Clasificacion y errores de interpretacion de lesiones meniscales

MR Imaging of the Knee: Expanded Classification and Pitfalls to Interpretation of 1 Memscal Tears
.

Mamed Mesgarzadeb, David S. Leder, MD Alexander Russonieio, Jamsbid Tebranzadeb,

MD #{149} George MD MD

#{149}

#{149}

Ray Moyer, MD Revesz, PhD #{149} kbar A Bonakdarpour, Dan Guttmann, MD

MD

Findings knees cluding findings throscopic MR imaging tive were and 74% cus. teardictable Thus, meniscus even may if the be were

from not were only

magnetic interpreted signal correlated

resonance retrospectively intensity with performed were expanded for each potential imaging meniscus grade Medial from of the of the findings and meniscal the type

(MR)

imaging with a new recorded

examinations classification,

of

194 in-

but

morphologic
by a knee toeight of finding

abnormalities.
results from subspecialist. types in this and overall that

All
ar-

prospectively

examinations findings values

Meniscal study. Posiaccuracy be dif-

predictive calculated.

Several of MR medial

sources change 86% were were, and tears posterior or body posterior This suggestive

of error

could

ferentiated specificity for also when theAs the

through

pathologic

recognized. respectively, 90% for likelihood to occur to the horn should level horn their the

Sensitivity

95% and
lateral menisin a prehorn. medial suspected, of the be classification of confidence injury. of meniscal anterior

classification increased. extending tear

increased,

the

tended anterior expanded

pattern, is present, findings for MRuseful

a definite are

a tear not radiologists imaging

convincing. in expressing finding

in a particular

of meniscal

Abbreviations: Index terms:

ROC Knee, 1993; Departments

=

receiver

operating 452.4852
#{149}

characteristics, Knee, ligament,

SE

=

spin and

echo cartilage, 452.4852

injuries, 13:489-500

menisci,

RadloGraphics
I

From

theofDiagnostic Broad UT.). and From 1, 1993; Irvine February

Imaging Tioga the accepted 1991

(MM., Sts, RSNA March

D.S.L., scientific

G.R.,

A.B.,

D.G.) and

and

Orthopaedic 26, to MM.

Surgery of Radiology, 1992; revision

(R.M.,

AR.), University

Temple ofCalifornia,
RSNA,

University Irvine, 1993

Hospital,

Philadelphia. 8. Address

PA 19140; assembly.reprint

the Department

Receivedjune requests

requested

July 20 and received
C

489

U INTRODUCTION Use of magnetic resonance (MR) imaging to diagnose interarticular derangements of the knee joint has been well documented in severa! articles concerning the normal anatomy (1-3), osteochondritis dissecans (4), and injuries of the menisci (5-18) and ligaments (10, 19-23). There areseveral reports concerning the accuracy, sensitivity, and specificity of MR imaging in diagnosing meniscal injuries. In most of these reports, the meniscal MR imaging findings were divided into only three types according to the classification by Lotysch et a! (8) and Crues et a! (9). In these reports, the likelihood of false-positive or true-negative findings for a truncated or Severely smallmeniscus was not analyzed. Also, it was not determined whether a comminuted type of increased signal intensity or a bandlike increased signal intensity extending to one or both surfaces of the meniscus has the same or different accuracy. Furthermore, in those artides, arthroscopy was performed by several surgeons, and there were no prospective protocols used to report the arthroscopic results. Weanalyzed the results of MR imaging and arthroscopy in 194 consecutive knees. A prospective protocol was designed for arthroscopic findings, and the arthroscopist was a single surgeon in all cases who had extensive experience in knee arthroscopy. The meniscal MR imaging findings were expanded to eight types in this study (Fig 1). Our types 0, I, and II findings were identical to those described in...