Breast imaging reporting and data system standardized mammography lexicon: observer variability in lesion description.

Abstract

The American College of Radiology has recommended the Breast Imaging Reporting and Data System (BI-RADS) as a standardized scheme for describing mammographic lesions. The objective of this study was to measure inter- and intraobserver variabilities of radiologists' descriptions of mammographic lesions with the BI-RADS standardized lexicon.
Sixty mammographic studies with abnormal findings were independently evaluated by five radiologists. Readers described each lesion by selecting a single term from the BI-RADS lexicon for each of eight morphologic categories: calcification distribution, number, and description; mass margin, shape, and density; associated findings; and special cases. Additionally, each reader assessed the significance of each lesion on a five-point scale. One observer read each case twice. Inter- and intraobserver variabilities for each description and interpretation category of the BI-RADS lexicon were determined with Cohen's kappa statistic. Radiologists' specific use of calcification descriptors was evaluated in detail.
Substantial agreement was observed between readers for choosing terms to describe masses and calcifications (kappa value range, 0.50 +/- 0.02-0.77 +/- 0.03). Intraobserver agreement for these categories was similar (kappa value range, 0.57 +/- 0.07-0.84 +/- 0.09). Considerable inter- and intraobserver variabilities were noted for the "associated findings" and "special cases" categories (kappa value range, -0.02 +/- 0.14-0.38 +/- 0.12), a result that in part reflected the small number of cases to which these categories were assigned. Moderate interobserver variability and little intraobserver variability in the interpretation of lesion significance were noted when an assessment classification similar to that of BI-RADS was used. Use of terms to describe calcifications did not always conform to BI-RADS-defined levels of suspicion.
BI-RADS is moderately successful in providing a standardized language for physicians to describe lesion morphology. Efforts to reevaluate specific terms and the diagnostic significance assigned to calcification descriptors may prove useful in maintaining the promise of improved quality with the BI-RADS standardized mammography lexicon.

Full text

Page 1

773
Breast
Data
Mammography
Variability
Imaging
System
Reporting
Standardized
Lexicon:
in Lesion
and
Observer
Description
Jay
J. Kornguth1
E. Floyd,
A. Baker1
Phyllis
CareyJr.1’2
Received
September
This work was supported
17-94-J-4371
1Departmentof
July 13, 1995;
29, 1995.
acceptedafter revision
in part by grant DAMD
by the United
Radiology, Box
awarded StatesArmy.
Uni-3808, Duke
versityMedical Center,Durham,
to J. A. Baker.
of Biomedicai
Durham, NC 27710.
0361-803X196/1664-773
© American Roentgen
NC 27710.Ad-
dress correspondence
2Department
University,
Engineering,Duke
Ray Society
OBJECTIVE.
Imaging
ing mammographic
intraobserver
withthe
MATERIALS
wereindependently
selecting
gories:calcification
density;
thesignificance
twIce.
inter-
egory
of the
gists’specific
RESULTS.
todescribe
lntraobserver
0.07-0.84
“associated
0.38± 0.12), a result
egorieswere
ability
in the
classification
tionsdid not always
CONCLUSION.
guagefor
termsand
usefulin maintaining
mammography
TheAmerIcan
and
lesIons.
CollegeofRadiologyhasrecommended the
for describ-
Inter-
lesions
Breast
ReportingDataSystem
The
of
(BI-RADS)
objective
radIologIsts’
lexicon.
Sixty
by five
theBl-RADS
number,
andspecial
lesionon
lntraobserver
BI-RADSlexicon
use of calcification
Substantialagreement
massesandcaicifications
agreementfor
± 0.09).
Considerable
findings”and“special
thatin part
assigned.Moderate
interpretation
similarto that
conform
BI-RADS
physicianstodescribe
the diagnosticsignificance
thepromise
lexicon.
as a standardized
of thisstudy
descriptions
scheme
measurewas
of
toand
variabilities
standardized
ANDMETHODS.
evaluated
a singleterm
mammographic
BI-RADS
mammographic
radiologists.
lexicon
and
cases.
a five-point
variabilities
weredetermined
descriptors
was observed
studieswithabnormalfindings
lesionReaders
for each
description;
Additionally,
scale.
for each
with
wasevaluated
between
(kappavalue
categories
was
andintraobserver
cases”categories
reflected
thesmall
interobserver
lesionsignificance
of BI-RADSwas
to Bi-RADS-defined
is moderatelysuccessful
lesionmorphology.
assigned
of improved
described
of eight
mass
each
observer
eachby
frommorphologic
margin,
reader
read
interpretation
statistic.
in detail.
readersfor choosing
0.50
(kappavalue
variabilities
(kappavalue
numberof cases
variabilityandlittle
werenoted
used.
Use
of terms
levelsof suspicion.
in providing
Efforts
to calcification
qualitywiththe
cate-
anddistribution,
findings;
of each
shape,
assessed
each
associated
Onecase
cat-anddescription
Cohen’s
and
kappaRadiolo-
terms
± 0.03).
0.57
for the
± 0.14-
these
range,
± 0.02-0.77
these
similarrange,
noted
-0.02
±
inter-were
range,
to which
intraobserver
whenan
to describe
cat-
van-
ofassessment
calcif lea-
a standardized
reevaluate
descriptors
BI-RADS
Ian-
tospecific
may
standardized
prove
AJR
1996;166:773-778
Aspartof ongoing
College
Data
of breast
lexicon
to provide
hasrecommended
mammograms.
reports[1-3].
DespiteACR
datehaveaddressed
whenthe BI-RADS
siderablevariability
efforts
of Radiology
(BI-RADS)
lesions
descriptive
a common
to improvequality
has devised
a standardized
categorizing
termsis
language
useof this
lexiconhas
in thepractice
the Breast
method
findings
part
description
of mammography,
Imaging
forevaluating
in anunambiguous
ofBI-RADS
andassessment.
by all physicians
in severalpreviously
the American
ingand
morphology
report.
intended
TheACR
preting
lished
(ACR)
[1],
Report-
Systemthe
andthe
Aofanintegral
lesion
nomenclature
reviewed
andis
for
theinter-
pub-Thisbeen
recommendations
the uniformity
lexicon
among
for
of descriptions
Although
in diagnostic
widespreadimplementation,
of mammognaphic
reports
interpretations
no
abnormalities
documented
of and
studiesto
is used.
radiologists
severalhavecon-
manage-

Page 2

774BAKERET AL.
AJR:166,April 1996
ment
evaluated
lesion
measure
descriptions
standardized
recommendations
variability
morphology.
.
inter-
for
radiologists’
study
intnaobserver
mammographic
lexicon.
breastlesions
descriptions
presented
variabilities
lesions
[4-6],
none
of
have
breast
TABLE
Lexicon
I : Terms Includedin the ACRBI-RADS Standardized
in
The hereendeavors
in radiologists
theBI-RADS
to
Category Term
and
of with Calcifications
Distribution Grouped
Linear
Segmental
Regional
Diffuse
<5
or clustered
Materialsand Methods
.
in this
diagnosis
is briefly
.
or scattered
Thedata analyzed study
in mammography
described
werecollected for a related
[7]. The
study Number
on computer-aided
acquisition
methodof
of the databelow.
?
DescriptionSkin
Case Selection
Vascular
Coarse
Large
Round
Spherical
Rim
Milk
Suture
Dystrophic
Punctate
indistinct
Pleomorphic
Fine
or “popcorn
rod like
like”
Sixtymammographic
patients
palpable,
selectedfrom
whomjoined
of thecases.
old).Twenty-three
of thesecancers
studieswererandomly
localization
breast
theradiologist
1990,
years
weremalignant
(35%).
selected
in 1990
lesion.
observers,
not familiar
(average,
(38%),
for
for a non-
Cases
study
fromwhounderwentneedle
detected
that
after
36-79
mammographically
1990
Duke
Patients
were
all of
any
years
eight
to ensure
University
were
of the 60 cases
werenoninvasive
or eggshell
of calciumwere
old
with
56.6
and
or amorphous
ParticipatingRadiologists
branching
Fiveboard-certified
participant
and three
in mammography.
phersand
Thefinal
ship training
togetherat the same
responsible
RADS
lexicon
mammography
ar withthe terminology.
radiologists
hadextensive
participants
Of these,
has practiced
participant
atthetime
were
practical
have
are dedicated
predominantly
in the fifth
ofthestudy.
institution.The
for trainingthe other
reviewed
fellow.Theremaining
invitedto readall 60 cases.
in mammog-
fellowship
academicmammogra-
as a general
of mammography
ofthesefive observers
experienced
three.Definitions
generalradiologist
participantswere
One
raphy,
has
other
experienceMasses
Margin
.
completedtrainingCircumscribed
Microlobulated
Obscured
indistinct
Spiculated
Round
Oval
Lobular
Irregular
Architectural
used
Fat containing
Low density
Equal
Highdensity
two
oneradiologist.?
wasmonth
Four
most
fellow-
work
was
or ill defined
observerShape
in partof the
and
BI-
thewerewiththe
alreadyfamil-
distortion(may
finding)
alsobe
as associated
Density
MammogramEvaluation
density(isodensity)
Eachreader
were
described
independently
clearly
the
andmediolateral-oblique
specialviews
describedeach
lexiconfor eight
number,and
associatedfindings
(e.g.,asymmetric
lexiconarelisted
Eachreaderestimated
on a five-pointscale
likelymalignant;
University.Four
(benign;probably
gestiveof malignancy).
nant”categories
classificationlabeled
equivalenttothe
RADSlexiconhas
caseswithoutfindings,
evaluated
on the
lesions.
each of the 60 cases.
to ensure
Observerswere
viewsof each
at the timeof the examination.
by selectinga single
morphologiccategories:
description;massmargin,
(e.g.,axillaryadenopathy);
breasttissue).All terms
in Table1 .
the likelihoodof malignancy
(1 , benign;
2, likely benign;
5,malignant),whichis routinely
Bi-RADSassessmentcategories
benign;suspiciousabnormality;
Whenthe“intermediate”
ofthefive-pointscale
“suspicious,”theresulting
fourBI-RADSassessment
oneadditionalassessment
whichwasnot included
The
lesions
pants
craniocaudal
as any
Readers
Bi-RADS
tribution,
sity;
cases
RADS
identified
same
filmsthat
provided
breast
all partici-
withOtherfindings
as wellAssociatedSkin
Nipple
Skin
Trabecular
Skin
Axillary
Architectural
Tubular
Intramammary
Asymmetric
Focal
retraction
retraction
thickening
otherobtained
lesiontermfromthe
dis-
den-
calcification
shape,
thickening
andlesion
andspecial
in the
adenopathy
includedBI-distortion
or solitary
lymph
breast
Specialcasesdensitydilatedduct
of eachlesionnode
3, intermediate;
used
arecomparable
andhighly
and“likely
combined
fourcategories
categories.
category
in this study.
4,tissue
densityandatDukeasymmetric
sug-Note.-ACR
Reporting
= AmericanCollegeof Radiology,Bi-RADS
= Breast imaging
malig-and Data System.
areintoone
are
BI-The
normal
To assess
rienced
least3 months
intraobserver
re-evaluated
after
variability
all 60 cases.
initialexercise.
withBI-RADS,
The
the
were
mostexpe-
forreaderfilmsrereadat
the

Page 3

AJR:166,April 1996
STANDARDIZEDMAMMOGRAPHYLEXICON
775
for choosingdescriptionsof mammognaphic
lesionmonphol-
Note.-Bl-RADS
= Breast ImagingReporting and Data System.
TABLE
of Calcification
2: Groupingof Specific BI-RADS Termsfor Description
Morphology
Category
Term
Typicallybenign Milk
Rim
Skin
Vascular
Spherical
Suture
Coarse
Large
Round
Dystrophic
Punctate
of calcium
rod like
intermediateIndistinct
Highprobabilityof malignancy Pleomorphic
Finebranching
Note.-Bl-RADS
= Breast
Imaging ReportingandData System.
StatisticalAnalysis
inter-andintraobserver
morphologic
Cohen’s
interpretations
statistical
or more
determines
agree and
solely
of 1.0, whereas
expected
agreement
scaleexists,
of agreement
?0.20,slight
variabilities
category
statistic
was
for choosing
BI-RADS
Variability
descriptive
lexicon
in observ-
Cohen’s
assess
ornominal
of selections
possibilityof
agreementis
valueof 0 indicates
A negativekappa
expectedby chance.
reportshavesuggested
observersforthe
0.21-0.40,fairagreement;
terms
evaluated
ers’
statistic
between
This
observers
attributable
kappa
agreement
cates
absolute
levels
ues:
from each
with
of the
[8-10].
calculated.
designed
forcategorical
proportion
forthe
Perfect
a kappa
chancealone.
thanthat
prior
between
agreement;
were
kappa
diagnosticalsokappa
is
two
a measure
observations
to agreement
data.
whichtechniquethefor
accounts
chance.
agreements
indicated
the level
value
Although
the
kappa
0.41-0.60,
tobya
valueof
bymdi-
lessno
following
indicatedval-
moderate
1 .00,
authors
to indicate
considered
agreement;0.61-0.80,
agreement
that
substantial
between
agreement;
observers
of ?0.50
a kappa
and
[11].
0.81-
Otheralmost
have
perfect
advocated
poor
to indicate
a kappa
and
reproducibility
valuebe considered
valueof ?0.75
[12].
reproducibilitythatbe
excellent
Evaluationof CalcificationDescriptors
Ofthe300 total lesion
evaluated
described
not alwaysagree
berof cases
evenlydivisible
groupsthe
broadcategories:
malignancy
the frequency
wascompared
the histologic
descriptions
microcalcifications
in 151 cases.
on the presence
describedto have
by thenumber
14 termsfor describing
typically
(Table2). To assess
of Bi-RADSterms
withthe diagnostic
diagnosesobtained
in this study(i.e., five participants
werethe only significant
the radiologist
of additionalfindings,
onlymicrocalcifications,
of observers.)The
calcificationmorphology
intermediate;and high probability
theusefulnessof these
chosento describe
interpretationsof the observers
at biopsy.
each
finding
60 lesions),
(Becauseobservers
the total
151 , is not
BI-RADS
into three
did
num-
lexicon
benign;of
broadgroups,
microcalcifications
and
Results
ObserverVariability
Thestatisticalanalysisof agreementbetweenobservers
ogy
agreement
masses
for selecting
(kappa
atively
In marked
tions,
mining
± 0.04)
even,
foronly
portion
gory,
to the
larylymphadenopathy.
cases”category
an actual
Results
of mammographic
therewas
is shownin Table
terms
calcifications.
oneof 14 terms
0.50± 0.02)
number
contrast
waspoor
the presence
andspecial
becauseterms
sixcases,
of the60 studies.
the onlydisagreement
presenceor absence
3. In general,
thatobservers
The
to describe
was
of choices.
to descriptions
agreement
of associated
cases(kappa
in these
thisdisagreement
For the
between
of architectural
Similarly,
concerned
on focalasymmetric
of intraobserven
abnormalities
moderateon substantial
therewassubstantial
todescribe
agreement
calcifications
of the
forchose
andlessimpressive
breast
becausevalue,
large
expectedrel-
of massesandcalcifica-
for deter-
value,
± 0.03).
eachselected
only
findings”
pertained
distortion
in the
opacityrepresented
density.
in choosing
shownin Table
agreement
therebetween
findings
value,
categories
observers
(kappa
0.16
were
involved
0.32
How-
a small
cate-“associated
any observers
and
“special
axil-
disagreement
whetheran
mass
variabilitydescriptions
4. Again,
between
are
the
TABLE
Mammographic
3:InterobserverVariabilityin ChoosingDescriptionsof
Abnormalities
BI-RADSCategoryKappaValue
Calcifications
Distribution
Number
Description
0.77
0.77
0.50
± 0.03
± 0.03
± 0.02
Masses
Margin
Shape
Density
0.63
0.65
0.62
± 0.02
±0.03
± 0.03
Otherfindings
Associated
Special
0.32
0.16
± 0.04
± 0.03cases
Note.-BI-RADS
= BreastImagingReportingand Data System.
TABLE
Mammographic
4:lntraobserverVariabilityin ChoosingDescriptionsof
Abnormalities
BI-RADSCategoryKappaValue
Calcifications
Distribution
Number
Description
0.80
0.84
0.57
± 0.11
± 0.09
± 0.07
Masses
Margin
Shape
Density
0.66
0.72
0.63
± 0.08
±0.09
± 0.10
Otherfindings
Associated
Special
-0.02
0.38
± 0.14
± 0.12
cases

Page 4

Fig.I .-Exampie
as punctate
areaof 5 x 3 mm
Patientwas
cer but no personal
terectomy
this study,
caicificatlons
ciuded in “typically
as “suspicious
excislonal
of nonpalpable
by three
andwere
old and
history
subareolar
observers.
mlcrocaiclflcations
Calcificatlons
enlarged
history
de-
scribed
total
of five
photographically
hadmoderate
of cancer.
hormonal
prior mammogram,
In this
group,each
Ductaicarcinoma
occupied
image.In this
of breast
undergone
months
53 yearsfamily can-
hys-She had previously
therapy. but was not taking
on most
were
Eighteen
approximately
Despite
observer
in situ was diagnosed
before
to 10
term
this lesion
recent
present
benign”
five
location. choosingIn-
Interpreted
for cancer.”
biopsy.
after
776
BAKERET AL.
AJR:166, April 1996
first
and
associated
than
this difference
significant.
cialcases”
present
kappa
ciated
ment
Interobserver
nificance
0.03).
pretation
and
calcifications.
secondreadings
In
for terms
each
intraobserver
agreement.
slight,and
reproducibility
“associated
againinvolved
for intraobserven
(kappa
no betterthan
agreement
of the 60 lesions
As expected,
wasmarkedly
usedto describe
category
agreement
However,
in no instance
for selecting
findings”categories
relatively
agreement
-0.02± 0.14)
expectedby chance
for assessing
wasmoderate
intraobserverconsistency
better(kappa
masses
except
was
mostinstances
wasit statistically
termsin the “spe-
was
cases.
withregard
indicates
alone.
thediagnostic
(kappavalue,
in lesion
value,0.65
descriptive for
findings, greaten
intenobservenin
was
Poor
andagain
andfewThe
to asso-
agree-
low
value
findingsvalue,
that
sig-
0.43
inter-
±
± 0.08).
CalcificationDescriptors
A comparison
microcalcifications
cernis shown
categories
diagnostic
benign,
ablybenign,
oflesions
63%of lesions
nancywere
Despite
egories,
lesion
describing
usedterms
28 lesions
picious
werefound
“punctate”
between
and
in Table
provided
classification
0% (0/7)
only
(9/85)
(10/16)
confirmed
the overall
participating
descriptors
the
included
for which
for cancer.”
malignant
on “round”
the
observers’
observers’ descriptions
of diagnostic
of 151
con-the
5. In general,
the radiologist
scheme.
weremalignant.
2%(1/43)
described
described
as malignancies
success
observers
asinitially
BI-RADS
in the “typically
the final
In addition,
on
was
level
the BI-RADS
observers
Oflesions
Of lesions
malignant.
suspicious
as highly
at histologic
of the Bi-RADS
didnot
proposed
nomenclature
benign”
diagnostic
threeof these
histologic
usedfor 26 of these
assessment
with a successful
classified
classified
Alternatively,
abnormalities
suggestive
as
as prob-
were
as
11%
and
of malig-
inspection.
assessment
several
ACR
Study
group
assessment
28 lesions
examination.
cat-
use
in the
[1].
BI-RADS
document
participants
to describe
was“sus-
(11%)
termThe
28 lesions.All
TABLE5:Radiologists’Useof BI-RADSCalcificationDescriptors
three
pie of such
terms
“high-probability-of-malignancy”
12 lesionsbelieved
Noneof these
logicallyproven
describedwith the term
category “probably benign.”
malignantlesionsweredescribed
in Figure
branching,”
group,
as punctate.
1 . Of equal
which
were
benign
Figure
threeof the five
but with the assessment
An exam-
interest,
represent
to describe
radiologists.
2 shows
radiologists
a lesionis shown
and
the
the“pleomorphic”“fine
used
to be probably
lesionswas
benignlesion
“pleomorphic”
by the
malignant.
that
a histo-
...
Calcification
Description
BI-RADS-
DefinedGroup
No. of B enignLesions/No.of MalignantLesio ns Classifiedas:
Benign
Probably
Benign
Suspicious
Abnormality
Highly
of Malignancy
Suggestive
Skin
Vascular
Coarse
Large
Round
Spherical
Rim
Milk
Suture
Dystrophic
Punctate
or “popcorn
rod like
like”
or eggshell
of calcium
Typically
Typically
Typically
Typically
Typically
Typically
Typically
Typically
Typically
Typically
Typically
benign
benign
benign
benign
benign
benign
benign
benign
benign
benign
benign
1/0
2/0
3/0
1/0
7/0
5/0
2/0
2/0
2/0
6/1
1/0
6/0
1/0
17/3
indistinctor amorphousintermediate6/09/2
Pleomorphic
Finebranching
High
High
probability
probability
of malignancy
of malignancy
11/0
1/0
37/4
5/0
6/4
0/6
Note.-Bl-RADS
= BreastIm agingReportingandDataSystem.

Page 5

Fig.2.-Example
described
area of 4 x 4 mm
was 43-year-old
historyof breast
“pleomorphic”
observer
malignancy
of microcalcifications
as pleomorphic
and
premenopausal
cancer.
is included
Interpretedthis
was found on excisional
(at 2-o’clock
of five
position
In right
breast)
cupled
Patient
family
term
each
by three
photographically
observers.
enlarged
withno prior
mammogram.
Lesion oc-
wasin this
personal
Although
image.
woman
her first
or
This
in “high-probability-of-malignancy”
lesionas “probably
biopsy.
was
group,
benign.”No evidenceof
AJR:166,April1996
STANDARDIZEDMAMMOGRAPHYLEXICON
777
Discussion
Although
a special
mography.
ensure
grams
the
national
The
Mammography
and requires
cation
However,
are of little
beconveyed
communication
BI-RADS,
andreporting
BI-RADS
of thought
inreports
useof the
putenized
tionand
hasalso
aided
ing mammograms
No studies
implementation
reports
ologists
thestudy
quality
focus
Several
high
but
also
ACR
standards
Mammography
is an obvious
of recent
programs
standards
for their
Mammography
for quality
Quality
Accreditation
UnitedStates
of mammography
high-quality
valueunless
cleanly
between
a standardized
mammographic
reasonthat
andimproves
[1 5].Moreover,
BI-RADS
national
patientmonitoring
beenused
diagnostictechnique
concern
for those
have
for the
for all radiologists,
involved
recentlybeen
production
Amongthe
Accreditation
assurancein mammography
StandardsAct of 1992
Program,
which
Foodand Drug Administration
facilities[14].
mammograms
theinformation
toconferringclinicians.
physicians,the
systemfor describing,
abnormalities.
a standardizedlexicon
patientcane
arecentproposal
lexiconas a tool
databaseformammography
[16].The
as a standardized
to assist
[17].
to date,however,have
ofthe
BI-RADS
haveprimarilyevaluated
in diagnosticinterpretations
presentedhereused
it is
attentionwith
instituted
of mammo-
major
established
mam-
to
not only
interpretation.
efforts,
Program
[13].
buildson the
wasvoluntary,
certifi-
andinterpretations
theycontain
To
ACRhas
interpreting,
Proponents
improves
by reducing
recommends
for constructing
thatcan
improve
devised
of
clarity
ambiguity
a com-
collec-data
BI-RADS
input
radiologists
nomenclature
for a computer-
in interpret-
evaluated
lexicon.
variability
of breast
BI-RADS
thepractical
severalUnlike
thatamong
lesions
nomenclature
radi-
[4-6],
the
to investigate
lesion
deciding
determine
Although
thisstudy
phology
This
thelesions
surgical
taming
sented.
recognized
of morphology,
conducted
ing in this
Indeed,
other
common
radiologists at other institutions.
Nevertheless,
intraobserveragreements
characteristicsof
of masseswithassociated
BI-RADSnomenclature.
wasfoundfor the
suchas focalasymmetric
accountedfor a small
Theseresults
cessfulin providing
to discusslesion
beusingallof
Overonequarter
study weredescribed
diagnosticgroups
not agree.
Severalpossible
descriptionsand
havemisunderstood
tionseemsunlikely
oftheparticipants
termsincludedin the
representbenign
itednumberof
alwaysfinda term
lesion.In thisscenario,
anyof thebroad
intendeddescription
mentthatmaynot agree
termselected.
Possiblesolutions
a widevarietyof lesion
suspiciononlyto
suchas milkof calcium,
cation, or skincalcification.
pathologic etiologies (e.g.,punctate, indistinct, and pleomor-
phic),radiologistsshouldcontinue
assessmentcategoryto avoid
variabilityin radiologists’
locating
benign
characteristics
likely
themiddle
abnormalities.
noteworthy
weresufficiently
obviously
vascular
itislikely
vastmajority
a studyaddressing
as well.Second,
study,fourwork
oneobserverwas
three.Therefore,
trainingand
descriptions
abnormality
malignant,
of
almost
of evaluating
of breast
butbefore
radiologists
abnormality.
simultaneously,
morphology.
whether
the
these
investigated
of mammographic
studyhas
evaluated
biopsy.
lesions
Although
by the
After
it
an
oris
nadiologic
actions
the
occur
stepthe mon-
two
limitations.First,mostof
abnormal
entities
to warrant
asfat-con-
not
would
on the
lesions
radiologists
togetherat one
responsiblefor training
mayreflect
somewhatfrom
Such
or
benign
calcifications
that
of radiologists
were
lesions
nepre-
suchbe
basis
thesecould
participat-
institution.
be
of the five
closely
in part
results
vary
the
the
may
somelevelof
resultsfor
theresultsdemonstrate
determining
calcifications,
microcalcifications
Considerably
involving
densities.
minorityof all the
suggestthat
a standardized
morphology.
thedescriptive
of themicrocalcifications
withBI-RADS
withwhich
good
the
and
inter-and
formorphologic
combinations
by use
agreement
typesof lesions,
However,
casesevaluated.
is moderately
languagefor physicians
However,radiologists
termsasinitially
evaluated
termsincluded
theevaluating
masses,
of the
poorer
fewcasesother
thesecases
BI-RADSsuc-
may
intended.
not
in this
in broad
radiologistsdid
explanations
diagnoses
the definition
given
in this study.
“typically
lesions.
descriptive
that
a reader
groups
and
for the
First,
of the terms.
levelof mammography
Second,
benign”group
possibility
choices,
believeadequately
must“settle”
mostclosely
provide
the BI-RADS
discrepancy
the
between
exist.radiologists
This
may
explana-
training
BI-RADS
not always
given
may
describes
for the
approximates
a diagnostic
assessment
the
someofthe
may
Athirdis that,a lim-
radiologistsnot
theya
interm
that
then
with
the
assess-
of the
to these
descriptors
thosedescriptors
vascular
problemsinclude
assigning
known
maintaining
a level
etiologies,
suturai
without
but
with
of
calcification,
descriptors
calcifi-
knownFor
toassignaBI-RADS
in dic-unnecessaryambiguity