8/6/2019 MRI Artefacts AJR115

http://slidepdf.com/reader/full/mri-artefacts-ajr115 1/4

11 5

0361-803X/93 /1611-0115

© Am e ric an R oen tg en R ay S oc ie ty

MR D iagnos is o f In te rna l

D e rangem en ts o f the K nee : E ffec to f F ie ld S treng th on E fficacy

M ichae l J . B a rne tt1

R ece ive d O c tob e r 9 , 199 2 ; ac ce p te d a fte r rev i-

s io n F eb ru ar y 9 , 1 99 3.

1 M ed ica l D ia gn os tic C en te r, 18 1 7 R iv e rd a le S t.,

W e st S prin g f ie ld , M A 01089 . Ad dress co rn esp on -

de nce to M . J . Ba nne tt .

OBJECT IVE . The pu rpose o f th e s tudy w as to tes t th e hypo th es is th a t d iagnoses o f

in te rn a l d e rangem ents o f th e knee based on fin d ings on MR Im ages ob ta in ed a t 0 .5 T

a re as accu ra te a s those based on find ings on M R im ages ob ta in ed a t 1 .5 T .

MATER IA LS AND M ETHODS . The M R find ing s in 118 consecu tive pa tien ts w ho had

MR im ag ing o f the kn ee a t 0 .5 T and subsequen t a r th roscopy w ere rev iew ed re tro -

spec tive ly . N in e ty -tw o pa tien ts h ad a rth ro scop ica ily p ro ved te a rs o f th e m ed ia l

m en is cus , la te ra l m en iscus , and /o r an te r io r c ru c ia te lig am en t. S ens itiv ity , s pec if ic ity ,

a nd accu ra cy fo r M R eva lua tio n o f th e m ed ia l a nd la te ra l m en isc i and an ter io r c ru e l-

a te lig am en t w e re de te rm in ed by us ing arth ro scopy as th e go ld s tandard . The re su lts

w e re com pared w ith s im ila r d a ta from 11 recen tly pub lish ed s tud ie s in w h ich M R

im ages ob ta in ed a t 1 .5 T w ere used . To accoun t fo r am b igu ity in M R in te rp re ta tio n ,

the M R find ings in th is s tudy w ere rean a lyzed by us in g a five -po in t sca le o f d iagnos -

t ic ce rta in ty , a nd re ce ive r-ope ra tin g -charac te ris tic cu rv es w ere constru c ted fo r e ach

o f th e th re e anatom ic s tru c tu re s exam ined .

RESULTS . Fo r te a rs o f th e m ed ia l m en is cus , th e sens itiv ity , s p ec ific ity , a nd accu -

rac y o f M R at 0 .5 T w e re 93% , 90% , and 92% , re spec tive ly . F o r te ars o f th e la te ra l

m en is cus , th e sens itiv ity w as 81% , the spec ific ity w as 97% , and th e accu ra cy w as

9 3% . The sens it iv ity , s pec if ic ity , a nd accu ra cy fo r d e te c tin g com p le te tea rs o f th e

an te rio r c ru c la te lig am en t w ere 100% , 9 7% , and 97% , re spec tive ly . N o c lin ic a lly s ign if-

ic an t fie ld s treng th -d ependen t d iffe ren ces w ere found .

CONCLUS ION . The resu lts sugges t th a t, a llow ing fo r n ecessary d is crepanc ie s in

im ag ing p ro to co l, m agnetic f ie ld s treng th is no t a s ign ifican t d e te rm in an t o f d iagnos -

t ic re liab ility o f M R assessm en ts o f in te rna l d erangem en t o f th e knee .

AJR 1993 ;161 :115 -118

D esp ite w idesp read use o f m id-fie ld -s treng th M R un its , little in fo rm a tion is ava il-

ab le on the accu racy o f us ing m id-fie ld -s treng th M R im ag ing to eva lua te in te rna l

de rangem en ts o f the knee . E leven s tud ie s pub lished s in ce 1987 in wh ich 1 .5 -T

m agne ts w e re used [1 -ii] and one in wh ich a 1 .0 -T m agne t w as used [1 2 ] p rov ide

s ta tis tica l da ta on the e fficacy o f M R im ag ing fo r the d iagnos is o f tea rs o f the

m en isc i o r an te rio r cm uc ia te ligam en t. O n ly fiv e s tud ie s pub lished a fte r 1 987 used

m agne ts o f le ss than 1 .0 -T fie ld s treng th [11 , 13 -16 ], and none w e re pub lished in

the Am e rican rad io log ic lite ra tu re . O ne study [11 ] com pa red resu lts ob ta ined w ith a

0 .35 -T sys tem w ith those ob ta ined w ith a 1 .5 -T sys tem . The au tho rs found no s ig -

n ifican t d iffe re nce in eva lua tions o f team s o f the an ter io r c ruc ia te ligam en t on la te ra l

m en iscus bu t conc luded tha t im ages ob ta ined w ith the 1 .5 -T sys tem w e re m o re

re liab le fo r eva lua tion of the m ed ia l m en iscus . The s tudy , how eve r, w as b iased .

W ith the 0 .35 -T un it, on ly sag itta l im ages w e re ob ta ined , and a poste rio r, re ce ive -

on ly co il w as used , w he reas w ith the 1 .5 -T un it, s ag itta l and co rona l im ages w e re

ob ta ined , and a quad ra tu re co il w as used (pe rsona l commun ica tion ).

A t leas t one recen t a rtic le [2 ] sugges ts tha t the accu racy o f d iagnoses based on

M R find ings m ay depend on the fie ld s treng th o f the m agne t used for im ag ing .

T he pu rpose o f th is s tudy w as to com pa re the re liab ility o f im ages ob ta ined w ith a

0 .5 -T M R sys tem w ith those ob ta ined a t 1 .5 T fo r d iagnos ing tea rs o f the m en isc i

and an te r io r c ruc ia te ligam en t.

8/6/2019 MRI Artefacts AJR115

http://slidepdf.com/reader/full/mri-artefacts-ajr115 2/4

116 BARNETT AJR:161 , J uly 1 99 3

M ate ria ls and M e thods

B etw ee n O cto be r 1 99 1 and August 1 992 , 363 patien ts had M R

eva lua tion o f the knee a t th is in s titu tion ; 1 18 o f these subsequen tly

had arth roscopy , w h ich w as pe rfo rm ed by any one o f 22 o rthoped ic

su rgeons w ith a m ean o f 9 .4 yea rs (range , 1 -20 yea rs ) o f pos tre s i-

den cy a rth ro scop ic e xpe rien ce . S ligh tly m ore tha n ha lf (6 0 ) th e

a rth roscop ies we re pe rfo rm ed by e ithe r o f tw o knee surgeons w hohad fe llow sh ip tra in ing in th is p rocedu re . The tim e in te rva l be tw een

MR im ag ing and a rth roscopy va r ied from 1 to 192 da ys .

The M R stud ie s o f these 118 patien ts w e re re tro spec tive ly re -

v iew ed by a rad io log is t w ith 6 yea rs pos tre s idency expe rience in

M R in te rp re ta tion . Fo r in itia l ana lys is , each case was ca tego rized as

pos itive on nega tive by fo rced cho ice w ith rega rd to show ing tea rs o f

the m ed ia l and la te ra l m en isc i and com p le te tea rs o f the an te r io r

c ruc ia te ligam en t (pa rtia l tea rs o f the ligam en t w ere ca tego riz ed as

nega tiv e because o f the d ifficu lty in ve rify ing som e o f the a rth ro -

scop ic repo rts ). T o m ain ta in ob je c tiv ity , the s tud ie s w ere in te rp re ted

w ithout know ledge o f the o rig ina l rad io log is t’s repo rt o r the a rthro -

scop ic find ings . T he re tro spec tiv e find ings w ere then com pa red w ith

the a rth roscop ic repo rts ; the a rth roscop ic desc rip tions we re used as

the go ld s tanda rd . Fo r each o f the th ree s truc tu res exam ined , each

case w as des igna ted true -pos itive if a tea r w as seen on M R im ages

and con firm ed a rth roscop ica lly , fa lse -pos itiv e if a tear suspec ted on

the bas is o f M R find ings was no t a rth roscop ica lly confirm ed , fa lse -

nega tiv e if a tea r seen a t a rthnoscopy w as no t suspec ted on the

bas is o f M R find ings , o r true -nega tiv e if no tea r was suspec ted on

the bas is o f M R find ings and no tea r w as seen at a rth roscopy .

T hese da ta a re summ arized in Tab le 1 ; from th is in fo rm a tio n , s en s i-

t iv ity , sp ec ific ity , an d a ccu ra cy w ere ca lcu la ted .

Fo rced cho ice is a som ew ha t a rtific ia l s itua tion ; ty p ica lly , a rad io l-

og is t’s repo rt a ls o re fle c ts a con fidence le ve l. T o accoun t fo r th is

va riab le , the MR stud ies w e re re in te rp re ted by us ing a five -po in t

sca le o f d iagnos tic ce rta in ty . E ach case w as c la ss ified as e ithe r de f-

in ite ly pos itive , p robab ly pos itive , equ ivoca l, p robab ly nega tive , o r

de fin ite ly ne ga tive w ith reg a rd to te a rs o f the th re e s truc tu re s e xam -

m ed . The a rth roscop ic resu lts , the resu lts o f the fo rced -cho ice ana l-

ys is , and the o rig ina l rad io log is t’s in te rp re ta tions we re unknown by

the pe rson who c la s s ified the cases . The M R find ings w e re then

com pa red w ith the a rth roscop ic repo rts ; these da ta a re sum marized

in Tab le 2 . W ith the se da ta , re ce ive r-op e ra ting -ch a rac te ris tic (RO C )

cu rves w e re cons tru c ted fo r tea rs o f the m ed ia l m en iscus and a t-

ena l m eniscus and fo r com p le te te a rs o f the an te rio r c ruc ia te liga -

m en t, b y us in g a com pu te r p rog ram des ig ned by C harles M etz

(D epa rtm en t o f R ad io logy , U nive rs ity o f C hicago ). T he re la tive e ffi-

ca cy o f a te s t (in th is ca se , M R im ag ing o f the knee a t 0 .5 T ) is p ro -

po rtio na l to th e a re a unde r th e ROC curv e [17 -1 9 ].

TABLE 1 : R esu lts o f Fo rc ed -C ho ice M R Eva lu a t io n o f th e Knee

at 0 .5 T in 118 Pat ie n ts

Loca tion o f T ea rs (N o. o f P atien ts )Resu l t s Med ia l

Men is cus

Latera l

Men is cus

A n te rio r C ru cia te

L igamen t

True -pos i t ive 7 1 21 26

T rue-nega tiv e 38 89 89

False-pos i t ive 4 3 3

Fa lse -nega tiv e 5 5 0

Note-True-pos it ive = te a r s een on M R im ages an d co n firm ed a rth ros co p -

i ca l ly ; t ru e -n e g at iv e = no tea r s ee n o n M R im age s o n ar th roscop ica lly ; fa ls e -

pos it ive = te a r s usp ec ted on bas is o f M R im ages b u t n o t c on firm ed ar thn o -

s c o pi ca l ly ; f a ls e -n e g at iv e = te a r n o t s usp ec ted o n bas is o f M R im ages bu t

s e e n a n th r os c o pi ca l ly .

TABLE 2 : Ana lys is w ith a F ive -Po in t-S ca le o f D iagnos tic

C erta in ty o f R esu lts o f MR Eva lua tion o f the Knee a t 0 .5 T in 1 18

Pat ients

Fi n d in gs o n A rth no sco py

Find ings

on M R

Med ia l

Men i s c u s

La te ra l

Men i s c u s

A nte rio r C nu cia te

L ig amen t

T No trnTorn

N otTorn

Torn

N otTorn

Torn

De fin ite ly to rn 6 3 3 16 1 27 3

P robab ly to rn 7 2 3 0 1 0

Equ ivoca l 1 1 2 2 0 1

P robab ly n o rm a l 2 6 1 6 0 1

De fin ite ly no rm a l 3 3 0 3 84 0 85

The m ajo r c r ite rion used fo r d iagnos ing a m en isca l tear w as de fi-

n ite e x te ns ion o f in tram en isca l inc re ase d s ign a l o n an y M R im age to

the supe rio r o r in fe rio r m en isca l su rfa ce ; suppo rting cr ite ria in c luded

b lu n t ing o f a fre e e dg e , m en isca l d e fo rm ity , a nd s ig na l in the m idd le

th ird o f a m en iscus w ith in 10 mm o f the m en isca l pe r iphe ry on sag it-

ta l im ages . A n an te r io r c ruc ia te ligam en t was ca lled com p le te ly to rn

if n o low -s ig na l-in te ns ity fib e rs co u ld be de te c te d in c on tin u ity from

the fem u r to the tib ia a long the expec ted cou rse o f the ligam en t;

supporting c rite ria inc lud ed con tu s io n s o f the m idpo r tion o f th e la t-

e ra l fem ora l condy le o r o f the pos te r io r tib ia l p la teau , so ft-tissue

th icke n ing a lo ng th e e xpe c te d cou rse o f th e lig am en t, p ro g ress ive ly

inc rea s ing s ign a l a lo n g th e e xpe c te d cou rse w ith in c re as ing T2

we ig h ting , a n d p rom in en ce o f th e la te ra l fem ona l n o tc h . T hese cn ite -

na a re s im ila r, a lthough no t iden tic a l, to those used in s tud ie s [1 , 2,

4- i l ] in w h ich i.5 -T m agne ts w e re u sed .

A ll pa tien ts w ere im aged w ith a 0 .5 -T s ys tem (M ax ; G ene ra l E le c -

tn ic M ed ica l S ys tem s , M ilw aukee , W I). A c ircum fe ren tia l transm it-

re ce ive co il (M ed ica l A dvances , In c ., M ilw aukee, W I), a 15 -cm fie ld

o f v iew , a 1 92 x 1 92 m a trix , and a 1 -m m in te rs lice gap we re used .

Fo r a ll pa tien ts , sag itta l p ro ton dens ity -w eigh ted and T2 -we igh ted

im a ge s ( 20 00 /2 0-9 0 [TR ITE ], tw o ex c ita tion s , 5 -m m s lice th ic k -

ness ), c o rona l T i -w e igh ted im ages (7 00 /25 , tw o e xc ita tio n s , 5 -mm

s lic e th ic kness ), and ax ia l g rad ien t-e cho im ages (700 /25 , 40# {176}lip

a ng le , tw o e xc ita tio ns , 5 -mm s lice th ic kne ss ) w e re ob ta in e d o f th e

en tire knee , and m en is ca l w indow s w ere film ed o f the sag itta l p ro -

ton dens ity -w e igh ted and co rona l T i -w e igh ted im ages. O b lique

sa g it ta l T i -w e ig h te d im age s (5 50 /30 , fou r ex c ita tio ns , 3 -m m slice

th ickness) o f the an te rio r c ruc ia te ligam en t w e re a lso acqu ired in

eve ry case . In patien ts w ho had traum a 1 m on th on less be fo re

im ag ing , a co ron a l g rad ien t-e ch o se quen ce (6 00 /30 , 30 # {1 76 }lip a ng le ,

tw o exc ita tions , 5 -m m slic e th ickness) w as added fo r be tte r eva lua -

t io n o f th e co lla te ra l l ig am en ts and a sse ssm en t o f m en isco cap su la n

sepa ra tions . To ta l tim e requ ired fo r im age acqu is itio n ranged from

28 mm 48 sec to 32 m m 38 se c ; in m os t ca se s , im age s we re

o bta in ed w ith in a n a llo ca te d 1 -h r tim e s lo t.

Resu l ts

The sens itiv ity o f M R im ages ob ta ined a t 0 .5 T fo r de tec t-

ing to rn m ed ia l m en isc i w as 93% , the spec ific ity w as 90% ,

and the accu racy was 92% . The co rrespond ing RO C cu rve

is show n in F igu re 1 . The sens itiv ity , spec ific ity , and accu -

racy o f M R im ages ob ta ined a t 0 .5 T fo r the d iagnos is o f

team s o f the la tera l m en iscus w e re 81% , 97% , and 93% ,

respec tive ly . F igu re 2 show s the assoc ia ted ROC cu rve . T he

sens itiv ity o f M R im ages ob ta ined a t 0 .5 T fo r de tec tion o f

8/6/2019 MRI Artefacts AJR115

http://slidepdf.com/reader/full/mri-artefacts-ajr115 3/4

Stu d y

H e ron and C a lve rt [1 ]

R uw e e t a l. [2 ]

Quinn and B rown [3 ]

A rak i e t a l. [4 ]

Vahe y e t a l. [5 ]

N iitsu e t a l. [6 ]

M in k e t a l. [7 ]

Lee e t a l. [8 ]

R aune s t e t a l. [9 ]

Qu in n e t a l. [10 ]

F isch e r e t a l. [1 1 ]

N o . o f C a se s R eg ion (s ) E va lu a te d

io o Men is c i

ACL

103 M en isc i a nd AC L

254 M ed ia l m en is cus

L ate ra l m en is cu s

40 M en isc ia

Men i s c i b

81 A cu te ACL

O ve ra ll AC L

52 ACLC

242 M en is c i

A CL

79 ACL

50 M en isc i

259 M ed ia l m en iscus

630 M ed ia l m en iscus

L ate ra l m en is cu s

AC L

Sens it iv i ty (% )

97

92

92

70

82

10 0

10 0

92

96

95

10 0

94

64

90

97

70-71

92-97

Spec if ic i ty (% )

94

96

82

92

10 0

10 0

93

93

92

91

96

i 00

83

95

79 -88

94-96

93

Ac cu ra c y (% )

94

88

89

10 0

96

93

93

95

10 0

78

92

88-93

87-92

93-94

advantages o f sm a lle r m agne ts unde rsco re the im po rtance becom es less s ign ifican t.

;_

I 0 .8

0 .9 0 .7 I

0 .9 960.8

0.6

0 .7 0.9 9m a. p os itiv e

0.6 fraction 0 .986T rue T rue

positive 0.6 positive 0.98

fraction 0.4 fraction

0 .976

0.30.97

0.2

0 .1 0.9 66

0.96

0 0 .6

Fa lse p os *lv e f ra ct io n

F i g . 1 . -R e c e iv e r- ope ra t in g - c har a c te r -

is tic c u rve fo r d ia gn os is o f m ed ia l m en is -

cus te a rs based on fin d ings on M R Im ages

ob ta in ed a t 0 .5 T . A rea unde r cu rve is

0 .9585 ; s tand ard dev ia tio n is 0 .01 75 . F in al

v a lu e o f p a ram ete r A Is 2 .3365 ; f in a l v al ue

o f p aram ete r B Is 0 .9034 .

A JR :16 1 , J u ly 199 3 MR OF IN TERNAL DERANG EM ENTS O F KNEE 117

TABLE 3 : E ff ic a cy o f M R Im ag ing a t 1 .5 T in E va lu a tio n s o f K nee M en isc i and AC L In 11 S tud ie s Pub lish ed A fte r 19 87

Note . -ACL = an te r io r c ruc ia te l igam en t; d ash es = in fo rm atio n n ot p ub lish ed .aTwod im ens iona l Fo u rie r t ran s fo rm on ly .

bC om bined tw o -d im ens ion a l a nd th re e -d im en s io na l F ou rie r t ran s fo rm im ag ing .

C C in e i ma gin g.

com ple te tea rs o f the an te r io r c ruc ia te ligam en t was 100% ,

and the spec ific ity and accu racy we re bo th 97% . The com e -

spond ing ROC cu rve is show n in F igu re 3 .

Discuss ion

The p reva lence o f M R im ag ing sys tem s tha t ope ra te a t

le ss than 1 .0-T fie ld s treng th and the re la tiv e econom ic

0.3

0.2

0.1

0

of assess ing the d iagnos tic re liab ility o f im ages obta ined a t

low fie ld s treng ths vs those ob ta ined a t h ighe r fie ld

s treng ths . T he assum p tion tha t a h igh -fie ld -s treng th m agnet

is be tte r is based m a in ly on the d ire c t re la tionsh ip betw een

the s igna l- to -no ise ra tio and m agne tic fie ld s treng th . W ith

lowe r s treng th m agne ts , m uch o f th is can be com pensa ted

fo r by a cqu irin g m ore da ta , an d , as fas ten im ag ing tech -

n iques becom e ava ilab le , the add itiona l im ag ing tim e

0 0 .2 0 .4 0.8 0.8

Fa lse p os ft tv e fr ac ti on

F ig . 2 .-R ec eiv er -o pera t in g-c hara cte ris tic cu rve fo r

d iagnos is o f la te ra l m en is cus te ars based on fin d ings

on MR im ages ob ta in ed at 0 .5 T . A re a under cu rve is

0 .9 61 4; s ta nd ard d ev ia tio n is 0 .0285 . F In a l va lu e o f p a -

ram eter A Is 2 .1 668 ; fina l v a lu e o f p aram ete r B is 0 .7 0 95 .

0 0.2 0 .4 0.6 0 .8

Fats . positive f r act ion

F ig . 3 . -Re ce lv e r - op e ra t in g - c h a ra c te ri s -tic cu rv e fo r d iag no sis o f co mple te an ter io r

c ru c ia te lig am en t te a rs based on fin d ings

on M R im ages ob ta in ed at 0 .5 T . D a ta degen -

cra te ow ing to nea rly exac t f it. B inom ia l re -

ce lve r-o pe ra tin g -c ha rac te ris tic cu rve is

ve rtica l a t con s ta n t fa lse -po s it ive frac tio n o f

0 .0333 .

8/6/2019 MRI Artefacts AJR115

http://slidepdf.com/reader/full/mri-artefacts-ajr115 4/4

11 8 BARNETT A JR :1 61 , J uly 19 93

The pu rpose o f th is s tudy w as to de term ine if a ssess-

m en ts o f in te rna l de rangem en ts o f the knee we re as accu -

m a te w hen conven tiona l im ag ing techn iques and a 0 .5 -T

m agne t we re used as w hen a 1 .5 -T m agne t w as used . T o

m y know ledge , on ly one s tudy [11 ] a ttem p ted to com pa re

accu rac ie s o f d iagnos is o f knee abno rm a litie s based on find -

ings on M R im ages ob ta ined a t d iffe ren t fie ld s treng ths . T heconc lu s ions w e re m ixed and d ifficu lt to in te rp re t. H oweve r,

the au tho rs s ta ted tha t, fo r the d iagnos is o f tea rs o f the

m edia l m en is cus , im ages ob ta ined a t 1 .5 T we re superio r to

those ob ta ined a t 0 .35 T .

S ens itiv itie s , spec ific itie s , and accu rac ies for eva lua tions

o f tea rs o f m en isc i and an te r io r c ruc ia te ligam en ts in s tud ie s

pub lished a fte r 1987 tha t used 1 .5 -T im ag ing sys tem s a re

summ arized in T ab le 3 . L ike the p resen t s tudy , a ll u sed

a rth roscopy as the go ld s tanda rd (the va lid ity o f w h ich is lim -

ited , as pub lis hed m a tes o f accu racy o f a rth roscopy range

from 70% [20 ] to 97% [21 ], m ost like ly s im ila r to those

ob ta ined w ith M R im ag ing).

F o r eva lua tion o f m en isca l tea rs w ith M R im ages ob ta ined

a t 1 .5 T , repo rted accu rac ie s range from 78% [9 ] to 100%

[4 ] , sens itiv itie s from 64% [9 ] to 1 00% [4 ], and spec ific itie s

from 82% [3 ] to I 00% [4]. In th is s tudy , the ove ra ll a c cu racy ,

sens itiv ity , and spec ific ity fo r eva lua tion o f m en is ca l team s

w ith M R im ages ob ta ined a t 0 .5 T we re 93% , 90% , and 95% ,

respec t ive ly .

F o r eva lua tion o f team s o f the an te r io r c ruc ia te ligam en t

w ith MR im ages ob ta ined a t 1 .5 T , repo rted accu rac ie s

range from 93% [5 ] to 100% [8 ], sens itiv itie s from 92% [1 , 5,

1 1 ] to 1 00% [71 , and spec ific itie s from 92% [6 ] to 1 00% [8 ]. In

th is s tudy , fo r de tec tion o f com ple te team s o f the an te r io r c ru -

c ia te ligam ent w ith M R im ages ob ta ined a t 0 .5 T , the accu -

racy and spec ific ity w e re 97% and the sens itiv ity w as 100% .

The lack o f su rg ica lly s ign ifican t le s ions in the I 9 pa tien ts

who had no abno rm a litie s on M R im ages obta ined a t 0 .5 T

suppo rts the con ten tion tha t MR im ag ing a t 0 .5 T is a powe r-

fu l too l fo r sc reen ing cand ida tes fo r a rth roscopy , as has

been show n fo r im ag ing a t 1 .5 T [2 , 3 ].

The re tro spec tive na tu re of th is s tudy in troduces po ten tia l

sub je c tive b ia s ; I th in k th is has been con tro lled by ca tegon iz -

ing the MR find ings a p rio r i w ith regard to know ledge o f the

a rth roscop ic resu lts and w ithout the bene fit o f p rev ious

m ad io log ic in te rp re ta tions . The idea l s tudy w ou ld be doub le -

b lind and p rospec tive w ith the sam e sub jec ts im aged by

us ing m agne ts o f d iffe ren t fie ld s treng ths . S uch a s tudy

wou ld be cum be rsom e and cos tly and cou ld be pe rfo rm ed

on ly a t a lim ited num be r o f in s titu tions tha t have access to

h igh -fie ld - and m id -fie ld -s treng th sys tem s.

P e rhaps the m a jo r sou rce o f inhe ren t b ia s in s tud ie s tha t

re ly on an th roscop ic da ta fo r ana lys is ( in c lud ing a ll l is ted in

Tab le 3 ) is the la c k o f random iza tion o f the sam p le , as on ly

those chosen fo r a rth roscopy (p resum ab ly those w ith the

m ost seve re s igns and sym p tom s) a re in c luded . T h is p rob -

lem is un fo rtuna te ly unavo idab le , a lthough a t leas t one

g roup o f in ves tiga to rs [2 ] has c le ve rly add ressed the is sue

by in troduc ing c lin ic a l “ou tcom e ” as a dependen t va r iab le .

P resum ab ly , the onus o f longe r exam ina tion tim es a t

low e r fie ld s treng ths w ill a t leas t be pa rtia lly a lle v ia ted by

recen t te chn ica l deve lopm en ts .

B ecause o f the w idesp read use o f m id -fie ld -s treng th M R

equ ipm en t and the obv ious econom ic im p lica tions , fu rthe r

s tud ie s o f the im pac t o f fie ld s treng th on the d iagnos tic re li-

ab ility o f M R find ings is w a rran ted .

REFERENCES

1 . H eron CW , C alve r t P T. T h ree -d im en s io n a l g ra d ie n t-e ch o M R im ag in g o f

th e kn ee : co mp ariso n w ith a rth ros cop y in 10 0 pa t ien ts . Rad io logy 1992 ;

183 :836 - - 844

2 . R uwe P A , W righ t J , R anda ll R L , Lyn ch JK , J ok i P , M cCa rth y 5 . C an MR

im ag in g e ffe c tiv e ly rep lace d iagn os tic a r thn os co py? Rad io logy 1992 ;

183 :335 -339

3 . Q uin n SF , B rown TF . M en is ca l te a rs d iag nosed w ith M R im ag in g ve rsu s

a rth no sco py : ho w re lia ble a s tan dard is a rthn oscop y? Rad io logy 1991181 :

843-847

4. A ra k i Y , O ota n i F , T su kagu ch i I, e t a l. M R d iag nos is o f m en isca l tea rs o f

th e kn ee : va lue o f ax ia l th ree -d im en sio na l F ou rie r tran sfo rm ation G RA SS

images . M R 1992 ;158 :587 -590

5 . V ahey TN , B ro om e DR , Ka ye s K J , S he lbo u rne KD . A cu te a nd ch ro n ictears o f th e an te rio r c ru cia te lig am en t: d iffe ren tia l fe atu re s a t M R im ag in g.

Rad io logy 1991 1 8 1 : 2 5 1 - 25 3

6 . N iitsu M , A nno I, F u kuba ya sh i T , S h im ojo H , K uno 5 , A k is ada M . T ea rs o f

c ruc ia te ligam en ts a nd m en is c i: e va lu a tion w ith c ine M R im ag in g . Rad i -

o iogy 1991 ;178 :859 -864

7 . M ink JH , L evy T , C rue s JV . T ea rs o f th e an te rio r c ruc ia te ligam en t and

m en isc i o fthe kne e : M R im ag ing e va lua tion . Rad io logy i 988 ;167 :769 -774

8. Le e JK , Y ao L P he lp s C T , W ith C R , C za jk a J , L ozm an J . A n te rio r c ruc i-

a te ligam en t tea rs : M R im ag in g com pa re d w ith a rthn os cop y and c lin ica l

tes ts . Rad io logy 1988 ;166 :861 -864

9 . R aune s t J , O berle K , L oehne rt J , H oe tz in ge r H . T he c lin ica l va lue o f m ag-

n etic reso nan ce im ag in g in th e e va lu atio n o f m en isc al d iso rd ers . J B one

Jo in t S urg (A mJ 1 9 9 1 ; 73 - A :1 1 -1 6

10 . Q uinn S F, B row n T R, S zum ow sk i J . M en isc i o fth e knee : rad ia l M R im ag ing

co r re la ted with a rt hr os co py in 2 59 p at ie nt s. Rad io logy 1992 ;185 :577 -580

11 . F is ch e r 5P , F ox JM , De lP izzo W , F r iedm an M J, S nyd er SJ , F e rk e l RD .

A ccu ra cy o f d ia gn oses from m agne tic reson ance im ag in g o f the kne e . JBone Jo in t S urg [A m] 1 99 1 ;7 3-A :2 -1 0

12 . R eeden JD , M atz SO , B e cke r L , A nde lm an SM . M R im ag in g o fthe knee in

sa g it ta l p ro je c tion : com parison o f th ree -d im ens iona l g ra d ie n t-ech o and

s pin -e ch o s eq ue nc es . AJR 1 98 9 ;1 5 3 :5 3 7- 54 0

13 . S ilva I, S ilv e r D M. T ea rs o f th e m en iscu s a s re vea led b y m agne tic ne so -

n an ce im ag in g. J Bone Jo in t Su rg (Am ] 1988 ;70 -A:199 -202

14 . Boeree N A, W atk in so n A F, A ckno yd C E, Jo hn son C . M agne tic res onan ce

im a g in g o f m en is ca l an d cruc ia te in ju ries o f the kne e . J B one Jo in t Su rg

(Br ] 199173-6 :452 -457

15 . J am es P . B u irsk i G . M R im ag in g o f th e knee : a p rospe c tive tria l u s ing a

lo w fie ld s tre ng th m ag ne t. A ustra la s R ad io l 1990 ;34 :59 - 63

16 . G las how JL , K atz A , S chne ide r M , S co tt W N. D oub le -b lin d as ses sm en t o f

the va lu e o f m ag ne tic re son ance im ag ing in th e d ia gn os is o f an te rio r c ru -

c ia te a nd m en is ca l le sio ns . J Bone Jo in t Su rg (Am ] 1 98 9;7 1 -A :1 1 3-1 19

17 . M etz C E. RO C m eth odo lo g y in ra d io log ic im ag in g . I nv es t R a dio l 1986 ;

21 :720 -733

1 8 . B nism ar J . U nders tan d ing re ce ive r-o pe ra ting -ch a rac te ris tic c u rv es : ag ra ph ic a pp ro ac h. A JR 1991157 :1119-1121

19 . V in in g D J, G lad is h G W. R ece ive r ope ra ting c ha ra c te ris tic cu rves : a ba s ic

unders tand ing . Rad ioG raph ic s 1 9 92 ;1 2 :1 1 47 -1 1 54

20 . L ev in so hn EM , Ba ke r BE . P nean th no tom y d ia gno s tic ev a lua tio n o f th e

knee : re v iew o f 1 00 ca ses d iag no sed b y a rth nog raph y and a rth ro scopy .

AJR 1980 ;134 :107 -111

21 . S e le sn ick FH , N ob le H B, B achm an D C, S te inb e rg FL In te rna l de ra nge -

m en t o f th e kn ee : d ia gno sis by an th rog na phy , a nth ros cop y an d arth ro to my .

C lin O rth op 1985 ;198 :26 - 30