Mammography screening programs have been shown to be effective in decreasing breast cancer mortality through the detection and treatment of small breast cancers. This article addresses the technical factors needed to optimize visualization of small breast cancers and the subtle mammographic features of breast carcinoma, which might be difficult to detect.
is the Clinical Director of the South East London Breast
Screening Programme and
is a Clinical Fellow in Breast Imaging at King's College
Hospital, London, England.
Evaluation of the shoulder continues to be the most common
sonographic examination of the musculoskeletal system at most
institutions. Several advantages of sonography over alternative
imaging methods include availability, portability, and low cost.
Most patients prefer sonography over magnetic resonance (MR)
imaging, as it requires less time and it is more comfortable. This
is especially true in the older patient, who is more likely to
undergo evaluation for rotator cuff abnormality. Although learning
this imaging method initially takes time, thorough knowledge of
anatomy combined with standardized examination technique shortens
the learning process. Once skilled in this technique, a complete
shoulder sonographic examination can be easily completed in <5
minutes. Technologists can also be successfully trained in shoulder
sonography. With the use of technologists acquiring images and
documenting pathology, the sonographic images can be reviewed,
similar to other sonographic studies, in a matter of minutes.
From the perspective of a radiologist who is trained in both
sonography and MR imaging of the shoulder, shoulder sonography is
often preferred primarily because of the improved resolution of
sonography over MR imaging. For example, a tendon abnormality as
small as several millimeters can be seen well, and the critical
distinction between articular surface and isolated greater
tuberosity extension can be made easily. This is important, as the
tendon abnormality in the latter situation cannot be visualized at
arthroscopy. From an orthopedic surgeon's perspective, accuracy of
rotator cuff sonography can meet, and even exceed, that of MR
It is important, however, that the strengths and limitations of
shoulder sonography are understood. For example, sonographic
evaluation of the rotator cuff is very effective.
For evaluation of the glenoid labrum and ligamentous structures, MR
arthrography is indicated. The proper utilization of sonography
relies on patient selection. At our institution, a patient under
the age of 35 years will typically have MR arthrography to evaluate
cartilaginous structures, while a patient over the age of 40 years
will more likely have sonography to evaluate the rotator cuff.
Although most current ultrasound machines can perform
musculoskeletal sonography, proper transducer selection is critical
to success. For sonography of the shoulder, a linear transducer is
preferred. More importantly, a frequency of >=10 MHz is
recommended. There is a trade-off, as high frequency transducers
offer the highest resolution, but at the expense of depth
penetration. At our institution, a 12-MHz linear transducer with a
5-cm footprint is used for the rotator cuff. However, a 10-MHz or
7-MHz transducer may be used in patients with a large body habitus.
All images used in this article were obtained on a dedicated
musculoskeletal ultrasound unit (Model HDI 5000, Advanced
Technology Laboratories, Bothell, WA).
To ensure complete evaluation of the shoulder, specific patient
positions and transducer placements are used to optimize
visualization of each structure.
To evaluate the supraspinatus, it is important that the patient
internally rotates and hyperextends or adducts the arm so that the
tendon is pulled down from beneath the acromion for visualization.
This is accomplished by asking the patient to place the dorsum of
the hand in the small of the back, or alternatively the palm of the
hand at the ipsilateral hip. This latter position allows
visualization of the intra-articular portion of the long head of
the biceps brachii tendon. This is an important landmark when
assessing for and describing the location of supraspinatus tendon
abnormalities. Identification of the biceps tendon also ensures
that the full anterior extent of the supraspinatus tendon has been
evaluated, as this area is commonly torn. The subscapularis is
optimally visualized with the arm in external rotation to eliminate
anisotropy. The infraspinatus and teres minor are identified from a
posterior approach, while the long head of the biceps brachii
tendon is identified in the bicipital groove from an anterior
approach with the hand supinated in front of the patient.
Normal sonographic anatomy
Normal tendons appear hyperechoic with a fibrillar echotexture
when imaged perpendicular to the ultrasound beam (Figure 1).
It is important to recognize that tendon fibers oblique to the
ultrasound beam will appear artifactually hypoechoic, which is
This should not be confused with a tendon tear, which may also be
hypoechoic. Only that segment of tendon that is perpendicular to
the ultrasound beam should be evaluated for tendon abnormality.
Normal muscle appears relatively hypoechoic at sonography, with
interspersed hyperechoic fibroadipose septa (Figure 1). Bone cortex
is highly reflective and will appear hyperechoic with posterior
acoustic shadowing or reverberation artifact if a flat segment of
cortex is imaged perpendicular to the ultrasound beam. Hyaline
cartilage appears hypo-echoic, while fibrocartilage, such as the
glenoid labrum, appears hyperechoic. The subacromial-subdeltoid
bursa is located between the supraspinatus tendon and deltoid
muscle and is outlined by hyperechoic peribursal fat and bursal
walls. This bursa continues beyond the greater tuberosity, which
allows differentiation from the adjacent rotator cuff inserting on
The strength of shoulder sonography lies in its ability to
accurately evaluate the rotator cuff.
It has been shown that sonography can perform at least equal to MR
imaging in this application.
The following discussion will review the sonographic appearances of
various rotator cuff pathologies, and will approach this from a
differential diagnosis perspective. Abnormalities of the
supraspinatus tendon are emphasized, as this tendon is most
Focal hypoechoic or
anechoic tendon abnormality
When a focal hypoechoic or anechoic abnormality of a tendon is
identified, one must first differentiate tendinosis from a tendon
tear. Tendinosis, or tendinopathy, is tendon degeneration, and this
term is used instead of tendonitis, as active, inflammatory cells
In this differentiation, tendon tears are more likely well-defined
(Figure 2), while tendinosis is more likely ill-defined (Figure 3).
In addition, tendon tears are more likely quite hypoechoic or even
anechoic (Figure 2), while tendinosis is more likely minimally
hypoechoic (Figure 3). Both processes, however, may coexist in the
continuum of a diseased tendon.
One indirect sign that is very helpful in differentiating tendon
tear from tendinosis is cortical irregularity of the greater
tuberosity at the supraspinatus tendon insertion (Figure 4). When a
focal and well-defined hypoechoic or anechoic supraspinatus tendon
abnormality is immediately adjacent to cortical ir-regularity of
the greater tuberosity, this likely represents a tendon tear.
Greater tuberosity cortical irregularity is seen with
full-thickness and partial-thickness supraspinatus tendon tears.
One should note, however, that cortical irregularity of the greater
tuberosity at the infraspinatus tendon insertion, and cortical
irregularity of the lesser tuberosity at the subscapularis tendon
insertion, does not carry the same significance and is commonly
seen even with a normal rotator cuff.
In the presence of cortical irregularity, a focal and
well-defined hypoechoic or anechoic supraspinatus tendon
abnormality is considered a tear--unless it contacts only the
greater tuberosity surface. In this case, it may still represent
tendon fiber disruption but does not extend to the articular or
bursal surfaces and cannot be visualized at arthroscopy or
is used in this situation, as it could either represent
intrasubstance tear or severe tendinosis (Figure 5). The
longitudinal view of the supraspinatus is most important in
determining if the tendon abnormality extends to the articular or
greater tuberosity surface.
Articular side abnormality
If the focal and well-defined hypo-echoic or anechoic
supraspinatus tendon abnormality contacts only the articular
surface, this indicates an articular side partial-thickness tear
(Figure 4). Many of these tears will demonstrate a mixed echogenic
appearance, with the hyper-echoic torn tendon fibers identified
within the hypoechoic or anechoic tendon gap.
If it is unclear if a tendon abnormality extends to the articular
surface, the presence of a hyperechoic reflective interface between
the tendon abnormality and hyaline articular cartilage of the
humerus (the cartilage interface sign) is helpful, as this
indicates articular extension (Figure 2).
If the normally convex superior surface of the supraspinatus
tendon is flattened or concave from tendon thinning or absence,
this indicates tendon fiber loss and may be seen with bursal side
partial-thickness tendon tear (Figure 6) or full-thickness tendon
tear (Figures 7 and 8).
Extension of the tendon abnormality from the bursal to articular
surface indicates full-thickness tear, while isolated bursal
surface extension indicates bursal side partial-thickness tendon
tear. This finding can also be applied to the other tendons of the
rotator cuff. Generally speaking, as the tendon tear becomes
larger, the degree of superficial tendon concavity or dipping the
deltoid muscle into the tendon gap becomes more pronounced. With an
acute tear, this tendon gap more likely fills with anechoic fluid
(Figure 2); with chronic tears, fluid is lacking and often only
tendon thinning or tapering is seen.
When there is complete nonvisualization of a tendon, this
indicates a massive full-thickness tear (Figure 8). In this
situation, the deltoid muscle is typically lying on the proximal
humerus in place of the torn rotator cuff.
The torn and retracted tendon stump is often located beneath the
acromion and not identified with sonography. It is important to
determine the extent of a full-thickness tear. On transverse
images, the intra-articular portion of the long head of the biceps
brachii tendon is used as a reference point for this purpose
(Figure 7B). A tendon tear that extends more than 2 cm to 2.5 cm
posterior to the biceps tendon indicates infraspinatus tendon
involvement. The presence of infraspinatus tendon atrophy should
also be noted, as this is a negative prognostic indicator.
If a rotator cuff tendon is heterogeneous in echotexture and is
diffusely hypoechoic without tendon thinning, this is
characteristic of diffuse tendinosis (Figure 9).
The involved tendon may be swollen with this process, which is a
helpful sign. Cortical irregularity of the greater tuberosity at
the supraspinatus tendon is typically absent. The patient may
experience significant pain with this condition during the
Focal hyperechoic tendon abnormality
If a focal hyperechoic tendon abnormality at the articular
surface is surrounded by a hypoechoic area, this indicates an
articular side partial-thickness tear.
If the hyperechoic abnormality is larger and displays posterior
acoustic shadowing, this is characteristic of calcium
hydroxyapatite deposition, as seen in calcific tendonitis (Figure
Although most of these deposits demonstrate shadowing, 7% display
no posterior acoustic shadowing.
If this situation is suspected, anisotropy is helpful in detection.
By imaging the involved tendon oblique to the ultrasound beam, the
tendon becomes hypoechoic from anisotropy while the calcium deposit
becomes more conspicuous, as it remains hyperechoic. If the calcium
deposit incites inflammation, then the term
is appropriate. Increased flow may be seen on color and power
Doppler imaging. Sonography has been proven effective in aspiration
and lavage of tendon calcium deposits.
Other shoulder pathology
An increase in the volume of joint fluid can be visualized in
the various joint recesses of the shoulder, including the long head
of the biceps brachii tendon sheath (Figure 11), the subscapularis
recess, and the posterior joint recess. Joint effusion
visualization around the biceps tendon is common. This should be
differentiated from a true biceps tenosynovitis. A focal collection
of fluid associated with pain and hyperemia is characteristic of
this latter condition. The presence of an isolated joint effusion
is not specific for one type of pathology, and has low sensitivity
and specificity for the diagnosis of rotator cuff tear.
Joint fluid may range from anechoic simple fluid to hyperechoic
Subacromial-subdeltoid bursal fluid
Fluid within the subacromial-subdeltoid bursa typically appears
anechoic or hypoechoic, surrounded by hyperechoic peribursal fat
and bursal walls (Figure 12). Bursal wall distention <2 mm is
Abnormal quantities of bursal fluid may be seen with impingement,
as well as full-thickness and bursal side partial-thickness tendon
tears. The presence of subacromial-subdeltoid bursal fluid and
joint effusion has a high positive predictive value for the
diagnosis of rotator cuff tear.
Bursal fluid may range from anechoic reactive fluid to hyperechoic
complicated fluid. It is important not to mistake hyperechoic
synovium within the subacromial-subdeltoid bursa as rotator cuff.
Extension of this tissue beyond the greater tuberosity indicates
bursal thickening as the etiology (Figure 5).
Biceps tendon subluxation and tear
The long head of the biceps brachii tendon is easily assessed
with sonography. Its characteristic location within the bicipital
groove of the proximal humerus aids in its identification. Absence
of tendon within the bicipital groove may indicate either
full-thickness biceps tendon tear (Figure 13) or biceps tendon
subluxation or dislocation (Figure 14).
In the setting of a complete tendon tear, the distal retracted
tendon stump can be visualized more distally, which confirms this
diagnosis (Figure 13B). The biceps tendon may subluxate over the
lesser tuberosity (Figure 14A), or dislocate into a subscapularis
tendon tear and into the glenohumeral joint (Figure 14B). Abnormal
medial position of the biceps tendon may indicate coracohumeral and
superior glenohumeral ligament abnormality. With an empty bicipital
groove, it is important to scan distally for a retracted tendon,
and medially within the joint to differentiate tear from
Glenoid labrum tear and cyst
MR arthrography is considered the imaging study of choice in
evaluation of the glenoid labrum. However, the labrum can be
identified with sonography, especially the posterior aspect.
Evaluation of the posterior labrum can occur simultaneously during
evaluation of the infraspinatus, although a 7-MHz transducer may be
needed to optimally delineate the hyperechoic fibrocartilage
labrum. When a well-defined hypoechoic or anechoic cleft is seen, a
labral tear can be suggested. The presence of a hypoechoic or
anechoic glenoid labral cyst adjacent to the labrum is more
evidence that the labrum is torn (Figure 15).
Because symptoms related to a glenoid labral cyst may simulate that
of rotator cuff pathology, it is important to consider this
diagnosis, especially in the setting of infraspinatus and possible
supra-spinatus muscle atrophy. Sonography has been used to guide
cyst aspiration about the shoulder.
Because of the limitation of sonography in evaluation of the
labrum, especially at the biceps anchor, any suspected labral
abnormality is typically followed by MR arthrography for
Greater tuberosity fracture
Nondisplaced greater tuberosity fractures of the proximal
humerus are often overlooked on radiographs due to in-
adequate patient positioning, suboptimal radiographic technique,
and subtlety of the fracture line. In the setting of a reported
negative radiograph, a patient with continued shoulder pain after
trauma may be evaluated with sonography to exclude
rotator cuff tear. At sonography, a greater tuberosity fracture
will appear as dis-ruption of the echogenic cortex with step-off
deformity (Figure 16).
It is this step-off deformity that helps differentiate the true
fracture from the greater tuberosity cortical irregularity that is
seen with rotator cuff tears. With a greater tuberosity fracture,
point tenderness is present with transducer pressure and the
rotator cuff is often normal.
Sonography has been proven effective in the evaluation of the
rotator cuff. Thorough knowledge of anatomy, a standardized
technique, and familiarity with pathology as seen on sonography are
important for success. The advantages of sonography include
availability, low cost, patient comfort, and superior image