

SN

Bal

one

(ifS

LNne

gativ

e)

2,01

195

.6 mon

ths

(mea

n)

Tota

leve

nts,

16.7

%;

336

of2,

011

even

ts8-

yrO

S,

90.3

%;

95%

CI,

88.9

to91

.8

QO

LN

RA

llerg

icre

actio

n,8%

(blu

edy

e)S

houl

der

abdu

ctio

nde

ficits

10%

peak

edat

1w

eek

for

SN

B

ALN

D(7

5%)

and

SN

B(4

1%)

grou

psA

rmvo

lum

edi

ffer

ence

s

10%

at36

mon

ths

wer

eev

iden

tfo

rS

NB

ALN

D(1

4%)

and

SN

B(8

%)

grou

psN

umbn

ess

and

tingl

ing

peak

edat

6m

onth

sfo

rS

NB

ALN

D(4

9%an

d23

%)

and

SN

B(1

5%an

d10

%)

grou

ps

FNR

,38

.6%

(268

of69

4);

95%

CI,

35.0

to42

.4

97.1

%(2

,544

of2,

619)

;95

%C

I,96

.4%

to97

.7%

(of

75pa

tient

sw

ithne

gativ

eS

NB

with

node

-po

sitiv

eA

LND

)

DFS

,16

.7%

;33

6of

2,01

1H

R,

1.19

;95

%C

I,0.

95to

1.49

;P

.13

NP

V,

9.8%

(75

of76

6);

95%

CI,

7.8

to12

.1D

ista

nt,

3.2%

;64

of2,

011

PP

V,

NR

LR,

2.4%

;49

of20

11

Reg

iona

lnod

e,0.

7%;

14of

2,01

1E

FS,

83.3

%;

1,67

5of

2,01

1C

ontr

alat

eral

,2.

2%;

44of

2,01

1D

eath

s,8.

4%;

169

of2,

011

Dea

ths

(no

BC

),2.

8%;

56of

1,97

5S

NB

ALN

D1,

975

Tota

leve

nts,

16%

;31

5of

1,97

58-

year

OS

,91

.8%

;95

%C

I,90

.4to

93.3

D

FS,

315

of1,

975

Dis

tant

,2.

8%;

55of

1,97

5LR

,2.

7%;

54of

1,97

5

HR

,1.

20;

95%

CI,

0.96

to1.

50;

P

.12

Reg

iona

lnod

e,0.

4%;

8of

1,97

5C

ontr

alat

eral

,2.

8%;

56of

1,97

5E

FS,

84.1

%;

1,66

0of

1,97

5D

eath

s,7%

;14

0of

1,97

5D

eath

s(n

oB

C),

2.3%

;5

of19

75

Abb

revi

atio

ns:

AC

OS

OG

,A

mer

ican

Col

lege

ofS

urge

ons;

ALM

AN

AC

,A

xilla

ryLy

mph

atic

Map

ping

Aga

inst

Nod

alA

xilla

ryC

lear

ance

;A

LND

,ax

illar

yly

mph

node

diss

ectio

n;B

C,

brea

stca

ncer

;D

FS,

dise

ase-

free

surv

ival

;EFS

,eve

nt-f

ree

surv

ival

;FA

CT-

B

4,Fu

nctio

nalA

sses

smen

tof

Can

cer

Ther

apy

Bre

ast,

scal

eve

rsio

n4;

FNR

,fal

se-n

egat

ive

rate

;GIV

OM

,Gru

ppo

Inte

rdis

cipl

inar

eV

enet

odi

Onc

olog

iaM

amm

aria

;GS

I,G

loba

lSev

erity

Inde

x;H

R,h

azar

dra

tio;I

BC

SG

,Int

erna

tiona

lBre

ast

Can

cer

Stu

dyG

roup

;LR

,loc

alre

curr

ence

;LR

R,l

ocor

egio

nalr

ecur

renc

e;M

AC

,Men

talA

djus

tmen

tto

Can

cer;

NP

V,n

egat

ive

pred

ictiv

eva

lue;

NR

,not

repo

rted

;NS

,not

sign

ifica

nt;N

SA

BP

,Nat

iona

lSur

gica

lAdj

uvan

tB

reas

tan

dB

owel

Pro

ject

;OR

,odd

sra

tio;O

S,o

vera

llsu

rviv

al;P

GW

B,P

sych

olog

ical

Gen

eral

Wel

lBei

ngIn

dex;

PP

V,p

ositi

vepr

edic

tive

valu

e;Q

OL,

qual

ityof

life;

RA

CS

,Roy

alA

ustr

alas

ian

Col

lege

ofS

urge

ons;

RC

T,ra

ndom

ized

cont

rolle

dtr

ial;

RR

,ris

kra

tio;S

F-36

,Sho

rtFo

rm36

;SLN

,sen

tinel

lym

phno

de;S

NA

C,S

entin

elN

ode

Ver

sus

Axi

llary

Cle

aran

ce;

SN

B,

sent

inel

node

biop

sy;

TOI,

Tria

lOut

com

esIn

dex.

E

valu

ated

and/

orra

ndom

lyas

sign

edw

ithfo

llow

-up.

Inc

lude

sde

aths

.U

nadj

uste

dH

R.

Adj

uste

dH

Rfo

rad

juva

ntth

erap

y(c

hem

othe

rapy

,en

docr

ine

ther

apy,

and/

orra

diat

ion

ther

apy)

and

age.

Bec

kD

epre

ssio

nIn

vent

ory.

S

tate

-Tra

itA

nxie

tyIn

vent

ory

#Eur

opea

nO

rgan

isat

ion

for

Res

earc

han

dTr

eatm

ent

ofC

ance

rB

reas

tC

ance

rM

odul

e(E

OR

TC-Q

LMB

R23

).





Table

2.R

CTs

:Q

ualit

yA

sses

smen

t

Stu

dyA

dequ

ate

Ran

dom

izat

ion

Con

ceal

edA

lloca

tion

Suf

ficie

ntS

ampl

eS

ize

Com

para

ble

Gro

ups

Blin

ded

Val

idat

edan

dR

elia

ble

Mea

sure

sA

dequ

ate

Follo

w-U

pIn

sign

ifica

ntC

OIs

Ove

rall

Ris

kof

Bia

s

Can

aves

eet

al3

(200

9)X

?

X

X?

??

Giu

liano

etal

4(2

011;

AC

OS

OG

Z001

1)?

?X

XP

artia

llyX

?X

Low

Man

sele

tal

5(2

006;

ALM

AN

AC

tria

l)X

XX

X

Par

tially

XX

?Za

vagn

oet

al6

(200

8;S

entin

ella

/GIV

OM

)X

Par

tially

XX

P

artia

lly?

XLo

wS

mith

etal

7(2

009)

?

XX

X

?X

Inte

rmed

iate

Sm

ithet

al7

(200

9),

Ung

etal

8(2

004;

RA

CS

/SN

AC

)?

?X

?P

artia

lly

?

X?

Ver

ones

iet

al9,

10(2

010;

NC

T009

7098

3)X

XX

X

X?,

part

ially

XLo

wP

urus

hoth

amet

al11

(200

5;C

ambr

idge

/Eas

tA

nglia

Stu

dyG

roup

)X

XX

XP

artia

lly;

was

not

doub

lebl

ind

XX

XLo

w

Gal

imbe

rtie

tal

12(2

013;

IBC

SG

23-0

1)X

?P

artia

lX

X

XX

Inte

rmed

iate

Kra

get

al13

,14

(201

0;N

SA

BP

B32

)X

??

X?

XX

Not

rele

vant

Low

NO

TE.

Xin

dica

tes

crite

rion

was

met

;

indi

cate

scr

iterio

nw

asno

tm

et;

?in

dica

tes

insu

ffici

ent

deta

il,no

tre

port

ed,

and/

orun

cert

ain

risk

ofbi

as.

Abb

revi

atio

ns:A

CO

SO

G,A

mer

ican

Col

lege

ofS

urge

ons;

ALM

AN

AC

,Axi

llary

Lym

phat

icM

appi

ngA

gain

stN

odal

Axi

llary

Cle

aran

ce;C

OI,

confl

ict

ofin

tere

st;G

IVO

M,G

rupp

oIn

terd

isci

plin

are

Ven

eto

diO

ncol

ogia

Mam

mar

ia;I

BC

SG

,Int

erna

tiona

lBre

astC

ance

rStu

dyG

roup

;NS

AB

P,N

atio

nalS

urgi

calA

djuv

antB

reas

tand

Bow

elP

roje

ct;R

AC

S,R

oyal

Aus

tral

asia

nC

olle

geof

Sur

geon

s;R

CT,

rand

omiz

edco

ntro

lled

tria

l;S

NA

C,

Sen

tinel

Nod

eV

ersu

sA

xilla

ryC

lear

ance

.S

ever

alite

ms

wer

eno

tre

port

ed;

tria

lnot

fully

publ

ishe

d.I

nter

iman

alys

is.

Int

erim

publ

icat

ion

(n

150)

.

Lyman et al




Survival/mortality. Of the five trials reporting OS and/or mor-tality, none of the studies reported statistically significant differ-ences.3,6,5,9,16 For example, the largest study13 reported an actuarial8-year survival rate for SNB and ALND of 91.8% (95% CI, 90.4 to93.3), compared with 90.3% (95% CI, 88.8 to 91.8) for SNB alone.All-cause mortality was 4% in each arm.16

DFS/EFS. Of the four trials reporting DFS and/or EFS, nonereported statistically significant differences.3,6,9,16

Recurrence. Five trials reported on recurrence.3,5,6,9,16 Theyfound that the rates of in-breast local recurrence, axillary recur-rence, and distant recurrencewere similarwith SNB alone andwithSNB plus ALND. For example, Canavese et al3 showed the overallannual rate of events per 1,000 (including deaths) was 16.2 withSNB and 18.6with ALND. InNSABPB32, with locoregional recur-rence as the first event, local recurrence was 2% in both arms, anddistant recurrence was similar.

Adverse events. Five trials reported adverse events.5-8,11,18,20 Inmost studies, results of adverse events/effects were higher for thosewith tumor-free SLNs who underwent ALND than for those whounderwent only SNB. Important adverse effects includedlymphedema, infections, seroma, and neurologic and sensory def-icits, including paresthesia and shoulder pain and/or impairmentof motion.

In the four studies reporting lymphedema, a clinically importantadverse event, lymphedema was clearly identified to occur even withSNB alone but at lower rates than those found among patients under-going ALND. In the ALMANAC trial, lymphedemawasmeasured byparticipant self-assessment and reported as moderate or severe at 12months in 1% with SNB versus 2% with ALND (P .001).5 Inanother study, lymphedema was statistically significantly lower with-out ALND.6 The NSABP B32 study assessed lymphedema by mea-surement of arm volume a 10% difference from baseline. Grade 3and 4 rates of lymphedema were not notably different. Overalllymphedema was higher in ALND arm.24

Another set of adverse events reported in several studieswere neurologic/sensory deficits. In the four studies reportingthese,5,6,11,13,20,21 the percentages of patients experiencing thesedeficits were statistically significantly lower in the SNB-alonearms.5,6,11,18,20 For example, residual armtingling andnumbness at 36months were significantly lower in favor of SNB. Rates of sensory andmotor neuropathy in NSABP B32 were higher with ALND. In someserious adverse-event categories, therewere no significant differences.For example, in one study, seroma was stratified by nodal status, andthe proportion of patients who developed a seroma with SNB wassignificantly smaller than the proportion in the SNB plus ALNDcontrol group; the difference in those requiring aspirations was alsosignificant.11 InNSABPB32, therewas not a difference in the percent-age of patients who had grade 3 surgery-related events.

Other OutcomesPerformance. Six of the trials to date have reported on perfor-

mance aspects of SNB.3,5-9,16 Performance outcomes extracted in thesystematic review include FNR, negative predictive value (NPV), andoverall accuracy. FNRs reported in the six studies ranged from ap-proximately 4.6%9 (ASCO staff calculation) to 16.7%.6 Three studiesreported NPVs ranging from 90.1%3 to 96.1%.6,16 Four studies re-ported overall accuracy of SNB results ranging from 93%3

to 97.6%.6,16,19

QOL. Three studies reported on QOL.5,11,18,20 The trials usedsuch instruments as the Trial Outcome Index (TOI), Functional As-sessment of Cancer TherapyBreast, scale version 4 (FACT-B4),Psychological General Well Being Index (PGWB), Short Form36(SF-36), and the simple visual analog QOL, as well as tools for mea-suring psychological morbidity. Either there were no significant dif-ferences, or results favored the SNB-alone arm. For example, in theALMANAC trial, there were statistically significantly better outcomesfor participants in the SNB-alone arm; also reported were changes inFACT-B4 scores favoring SNB alone.18

Clinical InterpretationThe Z0010 study25 was not included in final systematic review,

because it was a prospective multicenter cohort, not an RCT. Itsprimary end point was OS, and secondary end points included DFSand axillary recurrence; however, theywere not reported in the Z0010publication used in this guideline.25 The study alsomeasured adverseevents at 30 days and at 6-month intervals up to year 3 and annuallythereafter.A total of 5,539patientswere enrolled, and therewere5,327available for analysis. In amultivariable analysis, age (at a cutoff of 70years) was a predictor for axillary seroma. In another outcome of amultivariable analysis, decreasingagewas associatedwith thepresenceof paresthesia.

CLINICAL QUESTION 2

Is ALND necessary for all patients with metastatic findings on SNB?

Literature Review and AnalysisThis section summarizes the efficacy results from the trials pub-

lishedsince the systematic reviewfor the2005guideline for this clinicalquestion. It is based on two RCTs that both reported efficacy andadverse-event results in patients with metastases in SLNs randomlyassigned to either completionALNDor noALND:ACOSOG (Amer-ican College of Surgeons Oncology Group) Z00114 and IBCSG 23-01.12 The ACOSOGZ0011 investigators reported on recurrence, OS/mortality, DFS, and adverse events in four articles included in thissystematic review. The IBCSG 23-01 investigators reported on recur-rence, OS/mortality, DFS, and adverse events in one article publishedafter theASCOsystematic review, but theUpdateCommittee deemedit appropriate to include because of its confirmatory nature.

ACOSOGZ0011was anoninferiorityRCT inwhich 446patientswere randomlyassigned toSNBwithno further axillary treatment and445 to SNB plus ALND. All patients hadmetastatic SLNs, and all hadT1 or T2 tumorsmanaged by lumpectomy and plannedwhole-breastirradiation. The extent of node involvement was micrometastatic(metastatic focus2mm) inapproximatelyhalf of the studypatients,but this informationwas not available for all trial participants.OSwasthe primary outcome. The primary outcome of IBCSG 23-01 wasnoninferiority in DFS, and this study was designed for patients withsentinel node micrometastases. It had a sample size of 964 partici-pants; most participants had only one metastatic node. Consistentwith the study design, 98% of the SLN tumors in both randomizedarms had micrometastatic disease only. Results in IBCSG 23-01 werenot analyzed by size ofmetastases.More than 80%of the participantsin theACOSOGZ0011andIBCSG23-01 trialshadestrogenreceptorpositive disease. Both of these studies were closed early because offailure to meet their accrual targets.





The recommendations in response to this clinical question weresplit into two separate recommendations. The first recommendationwas crafted to reflect the eligibility criteria of Z0011 (women withearly-stage breast cancer and one to two SLNmetastases, who under-went BCS with whole-breast radiotherapy).

CLINICAL QUESTION 2.1

Is ALND necessary for all patients with metastatic findings on SNBplanning to undergo BCS with whole-breast radiotherapy?

RECOMMENDATION 2.1Clinicians should not recommendALND forwomenwith early-

stage breast cancer and one or two SLNmetastases who will undergoBCS with conventionally fractionated whole-breast radiotherapy.Type: evidence based; benefits outweigh harms. Evidence quality:strong. Strength of recommendation: high.

Literature Review and Analysis: Clinical OutcomesMortality. There was no apparent negative impact onmortality

of omitting ALND (there was a statistically significant difference fornoninferiority inmortality) after amedian follow-upof 6.3 years. Thiswas thefirstRCT to show these results. Therewereno suchdifferencesin IBCSG 23-01.

DFS. In Z0011, DFS was not statistically significant betweenarms. In IBCSG 23-01, there was a 3.4% reduction in DFS for SNBalone, which was statistically significant for noninferiority forALND, with a 5-year follow-up. The per-protocol assessment ofDFSwas also statistically significant for noninferiority (P .0073).However, this study randomly assigned only 934 of an accrualtarget of 1,960 participants.12

Recurrence. Both studies reported on recurrence. In locore-gional, axillary, or distant recurrence, outcome results showed nosignificant differences.4 The differences were not statistically signifi-cant between arms in IBCSG 23-01 (eg, local recurrence rate at firstevent was 2% in both arms; distant recurrencewas 7% for ALNDand5% for no ALND).12

Adverse events. There were statistically significant differences inrates of adverse events between study arms in both studies. Usingslightly different numbers in the denominator than in the primaryreport extracted, another ACOSOG Z0011 publication reported onsurgical complications. The protocol did not include objective mea-surements of arm volume. However, subjectively reportedlymphedemawas higher with ALND than in those with SNB alone. Itwas statistically significantly higher starting at 12 months, althoughnot at 6 months or by arm measurement.26 There was statisticallysignificantly higher axillary seroma and parethesia in the immediateALND arm (P .001), but not impaired range of movement (thelatter two at 1 year), in a joint analysis of the cohort study Z0010 andRCT Z0011.27 In a third-line report from Z0011, adverse surgicaleffects were 45% less in the SNB arm, including lymphedema 12months.4,26,27 Infection inZ011was reported as lowerwith SNBalone(P .0026).26

In IBCSG 23-01, there were lower overall rates of lymphedema(13% v 3%; P .001) with no ALND. The percentages of patientsexperiencing neurologic/sensory deficits were also reported as higherwithALND(eg,motor neuropathy); however, the rates of grade 3 to 4

deficitswere low in IBCSG23-01.12 It is important tonote that adverseeventswerenot stratifiedbymicro- versusmacrometastases in IBCSG23-01. Infection was higher in the ALND arm (P .0016) in Z0011;there was one case of infection in the ALND arm of IBCSG 23-01.There was not a subgroup of analyses of those whose SLN tumorswere 2mm in IBCSG 23-01 (2% in each arm).

Other FindingsPerformance and QOL. There are not yet reports of findings on

performance or QOL to inform this recommendation.

Clinical InterpretationTheeligibility criteria forZ0011 includedafindingof ametastatic

SLNby frozen section, touchpreparation, or hematoxylin-eosin (HE)staining; all participants had lumpectomies. Exclusion criteria in-cluded having three metastatic SLNs. Most patients entered thestudy after undergoing SNB, but 287 of 891 had not yet undergoneSNB and were randomly assigned after surgeons found metastaticSLNs intraoperatively.4Thegroupwith threemetastatic SLNscom-prised 21% of the patients in the ALND arm and 3.7% in the SNB-alone arm; there was a higher proportion of patients with zero to twometastatic nodes in the SNB-alone arm. Although this was set an asineligibility criterion, some of the patients were enrolled before nodalstatus was known; if it turned out they had three SLNs, thesepatients were included in the analysis.

By contrast, in IBCSG 23-01, 1% of those in the ALND armandnoparticipant in theno-ALNDarmhad threemetastatic SLNs.According to the original eligibility criteria, participants could haveonly onemicrometastatic SLN. The criteria for eligibility were broad-ened after the study began to include patients with one metastaticSLN, with multicentric or multifocal tumors (formerly only unicen-tric), andwhose largest lesion sizewas5 cm(formerly3 cm). Fivepercent of participants undergoing ALND and 4%of those randomlyassigned to no ALND had two metastatic SLNs. Most of the partici-pants had one metastatic node; however, 98% of the SLN tumors inboth arms were 2mm (micrometastatic).

In the expert opinion of the Update Committee, ALND can beavoided in cases of BCS, but only when conventionally fractionatedwhole-breast radiation therapy is planned. Although there were pa-tients in the IBCSG study in the group managed without axillarydissectionwhounderwentBCSwithnoradiation therapy(12patients;3%) or intraoperative partial-breast irradiation (80 patients; 19%),these subgroups were deemed too small to influence the decision bythe Update Committee not to extend the recommendation beyondpatients treated with conventionally fractionated whole-breast irradi-ation. This recommendationdoes not apply to patientswhose axillarynodal positivity is documented by axillary fine-needle aspiration(FNA); clinicians may still perform SNB if the abnormal lymph nodeis removed. Data are insufficient to address whether FNA-positivepatients can undergo SNB (under the assumption that resected SLNsrepresent the FNA-biopsied node) and then avoid ALND after resec-tion of the metastatic SLN. Patients with FNA-positive nodes mayhave ahigher axillary tumorburdencomparedwith thosewithdiseaseapparent on dissection of an SLN.

Clinicians may also consider this recommendation with cautionin cases of womenwith large or bulkymetastatic axillary SLNs and/orthosewith gross extranodal extension of the tumor. Thesewere exclu-sion criteria for Z0011. The former were represented by few patient

Lyman et al




cases in the ACOSOG Z0011 study, and the latter were specificallyexcluded. These uncommon patient cases may or may not carry ahigher risk of axillary recurrence than those largely represented inthe trial.

CLINICAL QUESTION 2.2

Is ALND necessary for all patients with metastatic findings onSNB who are planning to undergo mastectomy?

RECOMMENDATION 2.2Clinicians may offer ALND for women with early-stage breast

cancer with nodal metastases found on SNB who will undergo mas-tectomy. Type: evidence based; benefits outweigh harms. Evidencequality: low. Strength of recommendation: weak.

Literature Review and AnalysisThis recommendation is based on a subgroup of participants in

IBCSG23-01.Ninepercent of theparticipants in each armunderwentmastectomy (ALND,n44; noALND,n42). In the study analysesby subgroup, the results by type of surgery showed a lower risk ofeventswithnoALNDandBCS ( radiation therapy).Adverse events,QOL, and performance were not reported.

Clinical InterpretationThis recommendation is based on one subgroup of one study,

involving a small number of participants. Adverse events were greaterwith ALND. There are currently conflicting data. There may be re-search in the future that could further inform this recommendation.

CLINICAL QUESTION 3

What is the role of SNB in special circumstances in clinical practice?Note: Because of lack of evidence inmany areas, the scope of this

section was narrowed to the following circumstances. Recommenda-tions from2005areavailable inDataSupplement7.These recommen-dations are limited by the insufficiency and paucity of data.

RECOMMENDATION 3.1: MULTICENTRICClinicians may offer SNB for women who have operable breast

cancer and the following circumstance: multicentric tumors. Type:evidence based; benefits outweigh harms. Evidence quality: interme-diate. Strength of recommendation: moderate.

Literature Review and AnalysisFor the question of patients with multicentric tumors, the sys-

tematic review found five observational studiesmeeting the inclusioncriteria. One study reported on survival and axillary recurrence; theresults were not statistically significant. One reported on survival orDFS, two on recurrence, and three on performance. No significantdifferenceswere reported.Onewas a substudy of theALMANAC trial(which was included as an RCT informing Clinical Question 1), inwhich all patients underwent SNB. There were no significant differ-ences for either FNR or number of failed localizations for those withmultifocal metastases.19 In another substudy of ALMANAC,28 therewere no significant differences in performance. There is one registry

study examining risk factors for axillary recurrence. The investigatorsdivided patients with tumor-free SNB into four subgroups (smallunifocal, large unifocal, smallmultifocal, and largemultifocal [metas-tases]). There were axillary recurrences in 0.4% of the patients withsmall unifocal, 1.6% of those with small multifocal, and 0% of thosewith large unifocal or large multifocal tumors.29

In another study, overall accuracywas similar betweenmulticen-tric and unicentric groups; the FNR was lower for the multicentricgroup. NPVwas 93.3% for multicentric and 97.9% for unicentric.30

Clinical InterpretationIn the observational data reviewed, no meaningful differences

were observed between studies using the terminology of multicentricormultifocal. The harms of omitting ALNDdonot seem to outweighthe benefits. Therefore, the Update Committee suggests that womenwith multicentric tumors be evaluated for SNB by the criteria inRecommendations 1 and 2.

RECOMMENDATION 3.2: DUCTAL CARCINOMAIN SITU

Clinicians may offer SNB for women who have operable breastcancer and the following circumstance: ductal carcinoma in situ(DCIS), when mastectomy is performed. Type: informal consensus;benefits outweigh harms. Evidence quality: insufficient. Strength ofrecommendation: weak.

Qualifying StatementsClinicians may perform SNB for DCIS diagnosed by minimally

invasive breast biopsy: one,whenmastectomy is planned, because thisprecludes subsequent SNB at a second operation; two, when physicalexamination or imaging shows a mass lesion highly suggestive ofinvasive cancer; or three, when the area of DCIS by imaging is large( 5 cm).

DCIS is characterized by proliferation of ductal epitheliumwith-out penetration or invasion through the ductal basementmembrane.DCISmay be present as a component of an invasive cancer, or it mayexistwithout any invasive cancer. In general, clinicians agree that pureDCIS without a coexisting invasive cancer cannot invade lymphaticsand spread to regional nodes.However, because of the sampling errorof minimally invasive biopsies, a substantial fraction of womenidentified with pure DCIS on a core-needle/vacuum-assisted min-imally invasive biopsy prove to have some component of invasivecancer at surgical resection.31 However, there is currently no vali-datedmethod to predict which patients will have invasive cancer inthis setting. This has been used by some to justify performing SNBin all womenwithDCIS identified by core-needle/vacuum-assistedminimally invasive breast biopsy. Retrospective series show that asmall percent have node metastases identified. However, a largemajority of these are classified as micrometastases or clusters ofisolated tumor cells ( 0.2 mm). Other data as reviewed in ourupdate show that SNB can be performed in women with a priorsurgical excision in the breast with success, and accuracy rates areequal to those in women with no prior breast excision.

Literature Review and AnalysisThere were no studies thatmet criteria for evaluation of SNB for

patients with DCIS, as confirmed by other recent systematic re-views.1,2 The rate of identification of metastatic SLNs defined in the





systematic reviews for patients proving to have pure DCIS on finalresection is 0.9% for pN1 and 1.5% for pN1mic (micrometastases).

Clinical InterpretationForwomenwith aminimally invasive biopsy showingDCISwho

arebeing treatedwithBCS, there isnoevidence to supportperformingSNB (see Recommendation 4.3). Performing SNB places patients atrisk for long-term complications including permanent lymphedema.SNBmay be performed as a separate second procedure in the womenin whom invasive cancer is found (reported in 10% to 20% of casesoverall, approximately half of which are limited to microinvasivecancer). Exceptions may include cases where breast imaging or phys-ical examination show an obviousmass characteristic of invasive can-cer or a large area of calcification without a mass (eg, 5 cm) wherethe probability of finding invasive cancer on the resection specimen ishigh. In addition, whenmastectomy is performed forDCIS, itmay bewarranted to perform SNB because of the possibility of finding aninvasive cancer in the resected breast, and the disruption of the lym-phatics by the mastectomy may preclude a subsequent SNB. Theserecommendations are primarily based on retrospective data. There-fore, the Update Committee does not endorse routine SNB for pa-tients with DCIS undergoing BCS.

RECOMMENDATION 3.3: PRIOR SURGERYClinicians may offer SNB for women who have operable breast

cancer and the following circumstance: prior breast and/or axillarysurgery. Type: evidence based; benefits outweigh harms. Evidencequality: intermediate. Strength of recommendation: strong.

Literature Review and AnalysisThe systematic review found two observational studies meeting

the inclusion criteria. These studies did not report survival or DFS.One study reported recurrence and found a slightly higher recurrencerate for patients who had undergone prior biopsy. Both studies re-ported performance and did not find differences in those outcomes.

The first was a nonrandomized study that had two temporalphases, each including two groups (those with nonpalpable lesionsand those with prior diagnostic biopsy). Distant recurrence for thosewith metastatic SLNs was as follows: prior biopsy, 5%; nonpalpablelesion, 1%. For patients with tumor-free SLNs, distant recurrencewasas follows: prior biopsy, 2.0%; nonpalpable lesion, 0.5%. Axillaryrecurrence for those who had undergone prior surgery was zero inboth thosewithmetastatic SLNs and thosewith tumor-free SLNs.TheSLNdetection rate for those having undergone prior surgerywas 96%versus 95%, and the FNR was 10% versus 5.6%.32 The second studywas a nonrandomized study that included two groups: one grouphadundergone prior excisional biopsies, and the comparison group hadundergone diagnostic core biopsies. Overall accuracy and sensitivitywas 100% in the first group, and there were no false negatives.33

Clinical InterpretationAlthough there are no randomized data or other additional evi-

dence that meet the eligibility criteria for this guideline update, theretrospective data are consistent with the feasibility and acceptableaccuracy of performing SLN biopsy in patients who have undergoneprior nononcologic breast/axillary surgery.

RECOMMENDATION 3.4: PREOPERATIVE/NEOADJUVANT SYSTEMIC THERAPY

Clinicians may offer SNB for women who have operablebreast cancer and the following circumstance: preoperative/neoadjuvant systemic therapy (NACT). Type: evidence based;benefits outweigh harms. Evidence quality: intermediate.Strength of recommendation: moderate.

Qualifying StatementsSNB may be offered before or after NACT, but the procedure

seems less accurate after NACT.

Literature Review and AnalysisThere were three cohort studies in the ASCO systematic review

on preoperative systemic therapy and/or NACT. None reported sur-vival/mortality, and one reported recurrence; the other data were onperformance.Most of the studies did not show statistically significantdifferences in the reported outcomes between those who receivedNACT and those who did not,34,35 including FNR in the two studiesthat reported it.34-36

The analysis of the first study excluded patients who had SNBonly (although included them in the study). Overall accuracy forthose who had received preoperative chemotherapy was 95.9%versus 92.6% for those who had not received preoperative chemo-therapy. Among patients whose SLNs were successfully identified,the overall accuracy rate was 93.7%. The FNRwas not significantlydifferent. For those who received preoperative chemotherapy, theNPV was 86.8%.34

There was a second smaller study of patients with locally ad-vanced breast cancer. The comparison group (patients with early-stage breast cancer) was from a previous RCT. The FNR for SNB afterNACT was 5.2%. The third study was small and retrospective/pro-spective. The sole recurrence information it reported was distant re-currence in one of 52 patients who had received neoadjuvantchemotherapy.Themediandetectionratewasnonsignificantbetweenthose who had not received neoadjuvant chemotherapy and thosewho had received it. The only statistically significant result was thenumber of lymphnodes removed, whichwas statistically significantlygreater in the non-NACT group.37

Clinical InterpretationSNBmay be offered before or afterNACT, but the FNR is higher

afterward, and therefore, the procedure seems less accurate afterNACT.Theoutcome forpatientswhohavemetastaticnodes that thenbecome negative has never been investigated. There were some otherstudies that were not included in the systematic review and, therefore,donot support the recommendationbutare selectivelydiscussedhere.

One study did not fit the ASCO systematic review criteria, be-cause it didnot compareparticipantswithandwithoutNACT.Aspartof theNSABPB27 trial ofNACTinparticipantswithclinical stage IIorIII breast cancer, axillary dissection was required for all patients. Thedates of this study coincided with ongoing studies from NSABP andACOSOG of SNB in patients with clinical node tumor-free breastcancer.The techniquewasnot standardized, and itwasnot included inthe actual B27 study. Although SNBwas not a component of the B27study, the NSABP determined that 428 participants had SNB beforethe required ALND at the treating center. The NSABP elected toreport the collective findings.However, the findings are limited by the

Lyman et al




fact that the techniqueof SNBwas selectedby the treating surgeonandwas not consistent (radioactive colloid alone, 15%; lymphazurinalone, 30%; both, 55%). At least one SLN was identified in 85% ofpatient cases. Among343patientswith both SNBandALND, 125hadmetastatic nodes. Among those with tumor-free SLNs, 15 had meta-static nonsentinel nodes (FNR, 10.7%).38

The benefits outweigh the harms in a given patient, because thepatient has already received systemic therapy, and the impact of a falsenegative is unlikely to lead to omission of radiation therapy. Cellsremaining in the SLNs after chemotherapy may be chemoresistant,and postoperative chemotherapy is variable and often not pursued.Radiation to the axillamay reduce the risk but is variably applied. Forpatients withmetastatic nodes beforeNACT, the FNRwith SNB aftertreatment may range from 10% to 30%, which, in the view of theUpdate Committee, is unacceptable. This may result in understagingand undertreatment of such patients. The SLN removed may not bethe same node previously biopsied and found to be metastatic. TheSENTINA (Sentinel Neoadjuvant) trial observed an FNR of 14.2%(and much higher rates if only one or two SLNs were removed).39

More than half of patients who became clinically tumor free afterNACT still had metastatic nodes pathologically. For example, theACOSOG Z1071 trial found FNRs 12% when only two SLNswere removed.40 There were FNRs of 31% if only one SLN was re-moved and 20% for single-agent lymphatic mapping (isotope or dyealone). Because of the publication dates of the reports, these studieswere not included in the systematic review.

For patients who have received NACT and are considering un-dergoing SNB (ALND), themost importantmeasure of outcome isaxillary local recurrence. TheUpdate Committee urges caution aboutSNB after NACT because of the fact that no studies to date havereported a positive result in axillary local recurrence.

In the expert opinion of the Update Committee, SNB is notrecommended in patients with T4d/inflammatory breast cancer whohave received NACT (regardless of patients clinical response toNACT), and data are insufficient to recommend SNB in patients withT4abcbreast cancerwhose cancer has been clinically downstaged afterreceiving NACT.

In addition, SLNbiopsy afterNACT is associatedwith a learningcurve and surgical expertise.41 Lastly, decisions regarding the use oflocoregional radiation therapy after NACTmay be influenced by thehistopathologic pre- as well as post-NACT nodal status. In the expertopinion of theUpdate Committee, data are presently insufficient, butthe committee will take under consideration any data forthcomingregarding SNB in patients who have node metastases at presentationby pretreatment axillary FNA biopsy and are planning to re-ceive NACT.

OTHER SPECIAL CIRCUMSTANCES

RECOMMENDATION 4.1: LARGE AND LOCALLYADVANCED INVASIVE TUMORS (T3/T4)

There are insufficient data to change the 2005 recommendationthat clinicians should not perform SNB for women who have early-stage breast cancer and have the following circumstance: large orlocally advanced invasive breast cancer (tumor size T3/T4). Type:informal consensus. Evidencequality: insufficient. Strengthof recom-mendation: weak.

Literature Review and AnalysisThere was one study found. The single small study meeting the

inclusion criteria included 64 patients with locally advanced breastcancer. The comparison group (patients with early-stage breast can-cer)was fromapreviousRCT.Theoverall accuracyresultswere96.7%for patients with locally advanced breast cancer and 93.0% for thosewith early-stage breast cancer. Other results included FNRs (locallyadvanced, 5.1%; early stage, 5.8%), NPV (locally advanced, 91.3%;early stage, 91.1%), and sensitivity (locally advanced, 88.1%; earlystage, 77.1%

Documents

J Clin Oncol (2014) Sentinel Lymph Node Biopsy for Patients With Early-Stage Breast Cancer- American Society of Clinical Oncology Clinical Practice Guideline Update