Jeffrey

R. Crass,

MD

#{149} . G

Lucy

van

de

Vegte,

MD

#{149}

Linda

A. Harkavy,

MD

Tendon

Echogenicity:

Ex Vivo

Study

Recent publications discussing the echogenicity of normal tendon have described it variously as hyperechoic on hypoechoic. Since the echogenicity of tendon has been used to define normality and abnormality, certain knowledge of the normal echogenicity of tendon is crucial. Fresh tendon and muscle from beef hock was scanned with sectorand linear-array-transducer imaging at multiple angles and distances. The echogenicity of tendon was found to be very angle-dependent, a characteristic known as anisotropy. Scanned perpendicular to its long axis with a linear-array transducer, tendon was significantly more echogenic than muscle. With a change in angle, echogenicity of tendon decreased relative to that of muscle (the echogenicity of muscle remained the same), becoming isoechoic at angles of 2#{176}-7#{176}hypoeand choic at greater angles. Tendon studied with a sector transducer exhibited varying echogenicity. If echogenicity is used as a diagnostic criterion, the angle of the interrogating ultrasound beam must be very specifically defined.
Index
(US),

1.

Figures
cross
don don

1, 2. section.

(1) Sonogram, When normal

bright compared

with linear-array to the transducer,

with muscle on
tendon

transducer, of muscle tendon is hyperechoic

both sides
decreases

(M)

and

tendon

(T)

relative

to muscle;

in tentenwere

is very (T)
in

(2) Sonogram,
cross

with

sector

transducer
Echogenicity

angled
of

10#{176} from muscle

of it. Sector perpendicular,

compared

section.

images were similar. of muscle (M) and with that of muscle;

central
similar.

tendon

is now

darker

than

two

flanking

specimens.

Linear-array

images

study of tendons sound has received interest. Studies of the ment, and the Achilles
HE

with ultraincreasing patellar liga(calcaneal),

above papers brought forward, we undertook a series of experiments to study the echogenicity of tendon in a controlled, ex vivo laboratory setting.

terms:
tissue

Tendons,
characterization,

46.92

#{149} Ultrasound

46.1298

Radiology

1988:

167:499-501

From

the
Center,

Department
Level

of Radiology.
4, State University

Health
of

Sciences

New York at Stony Brook, Stony Brook, NY li794. From the 1987 RSNA annual meeting. Received November 25, 1987; revision requested December 30; revision received January 13, 1988; accepted January 21. Address reprint re-

quests
RSNA,

to J.R.C.
1988

biceps, and rotator (musculotendinous) cuff tendons have all been meported (1-5). The appearance of the normal tendon has been described variously as hypoechoic (6), of lowlevel echogenicity (5), and as hyperechoic compared with that of muscle (1-3). In a recent review of the topic, Fonnage claimed that the normal tendon is a hypemechoic structure and that the hypoechogenicity noted by various other authors was an artifact of the incident angle of the intemmogating ultrasound beam relative to the tendon (7). We were intrigued, however, by some of the illustrations in the Fomnage paper in which more of the tendon length exhibited artifactual hypoechogenicity than the hypemechogenicity that he claimed was normal. Because of the uncertainties the

MATERIALS
The tissue for this was fresh muscle mestic cattle. The

AND
series

METHODS
of experiments from dotendon

and tendon large flexor

found
was the hock wise an

in the
dissected

hock
from hock

was
the

used.
proximal

The

muscle
part of lonas

if the additional was

included it; otherpiece of beef with

gitudinally
The the beef tendon. The tissue

oriented
trimmed samples

fibers
to the were

was

obtained.
size in

same

placed

a Wa-

ter bath. Sonogmams with and linear-array transducers 128 instrument (Acuson, Calif.) were obtained and

5-MHz sector of a model Mountain View, from longitudinal 0#{176}

at angles curved

(perpendicular)
cross-sectional,

to 20#{176}. Longitudinal,

images

were obtained at distances ranging from 2 to 5 cm. Experiments were performed at least three times during independent imaging 499

Figure 3. Demonstration of effect of changes in angle on sonographic appearance of tendon relative to that of muscle when images are obtained along axis of tissue fibers. Sonograms with linear-array transducer show that muscle (a) is less echogenic than tendon (b) when parallel to transducer face and therefore normal to ultrasound beam. (c) With change
in angle, echogenicity of muscle did not change but that of tendon decreased markedly.

Figure 4. Sonogram, with sector transducem, of tendon along axis of its fibers shows very echogenic zone (E) where tendon is
perpendicular to a line drawn region. from the apex

of the sector
trasound beam

and

thus
in that

is perpendicular

to ul-

sessions. Fresh tissue series of experiments. corded on sheet film.

was used for each Images were me-

RESULTS Cross-sectional Imaging

Imaging muscle and tendon in cross section yielded similar data with both linear-array and sector transducers. Examination of the muscle at varying angles and distances showed little variation in the echogenicity of the muscle. There was, however, a marked change in the appearance of the tendon. When the transducer was perpendicular to the tendon, the tendon was significantly hyperechoic relative to muscle. With an angle of as little as 2#{176}-3#{176}, the striking hyperechogenicity of tendon was lost; the tendon became isoechoic with muscle. This persisted until 7#{176}-8#{176}, at which point tendon echogenicity became definitely less than that of muscle. The relative echogenicity did not vary significantly with distance (Figs. 1, 2).

looked homogeneous on sector images, tendon echogenicity varied within the sector. Consistently, that portion of the tendon that was penpendiculam to a radius of the sectors arc was brightly hypemechoic. The tendon on either side of this bright area, however, rapidly decreased in echogenicity as the angle increased (Fig. 4). Curved tendon-The curved-tendon experiments revealed results similar to those with the sector scannem. Both linear-array and sector transducers yielded similar results. Muscle appeared homogeneous, as in the earlier experiments. Tendon vanied in its echogenicity in longitudinal section (Fig. 5). That portion of the tendon parallel to the face of the linear-array transducer (or perpendicular to a radius of the sector) was hypemechoic. As the angle increased, echogenicity decreased.

DISCUSSION
Our results show quite clearly that the appearance of tendon varies both with the type of transducer and with the angle of the tendon to the interrogating ultrasound beam. The cause for this variation can probably be found in the histologic structure of tendon. This tissue is composed of longitudinally organized collagen bundles with a scanty cellular stmoma. The tissue is very homogeneous. These featunes make the sonognaphic appearance of tendon angle-dependent. The collagen bundles, because of their length and smooth surface, act as specular reflectors of the ultrasound beam. When the tendon is normal to the axis of the transducer, a maximum of sound energy is returned to the transducer. When an interface

yielding a specular reflection is tilted relative to the transducer, the reflected sound reaching the transducer decreases rapidly. The homogeneity of tendon tissue results in few nonspeculam reflectors. As the angle of incidence of the ultrasound beam causes a loss of speculam reflections, the paucity of nonspecular reflectors results in a change from hyperechogenicity (when the beam strikes perpendicularly) to hypoechogenicity (with increasing angles). Muscle has a more complex histologic organization. There are fewer of the internal long interfaces that yield specular reflections and more nonspecular reflectors, whose amplitudes has little angle dependence. The result is little change in muscle echogenicity with change of the angle of the transducer. The variation of ultrasound interaction with tissues of a fibmillan natune has been called acoustic fiber anisotropy. It was, to our knowledge,
first mentioned in the world litena-

Longitudinal
Linear

Imaging

array-Longitudinal imaging with the linear-array transducer yielded the same results as cross-sectional imaging. When the transducer face was parallel to the tissue long axis, tendon was much more echogenic than muscle. With a change of angle, the echogenicity of tendon relative to muscle decreased consistently; tendon became isoechoic to muscle at 5#{176} and hypoechoic at 10#{176} (Fig. 3). Sector. Experiments with the secton transducer yielded quite different results from those with the linear-amray transducer. Whereas muscle

tune by Heuter in 1948 (8) and in the English literature by Dussik et al. in 1958 (9). These workers studied the ultrasound transmission chanactemistics of tissue in vitro and in vivo. The findings of our studies are consistent with the real-time sonogmaphic expression of the anisotnopic transmission characteristics these pioneers defined. The variation in echogenicity found in our studies is of clinical impontance. With a linear-array transducem, straight tendons such as the Achilles tendon and patellar ligament can be studied with the tendons parallel to the face of the transducem. They will appear hypemechoic.

500

#{149}

Radiology

May

1988

References
1. Fornage BD.

PM.

Sonography

2. 3.

preliminary 143:179-183. Fomnage BD. Achilles tendon: US examination. Radiology 1986; 159:759-764. Middleton WD, Remus WR, Totty WG, Melson CL, Murphy WA. Ultrasound of the biceps tendon apparatus. Radiology
1985; 157:211-216.

MD, Touche DH, Segal of the patellar tendon: observations. AJR 1984;

Rifkin

4.

5.

Figure
show thus

Crass JR. Craig EV, Thompson RC, Feinberg SB. Ultrasonography of the rotator cuff: surgical correlation. JCU 1984; 12:487-492. Middleton WD, Edelstein C, Remus WR, Melson CL, Murphy WA. Ultrasound of
the rotator cuff: technique and normal

5.

Sonograms,

with
zone to the

linear-array

transducer,

of curved
it is parallel

muscle
to the face

(a) and
of the

tendon
transducer

(b)
and 6.

very echogenic perpendicular

in the tendon where ultrasound beam.

anatomy.
552.

J Ultrasound

Med

1984; 3:549-

Khaleghian R, Tonkin U, De Geus JJ, Lee JPK. Ultrasonic examination of the flexor tendons of the fingers. JCU 1984; 12:547552.

Examination of curved tendons, howeven, will yield different results. Tendons such as the flexoms of the finger and the rotator cuff cannot be studied parallel to the face of a tmansducen-they are normally curved. Tendons such as these will vary in their echogenicity. Examinations done with a sector transducer deserve special notice. In many neal-time instruments, sector scanners are used as the high-resolution transducers. Sector scans of normal tendon will often me-

veal a bright spot in the tendon (Fig. 5); if echogenicity is a criterion for abnormality, this bright spot must be interpreted with great caution. In summary, the echogenicity of normal tendon varies with the angle of incidence of the ultrasound beam. The appearance of tendon is also quite different with sector scanners than with linear arrays. If echogenicity is to be used as a criterion of nonmality or abnormality, the type of transducer and the scanning technique must be rigidly defined. U

7.

Fornage BD. don: a pitfall.

22.

The

hypoechoic

normal

ten6:19-

J Ultrasound
Messung
tierischen

Med

1987;

8.

Heuter
sorption

TF.
in

der der

ultraschallabund ihre

geweben

abhangigkeit
1948;

von

frequenz.

Nature M, Sear tissues with

35:285-287.

9.

Dussik KT, Fritch RS. Measurements

ultrasound. 165. Am

DJ, Kyriazidou of articular

J Phys Med

1958;

37:160-

Volume

167

Number

Radiology

#{149} 501