357

Downloaded from www.ajronline.org by 50.32.148.132 on 11/02/15 from IP address 50.32.148.132. Copyright ARRS. For personal use only; all rights reserved

The Cruciate Ligaments of the Knee: Correlation Between MR Appearance and Gross and Histologic Findings in Cadaveric Specimens

Juerg

2

Parviz Haghighi3 Debra TrudeIl1 Donald Resnick1

OBJECTIVE. for differences

cruciate

Using anatomic and histologic correlation, we investigated in the MR signal intensity between the normal anterior

ligaments

and for the focally

increased

signal

intensity

the reasons and posterior

within these

ligaments

occasionally found in elderly patients. MATERIALS AND METHODS. The MR images of 18 anterior and 20 posterior cruciate ligaments obtained from cadavers (age at death, 56-88 years old; mean, 74 years) were evaluated blindly. Their overall signal intensity and focal regions of increased signal were noted. The findings then were correlated with corresponding anatomic and histologic

sections.

RESULTS. The anterior cruciate ligament had greater signal intensity than did the posterior cruciate ligament in 15 of 18 knees and had comparable signal intensity in three knees. No histologic basis to explain the different MR appearances of the anterior and posterior cruciate ligaments was found. Rather, differences in macroscopic anatomy appeared to be the best explanation. In the posterior cruciate ligament, the fibers were parallel, whereas they were diverging and twisted in the anterior cruciate ligament, causing volume-averaging artifacts. A focal area of signal increase was found in 29 of the 38 ligaments. In 17 of these 29 ligaments, mucoid and/or eosinophilic degeneration was found that appeared to have caused the focal MR signal change. In nine ligaments without focally increased signal on the MR images, only one had histologic evidence of degeneration. The predictive value of an abnormal finding on MR was 59%, and the predictive value of a normal finding on MR was 89%. The relationship between increased signal on MR and histologic degeneration was statistically significant (x2 test, p = .0126). CONCLUSION.

The

differences

in the

MR

appearances

of the

anterior

and

posterior

cruciate ligaments probably are due to differences in gross architecture. Increased MR signal within the cruciate ligaments occurring in elderly patients often can be attributed to degenerative

159:357-360,

AJR

Received October 21 vision February 24, 1992. This

work

Affairs Corp.,

was

,

1 991 : accepted

supported

after

re-

in part

by Veterans

grant SA 306, a grant from Ciba-Geigy and a grant

from University

Hospital,

Zurich,

Veterans

Affairs

Switzerland. I

Department

of

Radiology,

Medical Center(1 14), 3350 La Jolla Village Dr., San Diego, CA 921 61 . Address reprint requests to D. Aesnick. 2 Present address: Department of Radiology, University

Hospital,

Raemistr.

100, CH-8091

0361 -803X/92/1

592-0357

Roentgen

Ray Society

cruciate

the posterior

1992

ligament

cruciate

(ACL)

consistently

has a higher

signal

intensity

(PCL) on MR images [1 ]. To find an explanation

ligament

for this finding, we compared MR images of cadaveric knee joints with the corresponding anatomic sections and histologic findings. In elderly patients, focally increased MR signal within the cruciate ligaments can occasionally be found. This same finding was visible in many of our cadaveric specimens and was also correlated with the corresponding histologic sections.

Materials Twenty

Department of Pathology, Veterans Medical Center, San Diego, CA 92161. 3

anterior

August

and Methods

Zurich,

Switzerland.

© American

The than

changes.

Affairs

time

freshly

of death;

thawed.

frozen

mean,

Ti-weighted,

acquired.

The

weighted

images,

technical

cadavenc

74

years)

knees were

(1 0 cadavers

examined

proton

density-weighted,

factors

for the

2000/20

for proton

MR

by

imaging

after

and T2-weighted

examination

density-weighted

from persons MR

included

images,

56-88 the

years old at the specimens

were

spin-echo images were 500-600/20 (TRITE) for Tiand 2000/70 for T2-weighted

358

HODLER

Downloaded from www.ajronline.org by 50.32.148.132 on 11/02/15 from IP address 50.32.148.132. Copyright ARRS. For personal use only; all rights reserved

images. A matrix of 256 x 256 or 192 x 256, 1 6-cm field of view, 3mm slice thickness, and 1 -mm interslice gap were used. All sequences were acquired in the sagittal and coronal planes [2]. The first 10 knee specimens were sectioned in the sagittal plane and the last 1 0, in the coronal plane. In one specimen, the ACL was completely absent and in another it was completely torn, with only a small residual ligament visible.

These

two

ligaments

were

excluded

from

further

Therefore, eight ACLs and 1 0 PCLs were examined plane, and 1 0 ACLs and 1 0 PCLs were examined

evaluation.

in the sagittal in the coronal

plane. 20 knees were evaluated blindly by one of knowledge of the histologic findings. The overall signal intensity of ACLs and PCLs on Ti - and T2-weighted images, the presence of focally increased signal intensity, and the internal pattern (e.g. , the depiction and the course of single fascicles) were The

MR

the authors

described.

images

of the

without

Those

Ti-

and

T2-weighted

largest part of the respective anatomic sections then were logic

examination

(fixation

images

that

contained

the

ligament were used for analysis. The photographed and prepared for histo-

in 1 0%

formalin,

embedding

in paraffin,

and staining with hematoxylin and eosin). The histologic examination was performed by an experienced pathologist who did not know the MR findings or whether either an ACL or a PCL was being examined. Criteria used for the histologic examination were the extent and degree (mild, moderate, or severe) of degenerative changes (e.g., eosinophilic

or mucoid

degeneration),

obvious

differences

in the fiber

thickness of the connective tissue, the cellularity (normal or increased) of the ligament, and the presence or absence of interposed fat within the ligament. Ligamentous degeneration was defined as the presence of eosinophilic and/or mucoid degeneration of any degree. The MR imaging

results

of the gross

then

anatomy

were

compared

with

and the histologic

those

derived

from

analysis

findings.

Results On Ti -weighted images, the signal intensity of the ACL was greater than that of the PCL in 1 3 knees, and the two ligaments had comparable signal intensity in five knees. On T2-weighted images, the ACL had more intense signal than

ET AL.

AJR:159,

August

1992

the PCL in 1 5 knees, and the two ligaments had comparable signal in three knees. Histologically, no specific finding correlated with these signal differences; fiber thickness and arrangement were identical in the ACL and PCL. Also, no differences were apparent in the size, form, and number of cell nuclei, with the exception of two PCLs and one ACL, in which the cellularity was slightly higher. This last finding was probably explained by accompanying degenerative changes in the ACL and in one of the two PCLs. The internal architecture of the cruciate ligaments was only partially apparent on MR images. The failure to visualize individual fascicles with MR imaging presumably was related in part to the scarcity of slightly hyperintense fat and loose connective tissue between adjacent fascicles. Also, the diameter of the fascicles (250 m to several millimeters) [3] was relatively small compared with a theoretical spatial resolution of 0.6 mm accomplished with the imaging parameters used in this study. However, in the distal fanned portion of the cruciate ligaments close to their osseous insertions, single fascicles were identified with significant frequency; these fascicles were more evident in the ACL than in the PCL on MR in 1 4 cases (seven sagittal and seven coronal), and they were equally apparent on MR in four cases (one sagittal and three coronal). More optimal MR visualization of individual fascicles corresponded to better identification of individual fascicles on anatomic inspection and/or to the presence of interfascicular fat on histologic examination; eight of 1 8 ACLs and only four of 20 PCLs showed interfascicular fat histologically. On MR images, the PCL appeared as a homogeneous structure with parallel, well-delineated borders (Fig. 1). Conversely, the ACL had a more complex appearance, with a less homogeneous signal intensity and less well defined borders (Fig. 2). This appearance was more evident on sagittal than on coronal images. Also, anatomically, the PCL consisted mainly of parallel fascicles, whereas the fascicles of the ACL were more twisted, especially the posterior ones. Corre-

Fig. 1.-Eosinophilic degeneration of posterior cruciate ligament (PCL) in right knee of a 74-year-old woman. A, Sagittal Ti-weighted (500/20) MR image shows borders of PCL are well defined. Normal hypointensity signal in middle and distal portions of ligament (arrows). B, Sagittal T2-weighted (2000/70) MR Image shows that signal changes appear less extensive than in A. C, Anatomic section shows fibers are parallel to sagittal plane.

is maintained

proximally,

with

increased

MR

AJR:159, August 1992

Fig. 2.-Eosinophilic

degeneration

OF

CRUCIATE

LIGAMENTS

OF

359

KNEE

of anterior

cruciate ligament (ACL) in right knee of an 88year-old man. A, Sagittal Ti-weighted (500/20) MR Image

shows sharply demarcated

anterior border and

irregular posterior border. Normal hypointensity of anterior portion of ligament is maintained

Downloaded from www.ajronline.org by 50.32.148.132 on 11/02/15 from IP address 50.32.148.132. Copyright ARRS. For personal use only; all rights reserved

proximally and distally, with increased intermediate B, Sagfttal

part (arrows). T2-weighted

(2000/70)

signal of MR image

shows less extensive signal than in A. Increased signal within ACL is not as intense as that of joint effusion, which is visible anterior to ACL (arrowheads). C, Anatomic section. Whereas anterior fibers are parallel to Imaging plane (arrowheads), pos-

F.-.

terior fibers have a less steep course and are more obliquely oriented with regard to sagittal plane (arrows). D, Photomicrograph ophilic degeneration

shows (arrows).

extensive

eosin-

-

more extensive

images

of increased on Ti -weighted

which

images

appeared

to be

than on T2-weighted

images, were found in 29 ligaments (1 4 ACLs and 15 PCLs). In 1 7 of these 29 ligaments, the MR findings could be explained by eosinophilic degeneration alone (six cases, Figs. 1 and 2), by degeneration that was both eosinophilic and mucoid (seven cases), or by mucoid degeneration alone (four cases) (Fig. 3). Twelve of the 29 ligaments with focal signal changes

did not show

histologic

ever. Of the remaining nine changes, one had histologic consisted of focal eosinophilic eight ligaments were normal value of normal MR findings value of abnormal MR findings statistically

significant

signs

of degeneration,

how-

ligaments with no focal signal evidence of degeneration that degeneration alone. The other histologically. The predictive was 89%, and the predictive was 59%. A x2 test showed a

relationship

between

Discussion

Some of the differences tween

course

in all 1 0 PCLs.

signal,

-

-

.

D

posterior border of the ACL was not well eight sagittal images, whereas the anterior delineated on four of eight sagittal images. and posterior borders of the PCL were well

on the sagittal

Focal regions

.4

B

C

delineated

.

,.

A

..

spondingly, the delineated on all border was well Both the anterior

.

-

histologic

evi-

the ACL

and

of the ACL

in signal

the

PCL

with

regard

intensity

that exist

are explained

be-

by the oblique

to the sagittal

plane.

The

diagnostic value of MR imaging can be improved by altering the imaging plane [4]; the patient’s legs either are allowed to

rotate

into a comfortable

or a sagittal

imaging

plane

position with

of slight external appropriate

rotation

anteromedial

to

posterolateral angulation is chosen on the basis of an axial image. Even when these modifications in MR protocol are used, the ACL has a different appearance than the PCL. Anatomic studies [3, 5] and the present indicate that the differences in appearance

correlative probably

study are re-

lated to differences in ligament architecture. Although the PCL fascicles have a parallel course, the anatomy of the ACL is more

complicated.

The ACL

has a constant

anteromedial

bundle, and the remainder of the ligament is more twisted. The amount of twisting depends on the degree of knee flexion [5]. Although investigators disagree on the anatomic teristics of the ACL, the presence of distinct functional

characbands

dence of degeneration and MR evidence of increased signal (p = .Oi 26). The common odds ratio was 1 1 .3 (95% confidence interval = 1 .i 8, 532), indicating that a ligament with focally increased signal on MR is 1 1 .3 times more likely to

[6] seems to be generally accepted [3]. Some of these bands are likely to be obliquely oriented, and therefore cause volume-averaging artifacts during the MR examination. Moreover, the ACL is slightly thinner at its midportion than the PCL

have

is (1 1 .1 vs 1 3.4 mm [5]), a fact that would probability of volume averaging on MR imaging

histologic

increased

degeneration

signal on MR.

than

a ligament

with

no focally

increase the of the ACL.

HODLER

360

ET AL.

AJR:159,

August

1992

....

Downloaded from www.ajronline.org by 50.32.148.132 on 11/02/15 from IP address 50.32.148.132. Copyright ARRS. For personal use only; all rights reserved

4)J

..

,

,-

,,.1

A

B

of the ACL have a considerably

distally,

and

their

average

anteroposterior

attachment

to the

diameter

tibial

divergent

course

plateau

has

an

of 30 mm. The fat inter-

posed between the diverging fascicles explains the increased visibility of single fascicles distally. No significant histologic differences between the ACL and PCL have been found either in this study or by other authors

[7, 8]. Both ligaments

consist

mainly of dense fibrous

tissue

and therefore should appear hypointense on all MR images. In about 60% ofcases with focal signal changes, degenerative findings similar to those described in the rotator cuff [9] and in the menisci of the knee [1 0] were found.

MR imaging had a high negative analysis of the cruciate ligaments, showed abnormal histologic findings

predictive value in the as only one ligament when MR images were

interpreted as normal, and the degenerative changes in this ligament were mild. Possible explanations for the low positive predictive value include autolysis, changes that occur during the thawing of the specimen before the MR examination (e.g., diffusion of tissue fluid into the ligament), or architectural alterations due to reduced muscle tension. It is feasible that

the positive predictive value of MR imaging in living patients would be improved if such artifacts could be eliminated. Moreover, the radiologist may have overestimated focally increased signal, mainly because of volume-averaging artifacts. Also, the results by the use of oblique

volume-averaging Another servative

evaluation

may be simulated cystic

form

presence clinical

importance

limited,

the relationship

abnormalities

between

may be the con-

of the histologic findings by the patholoto avoid false-positive diagnoses of eosindegeneration. Eosinophilic degeneration

by irregular

of mucoid

of gaps

is probably

influencing

and microscopic

gist, who wanted ophilic or mucoid

could have been improved images, eliminating in part

effects.

explanation

MR findings

probably sagittal

staining

degeneration

caused

with eosin,

and the

may be simulated

by the

by the sectioning

of degeneration

but degeneration

_

C

Fig. 3.-Mucoid degeneration of posterior cruciate ligament (PCL) in left knee of a 71-year-old man. A, Sagittal TI-weighted (500/20) MR image shows increased signal intensity, mainly in proximal portion B, Sagittal T2-weighted (2000/70) MR image shows a subtle increase in signal intensity in PCL C, Photomicrograph shows extensive mucoid degeneration (arrows).

The fascicles

,

process.

of the cruciate

The

ligaments

could cause MR abnor-

of PCL (arrows).

malities that simulate those of tears. Indeed, the focal regions of higher signal in a cruciate ligament with intact borders and

a normal

course

could lead to a false diagnosis

of a partial

tear.

We conclude and

PCL

that the different

cannot

MR appearances

be explained

histologically.

of the ACL Variations

macroscopic anatomy leading to volume-averaging appear to explain these signal differences best. patients, frequently

degenerative changes occur in the cruciate

that are visible ligaments.

in

artifacts In elderly

on MR images

ACKNOWLEDGMENT We thank

Sandra

J. Underhill

for preparing

the histologic

sections.

REFERENCES 1 . Reicher High

MA, Rauschning W, Gold RH, Bassett LW, Lufkin RB, Glenn W. resolution MRI of the knee joint: normal anatomy. AJR 1985;

145: 895-902 2. Hodler J, Berthiaume M-J, Schweitzer ME, Resnick D. Focal changes of the articular cartilage in the femorotibial joint: magnetic resonance (MA)anatomic correlation and blinded analysis (abstr). Radiology 1991; 181 (P): 301 3. Burks AT. Gross anatomy. In: Dale DM, Akeson WH, O’Connor JJ, eds. Knee ligaments. New York: Raven, 1990:59-76 4. Mink JH, Levy T, Crues JV. Tears of the anterior cruciate ligament and menisci

of

the

knee:

MR

imaging

evaluation.

Radiology

1988:

167:769-774 5. Girgis

FG, Marshall

JL, al Monajem

ARS.

The

cruciate

ligaments

of the

knee joint. Clln Orthop 1975:106:216-231 6. Amoczky SP, Warren RF. Anatomy of the cruciate ligaments. In: Feagin JA, ed. The crucial ligaments. New York: Churchill Livingstone, 1988: 179-1 95 7. Amiel D, Frank C, Harwood F, Fronek J, Akeson W. Tendons and ligaments:

a morphological

1:257-265 8. Yao L, Sherwood Comput

Med Imaging

and biochemical

JW. Ex-vivo Graphics

comparison.

J Orthop

Res 1984;

MR imaging of the cruciate

ligaments.

1991;1 5:27-30

9. Kjellin I, Ho C, Haghighi P, et al. MA alterations of the rotator cuff and anterior glenoid labrum: correlation with histopathology (abstr). Radiology 1990;1 77(P): 103

10. Stoller DW, Martin C, Crues JV, Kaplan L, Mink JH. Meniscal tears: pathologic correlation with MR imaging. Radiology 1987;163:731-735