Keywords
rotator cuff - tendinopathy - tear - postoperative rotator cuff
The rotator cuff is a dynamic stabilizer of the shoulder, whereas the labrum, joint
capsule, glenohumeral ligaments, and osseous structures are static stabilizers.[1] Loss of integrity of the rotator cuff leads to progressive instability and eventual
degenerative changes at the glenohumeral joint, referred to as rotator cuff arthropathy.
Rotator cuff arthropathy is a primary indication for reverse total shoulder arthroplasty.[1]
Theories abound regarding the pathogenesis of rotator cuff injury, with extrinsic
and intrinsic impingement most commonly described.[2]
[3] In extrinsic impingement, rotator cuff tendinopathy and tearing is thought to result
from subacromial impingement due to structural abnormalities of the coracoacromial
arch, such as acromioclavicular (AC) joint osteophytes, enthesophytes, or abnormal
acromial configuration (e.g., hooked acromion or os acromiale).[3] Subcoracoid impingement is another form of extrinsic impingement that may result
from a narrowed coracohumeral interval, predisposing to tears of the subscapularis
tendon.[3]
Internal impingement frequently involves entrapment of the posterosuperior rotator
cuff tendons between the humeral head and posterior glenoid during abduction and external
rotation, eventually leading to tendon degeneration. Other theories suggest that tendons
simply undergo intrinsic degeneration due to repetitive use, causing destabilization
of the joint, subluxation of the humeral head, and potential for further injury.[2]
This article reviews the normal anatomy of the rotator cuff and illustrates distinct
types of rotator cuff tears. Key findings that should be included in the radiology
report for rotator cuff tears are discussed, highlighting features that may change
management. Finally, the various surgical techniques for rotator cuff repair are described
with example of cases to illustrate expected postoperative findings and complications.
Rotator Cuff Anatomy
The rotator cuff consists of the supraspinatus, infraspinatus, subscapularis, and
teres minor tendon–-muscle units. The supraspinatus tendon has an approximate width
of 23 mm, inserting on the superior and middle facet of the greater tuberosity.[4] The supraspinatus is primarily involved in shoulder abduction from 0 to 15 degrees,
acting in conjunction with the deltoid to abduct the arm beyond 15 degrees. The infraspinatus
tendon has a similar width and inserts on the middle facet with the anterior-most
fibers interdigitating with the posterior supraspinatus tendon. The infraspinatus
externally rotates the shoulder in conjunction with the teres minor that inserts on
the inferior facet. The subscapularis is a strong adductor and internal rotator composed
of four to six tendon slips inserting on the lesser tuberosity with a direct muscular
attachment at the inferior third; some fibers of the subscapularis cross the bicipital
groove to insert on the lesser tuberosity, forming the transverse humeral ligament.[3]
[5]
The subacromial-subdeltoid bursa lies superficial to the rotator cuff tendons and
deep to the acromion and deltoid. The subacromial-subdeltoid bursa can become distended
with fluid in the setting of a full-thickness rotator cuff tear or due to subacromial-subdeltoid
bursitis.
Imaging the Rotator Cuff: Pathology and Pitfalls
Imaging the Rotator Cuff: Pathology and Pitfalls
Radiographs
As described in the American College of Radiology Appropriateness Criteria, radiographs
are an appropriate first-line investigation for the assessment of rotator cuff injury
and demonstrate ancillary findings of rotator cuff disease.[6] Typical views obtained include anteroposterior (AP) external rotation, internal
rotation, and a scapular Y view. Radiographic findings of rotator cuff tear include
superior migration of the humeral head and narrowed acromiohumeral interval, < 6 mm,
best seen on the active abduction/AP external rotation view.[7] Osseous irregularity and cystic changes at the greater tuberosity and lesser tuberosity
are also an indicator of more chronic rotator cuff pathology.[7] The modified AP views show subacromial enthesophytes or inferiorly oriented osteophytes
from the AC joint, and a subscapular Y view is useful to assess acromial morphology.
Radiographs are also helpful to exclude alternative sources of pain, such as fracture
or osteoarthritis.
Ultrasonography
Ultrasonography (US) is a good first-line test for the assessment of rotator cuff
pathology due to its widespread accessibility, low cost, and superior spatial resolution.[6] A linear 12–15 Hz transducer ensures high spatial resolution without sacrificing
depth penetration, although lower frequency transducers may be required for larger
patients.[5] On US, the normal rotator cuff tendons are hyperechoic with fibrillar architecture.
The tendons should be evaluated in both the transverse and longitudinal imaging planes.
Dynamic assessment can also be performed with abduction of the shoulder to assess
for subacromial impingement of the supraspinatus tendon.
Magnetic resonance imaging (MRI) and US have similar sensitivities and specificities
for detecting rotator cuff disease in the native rotator cuff[4] ([Fig. 1]). US accuracy depends on technique, but in the hands of an experienced operator,
it approaches 100% for detection of a full-thickness tear and 91% for a partial-thickness
tear.[5]
Fig. 1 (a) Longitudinal ultrasonographic image of the left shoulder reveals an intermediate-grade
partial articular-sided tear at the supraspinatus footprint (arrow). (b) Corresponding coronal T2 fat-sat sequence confirms these findings.
There are several scanning pitfalls to be aware of when performing and interpreting
shoulder US. Anisotropy occurs when the normal hyperechoic tendon appears artifactually
hypoechoic when the tendon is not perpendicular to the sound beam,[5] most common at the supraspinatus footprint. Alternating hypoechoic bands can also
occur at the supraspinatus-infraspinatus junction or in the subscapularis tendon in
the short axis due to anisotropy.[5] Care should be taken to evaluate the full anteroposterior width of the greater tuberosity
to avoid missing small anterior supraspinatus tears near the rotator interval.[4]
Magnetic Resonance Imaging
MRI has higher interobserver reliability in determining tear size and degree of retraction
compared with US.[6] A meta-analysis in 2020 suggested that 1.5- or 3-T magnetic resonance arthrography
(MRA) has the highest diagnostic value to detect partial articular-sided and full-thickness
rotator cuff tears, followed by conventional non-MRA 3-T imaging; however, the marginal
increase in sensitivity with MRA may be offset by the added cost and resources required.[6]
[8]
The tendons of the rotator cuff should appear low signal on short and long TE sequences
with smooth margins and no areas of intervening high signal. Similar appearances are
expected on direct MRA studies. The supraspinatus and infraspinatus tendons are best
visualized on coronal and sagittal oblique images, and the subscapularis and teres
minor tendons on axial and sagittal oblique images.[3]
Tears are normally fluid signal intensity on T2-weighted images, although 10% of tears
can be low signal due to scarring or fibrosis. In these cases, ancillary signs, like
tendon retraction or muscle atrophy, can point toward the presence of a tear; these
types of tears are better seen on MRA.
Care should be taken not to mistake artifacts such as magic angle or volume averaging
for tendinopathy. Magic angle occurs on short TE sequences (proton-density [PD], T1-weighted
imaging) when the tendon is oriented 54.74 degrees from the main magnetic field, causing
intermediate to hyperintense signal in the tendon. Long TE sequences (T2-weighted
imaging) can help differentiate between true tendinopathy versus artifact, with the
tendon showing normal low signal in the setting of a magic angle artifact.
Computed Tomography Arthrography
Computed tomography (CT) arthrography is a reasonable alternative test in patients
who have a contraindication to MRI or when US is not available. Two studies showed
no statistical difference in the detection of rotator cuff lesions by CT arthrography
or MRA when compared against arthroscopy as the gold standard.[9]
[10] Full-thickness tears are easily identified as full-thickness gaps in the tendon
substance with extension of contrast into the subacromial-subdeltoid bursa. The articular
cartilage and osteoarthritic changes are also well delineated on CT. One of the drawbacks
of CT arthrography is the poor delineation of bursal and interstitial tears,[6] as well as slightly reduced sensitivity in the detection of infraspinatus and subscapularis
tears.
Tendinopathy
Tendinopathy, or tendinosis, refers to degeneration of the rotator cuff tendons and
is frequently a precursor to rotator cuff tears. On US, the tendon appears hypoechoic
and thickened in the setting of tendinopathy, often visually graded as mild, moderate,
or severe. On MRI, tendons are thickened and show intermediate signal on both short
(T1, PD-weighted imaging) and long (T2-weighted imaging) TE sequences.
Calcific tendinopathy results from deposition of hydroxyapatite crystals in the rotator
cuff tendons and is one of the most frequent causes of shoulder pain, with an estimated
prevalence of 2.7 to 10.3%.[3]
[11] Patients may present with shoulder pain and reduced range of motion in the chronic
setting; however, acute calcific tendinopathy can cause acute shoulder pain and may
be mistaken for infection due to the overlapping clinical presentation. Asymptomatic
calcific deposits are common, occurring in up to 8% of cases in one study[3]
[12]; middle-aged women 30 to 60 years old and those with calcifications measuring > 1.5 cm
are more likely to be symptomatic.[12] The distal supraspinatus tendon is most frequently involved in calcific tendinitis,
although the other tendons can also be affected.[3] Hydroxyapatite crystals can migrate into the bursa (calcific bursitis), musculature,
and bones.[3]
Typical US features of calcific tendinopathy include punctate or coarse echogenic
foci within the substance of the tendon with variable posterior acoustic shadowing.[6] On MRI, these calcifications appear as globular hypointense foci within the tendon
or surrounding structures. In the acute phase, there may be markedly increased signal
or edema around the calcification on fluid-sensitive sequences, reflecting inflammation.
Patterns of Injury
The prevalence of rotator cuff tears in the general population is ∼ 20% and increases
with age.[6]
[13] Chronic rotator cuff tears are most common and result from intrinsic degeneration
of the rotator cuff tendons from long-standing overuse, with or without impingement.
The typical pattern of progression is tendinopathy, followed by partial-thickness
tears and eventual full-thickness tears if left untreated.
Acute tears of the rotator cuff are less frequent, with a reported incidence of up
to 8%. They are often the result of a high-energy mechanism and tend to be large full-thickness
tears. Acute tears of the subscapularis are most common in the setting of anterior
shoulder dislocation, although any combination of tendons can be involved, depending
on the mechanism of injury.[14] Avulsion of the rotator cuff tendons from the greater tuberosity has also been reported
in the acute setting.
Supraspinatus tendon tears frequently occur at the anterior margin.[3] These tears can propagate posteriorly to the infraspinatus tendon or anteroinferiorly
through the rotator interval to involve the coracohumeral ligament and superior subscapularis
tendon.[3]
Infraspinatus tears often result from posterior extension of supraspinatus tendon
tears, although isolated tears can be seen in the setting of posterosuperior (internal)
impingement in overhead throwing athletes.[3]
[15] In posterosuperior impingement, the rotator cuff becomes entrapped between the greater
tuberosity and posterosuperior glenoid and labrum during abduction and external rotation.
Over time, this can lead to partial-thickness articular-sided tears of the posterior
supraspinatus and anterior infraspinatus tendons.[15] Teres minor tendon tears are rare and usually associated with infraspinatus tears.[3]
Although subscapularis tears were previously thought to be uncommon, the reported
incidence has increased with advanced arthroscopic techniques.[16] Subscapularis tears may result from anterior propagation of a supraspinatus tear,
or they may occur in isolation in the setting of subcoracoid impingement or anterior
shoulder dislocation.[17]
Rotator Cuff Tears: What to Include in Your Report
Rotator Cuff Tears: What to Include in Your Report
Important imaging features to describe in the radiology report for rotator cuff tears
include location, size, and thickness (partial or full thickness), tendon retraction,
tendon quality (delamination), and degree of muscle atrophy and fatty infiltration.
[Table S1] outlines a standardized approach to reporting cases of rotator cuff tear.
Full-thickness rotator cuff tears involve the entire tendon thickness, manifesting
as high T2/PD signal extending from the articular to bursal surface ([Fig. 2c, d]). Conversely, partial-thickness rotator cuff tears only involve a portion of the
tendon, occurring at the articular side, bursal side, or within the tendon substance
(referred to as intrasubstance or interstitial tears) ([Figs. 1] and [2a, b]). Several acronyms describe the subtypes of partial-thickness rotator cuff tears,
many of which originated in the arthroscopic literature (summarized in [Table S2], with accompanying examples in [Fig. S1] and [Fig. S2]). To prevent confusion, we encourage a standard description of tears (e.g., partial-thickness
articular-sided tear with intrasubstance delamination), rather than the use of acronyms.
Fig. 2 (a) Coronal T2 fat-sat sequences of the right shoulder showing intermediate-grade partial
articular-sided tear, (b) high-grade partial bursal-sided tear, (c) full-thickness tear without retraction, and (d) full-thickness tear with retraction of the supraspinatus tendon (solid arrows). There is intrasubstance delamination
of the partial tears shown in (a) and (b) (hashed arrows), with differential retraction
of the articular-sided fibers in (a). In the high-grade partial bursal-sided tear
(b), there are a few intact articular-sided fibers compared with the full-thickness
tears (c, d) where there are no intact articular-sided fibers. The tendon is medially
retracted to the superior glenoid in (d).
Grading of partial-thickness tears is based on the thickness of the tendon involved:
low grade (< 25%), intermediate grade (25–50%), and high grade (> 50%). Partial tears
at the supraspinatus or infraspinatus footprint and intrasubstance tears are important
to identify because these are often occult at arthroscopy ([Fig. S1]).[6] Care should be taken not to mistake the rotator cable, a normal thickened hypointense
band of signal frequently seen on the undersurface of the supraspinatus, as a partial-thickness
articular-sided tear.[18]
Tear size should be measured in both the anteroposterior and mediolateral dimensions.
Full-thickness tears can be more broadly classified as small (< 1 cm), medium (1–3 cm),
large (3–5 cm), or massive (> 5 cm) based on the AP dimension, as per the DeOrio and
Cofield classification.[3]
[19] The degree of medial retraction of the tendon stump can be measured (e.g., retracted
2 cm relative to the greater tuberosity) or described relative to surrounding structures
(e.g., at the superior humeral head, at the glenoid margin) ([Fig. 2d]).
Delamination refers to separation of the tendon layers in the setting of an intrasubstance
partial- or full-thickness tear. Delamination may result in differential retraction
of the articular-and bursal-sided fibers, with the articular-sided fibers often demonstrating
greater retraction ([Fig. 2b]). Delaminating tears should be noted in the report because studies suggest these
patients are more likely to fail conservative management.[20] The presence of delamination is also a negative prognostic factor for surgical repair.
Fluid dissecting into the tendon substance may undergo synovialization that reduces
tendon quality and healing potential.[20]
[21]
Delaminating tears can also propagate along the myotendinous unit and form small cysts
within the muscle belly, referred to as sentinel cysts[3] ([Fig. 3]). Studies evaluating the clinical significance of sentinel cysts have found an association
with both partial- and full-thickness tears, suggesting this is a helpful ancillary
feature for identification of small occult tears that may otherwise be missed.[22]
Fig. 3 (a) Coronal T2 fat-sat and (b) sagittal T2 fat-sat sequences of the right shoulder show a lobulated so-called sentinel
cyst at the infraspinatus myotendinous junction (a, solid arrow). This was tracking
from a pinhole low-grade partial articular-sided tear of the anterior infraspinatus
(a, b, hashed arrows).
Muscle atrophy, an important predictive factor for anatomical and functional outcome
after surgical repair, is associated with increased retear rates.[23] The degree of muscle atrophy can be quantified with the occupational ratio using
the method of Thomazeau et al[24] ([Fig. 4]). The cross-sectional area of the supraspinatus fossa is calculated on the T1-weighted
sagittal oblique Y view where the scapular spine and body intersect, with the undersurface
of the trapezius and distal clavicle delineating the superior margin of the supraspinatus
fossa.[23] This is compared with the cross-sectional area of the supraspinatus muscle, calculated
with the best fit line drawn around the outer margin of the muscle. The occupational
ratio (R) may be graded as follows: grade I, R ≥ 0.6 (normal or slightly atrophied);
grade II, R ≥ 0.6 to 0.4 (moderate atrophy); and grade III, ≤ 0.4 (severe atrophy),
with a smaller R indicating more advanced atrophic changes.[23] The tangent sign, also measured on the Y view, is another method to evaluate for
the presence or absence of atrophy. It is considered positive (atrophy present) when
the superior border of the supraspinatus lies inferior to a tangential line drawn
between the coracoid process and scapular spine.[25] Fatty infiltration is graded using the Goutallier classification system, where 0
is normal, 1 is mild fatty infiltration with a few fatty streaks, 2 is increased fatty
infiltration involving < 50% of the muscle, 3 is 50% fatty infiltration (equal fat
and muscle), and 4 is > 50% fatty infiltration.[26]
Fig. 4 T1-weighted sagittal oblique Y view demonstrating measurement of the occupation ratio, at the intersection of the scapular spine and body. The yellow circle denotes the
cross-sectional area of the supraspinatus muscle; the green circle denotes the area
of the supraspinatus fossa. The yellow area is divided by the green area to determine
the occupation ratio (R < 0.4, severe supraspinatus atrophy). Note mild fatty infiltration
of the supraspinatus (Goutallier grade 1). The infraspinatus (arrow) is also atrophic
with more pronounced fatty infiltration (Goutallier grade 3).
The “comma sign” is an arthroscopic finding seen with combined superior subscapularis
and anterior supraspinatus tears. Comma-shaped tissue composed of the superior glenohumeral
and coracohumeral ligaments can be followed to identify the retracted superolateral
edge of the torn subscapularis tendon.[16]
[27] Recent studies showed that the comma sign can be prospectively identified on preoperative
imaging, best appreciated in the coronal plane where torn subscapularis fibers are
displaced superiorly through the rotator interval, contiguous with a comma-shaped
fibrous structure lateral to the coracoid process[27] ([Fig. 5]). The presence of the comma sign should be noted because it may alter the arthroscopic
approach and repair technique.[16]
[27]
Fig. 5 “Comma sign.” (a) Coronal T2 fat-sat and (b) axial T2 fat-sat sequences of the left shoulder demonstrate a tear of the superior
subscapularis (solid arrows). The torn fibers are displaced superiorly through the
rotator interval, contiguous with a comma-shaped fibrous structure lateral to the
coracoid process (hashed arrow).
The position of the long head of the biceps should also be assessed in the setting
of a subscapularis tear. The biceps tendon can be subluxed into the substance of the
partially torn subscapularis tendon or may medially dislocate into the anterior glenohumeral
joint in the setting of an articular-sided or full-thickness tear.
The report should address findings that may predispose to subacromial impingement,
such as AC osteoarthritis, acromial enthesopathy, and abnormal acromial morphology
(in particular, lateral downsloping or anterior hooked configuration).
The presence of chondrosis and osteophytes at the glenohumeral joint should be noted
because it may have implications for surgical management. Patients presenting with
end-stage osteoarthritis may not be candidates for primary rotator cuff repair, instead
requiring reverse total shoulder arthroplasty. Superior migration of the humeral head
and resultant narrowing of the acromiohumeral interval often represents the first
stage of rotator cuff arthropathy.
Postoperative Imaging of the Rotator Cuff
Postoperative Imaging of the Rotator Cuff
Indications for primary rotator cuff repair include failure to respond to conservative
measures, ongoing pain, and functional deficits.[28] Small partial-thickness tears are typically treated with debridement, whereas high-grade
partial- or full-thickness tears often require reattachment at the greater tuberosity
with the single- or double-row repair technique.[4] Double-row repair involves anchor fixation of the tendon to the medial articular
margin with a second set of suture anchors at the lateral footprint, with lower reported
rates of retear compared with other techniques.[29]
Imaging of the postoperative rotator cuff is challenging due to distorted anatomy,
postsurgical soft tissue changes, and metal-related artifact, although this is decreased
with titanium or plastic hardware. When available, preoperative imaging and operative
reports should be reviewed to avoid misinterpreting normal postsurgical findings as
pathology. Occasionally, portions of the torn tendon may be left unrepaired by the
surgeon due to poor tendon quality, reduced tendon volume, or an inability to mobilize
the retracted tendon stump.[4] It is also important to consider the time interval between surgery and imaging.
The appearance of the repaired cuff varies, depending on the time elapsed since surgery.
High-resolution non-MRA at 1.5 or 3 T is usually sufficient for most cases of suspected
retear.[4] Although MRA has the highest sensitivity of all imaging modalities for detecting
retears, this examination is more time and resource intensive, and it may overestimate
cuff repair failure due to small areas of irregularity and fraying that are misinterpreted
as tears.[4] CT arthrography avoids the challenges of metal-associated artifact seen with MRI
and is highly sensitive for the detection of partial articular-sided tears postoperatively.[30]
In the early postoperative period, the repaired tendon may appear hypoechoic on US,
hyperintense on MRI, and variable in thickness ([Fig. 6a])[4]
[28]; this may be due to edema, inflammatory change, or early granulation tissue. The
intermediate to increased intratendinous signal on MRI can persist for months to years.[4] CT and MR arthrography must be interpreted with caution when evaluating the postoperative
rotator cuff because repairs are not watertight, allowing for seepage of gadolinium
into the subacromial space, even in an intact repair.[4]
[28]
Fig. 6 (a) Coronal T2 fat-sat sequence of the right shoulder shows intact postoperative rotator
cuff repair of the supraspinatus (arrow). Note the medial anchor at the edge of the
articular surface and more lateral anchor at the greater tuberosity, in keeping with
the double-row repair technique. Mild increased signal in the repaired tendon is an
expected postoperative finding. Because the graft is not watertight, fluid can travel
from the glenohumeral joint into the subacromial-subdeltoid bursa, as seen in this
case. Synovitis was noted at the axillary recess and near the biceps-labral anchor
(open arrows), likely reactive in the postoperative setting. (b) Coronal T2 fat-sat sequences of the right shoulder shows retearing of the previously
repaired supraspinatus, with retracted tendon stump adjacent to the superior labrum
(hashed arrow). Remodeling of the acromial undersurface is related to previous subacromial
decompression with a deficient inferior acromioclavicular (AC) joint, allowing fluid
to traverse from the subacromial-subdeltoid bursa into the AC joint.
Reported rates of rotator cuff retear range from 11% to 68%.[4] Larger rotator cuff tears, increased tendon retraction, and more severe tendon degeneration
are associated with worse clinical outcomes 1 year following repair.[31] One study found that the degree of tendon retraction and narrowing of the acromiohumeral
interval on preoperative MRI were independent predictive factors for retear following
cuff repair.[32] Additional risk factors for retear cited in the literature include advanced age,
diabetes, smoking, atrophy, delamination, longer time to repair, surgical technique,
and inadequate postoperative rehabilitation.[4]
[32]
[33]
Key features of recurrent tears are partial or complete gaps in the tendon, either
at the greater tuberosity or more proximally along the repaired tendon. Recurrent
tears on MRI are depicted by fluid signal extending into the tendon substance or completely
through the tendon with an associated gap ([Fig. 6b]). MRA may be useful in distinguishing between granulation tissue that shows intermediate
to high T2 signal intensity and true recurrent tears, the former not filling with
contrast material.[30] Other ancillary findings suggestive of retear include new or worsening tendon retraction
or hardware loosening.[28]
US may offer advantages over MRI for assessing rotator cuff integrity in the early
postoperative period, such as increased availability and accessibility and the opportunity
for dynamic assessment.[4] The postoperative tendon can appear heterogeneously hypoechoic, thickened, and irregular,
findings that can persist for many years postsurgery.[4] Presence of a fluid-filled defect or nonvisualization of the tendon at the greater
tuberosity attachment is suggestive of tendon retear.[4] Occasionally focal scarring and fibrosis may produce hypoechoic clefts in the tendon
substance mimicking a tear; however, unlike a true tear, these areas should show progressive
healing on follow-up US.[4] Limitations of US in evaluating the postoperative cuff include the operator-dependent
nature of this modality and artifact from postsurgical hardware and suture anchors.[4]
Other postoperative complications that can occur include broken sutures, displaced
anchors ([Fig. 7]), progressive muscle atrophy or fatty infiltration, recurrent subacromial spurs,
capsular thickening and edema, nerve injury, chondrolysis, osteoarthritis, and infection.[28] The open surgical approach for rotator cuff repair requires detachment of the deltoid
from the acromion to access the cuff tendons, with reattachment of the deltoid at
the end of surgery; stripping or dehiscence of the deltoid from the site of reattachment
can occur as a complication of an open approach. The risk of deltoid dehiscence is
reduced with a mini-open approach where a vertical split is made in the deltoid to
gain access to the rotator cuff. a postoperative hematoma or fluid collection can
occasionally be seen at the site of the vertical deltoid split.
Fig. 7 (a) Coronal proton-density and (b) axial T2 fat-sat sequences of the right shoulder show anchor loosening and pullout
at the site of previous rotator cuff repair (arrows). The repaired tendon is intact
(a, hashed arrow); mild thickening and altered in signal of the tendon is within normal
limits postoperatively. There is edema in the deltoid musculature surrounding the
retracted screw.
Subacromial decompression may be performed as part of rotator cuff surgery, with any
combination of the following: acromioplasty for hooked acromion or enthesopathy, subacromial-subdeltoid
bursectomy, resection of the os acromiale and resection of the distal clavicle (Mumford
procedure).[28] Suggestive findings on postoperative imaging include susceptibility artifact at
the AC joint, surgical defect along the acromial undersurface, subacromial scarring,
debrided distal clavicle, and pseudo-widening of the AC joint.[28] Fluid in the subacromial-subdeltoid space may persist for years after surgery.
Massive Rotator Cuff Tears and Repair
Massive Rotator Cuff Tears and Repair
The reported prevalence of massive tears is 10 to 40%,[34] defined as a tear measuring > 5 cm in either the anteroposterior dimension or complete
tears involving at least two adjacent tendons.[35]
Massive rotator cuff tears are difficult to repair surgically due to retraction, scarring
and fibrosis of the tendon, and associated muscle atrophy and fatty infiltration.[34]
[36] Although arthroscopic repair of chronic massive rotator cuff tears was shown to
improve pain, range of motion, and functional outcome, the retear rate is high, up
to 79% in one systemic review.[37]
The management of massive rotator cuff tears varies, depending on patient age, tendon
quality, muscle atrophy, and the presence of secondary glenohumeral degenerative changes.[36] Older patients with decreased functional demands may undergo debridement of the
tear with or without biceps tenotomy or tenodesis, or they may require reverse total
shoulder arthroplasty in the setting of rotator cuff arthropathy.[36] Reverse total shoulder arthroplasty is preferred in patients with a deficient rotator
cuff because there is reduced superior migration of the humeral component and greater
deltoid compensation compared with conventional total shoulder arthroplasty.[1]
In patients with minimal osteoarthritis and high functional requirements, several
surgical techniques can be used when primary repair is not possible: patch graft augmentation,
patch graft bridging, muscle-tendon transfer, and superior capsular reconstruction.[36]
[Table 1] summarizes these techniques.
Table 1
Surgical techniques for massive rotator cuff repair, expected postoperative findings,
and signs of graft failure
|
Patient selection
|
Tissue reconstruction
|
Normal magnetic resonance appearance
|
Common sites of failure
|
|
Patch graft augmentation
|
Irreparable defect (< 1 cm)
|
Autograft (fascia lata), allograft (acellular human dermal graft), xenografts (porcine
dermis), synthetic
|
Low to intermediate heterogeneous signal at site of previous tear
Medially attached to stump of torn tendon; laterally attached to greater tuberosity
|
Lateral attachment at greater tuberosity footprint
Medial row suture failure (if double-row repair technique used)
|
|
Patch graft bridging
|
Defects > 1 cm; inadequate excursion of native tendon due to scar
|
Same as above
|
Same as above
|
Medial tendon attachment; lateral less common
|
|
Muscle-tendon transfer
|
Younger patients; no arthrosis
|
Latissimus dorsi (posterosuperior tear), pectoralis major (anterosuperior tear)
|
Low to intermediate, heterogeneous signal of transferred tendon
|
Lateral attachment at greater tuberosity
|
|
Superior capsular reconstruction
|
Younger patients; no arthrosis
|
Dermal allograft, less commonly fascia lata allograft
|
Low-signal taut graft, uniform thickness (3 mm) with intact attachments to superior
glenoid tubercle medially and greater tuberosity laterally
Small suture holes or fenestrations are expected postoperative finding
|
Lateral attachment at greater tuberosity
Less frequently at medial glenoid attachment or intrasubstance
|
Grafts are used to augment the repair at the footprint when the native cuff cannot
provide full coverage, when the tendon is poor quality, or in the setting of severe
retraction with inability to mobilize the tendon.[29] Patch graft augmentation is often used for small irreparable defects; patch graft bridging is preferred for larger defects with scarring and inelasticity of the native tendon.
Superior capsular reconstruction aims to restore glenohumeral instability and function
by reconstructing the superior joint capsule, and it has lower failure rates than
conventional patch grafting.[38] The original graft used for this procedure was tensor fascia lata autograft; more
recently, dermal allograft has been used to reduce complications at the donor graft
site and improve surgical time.[36]
Muscle-tendon transfer is an option for younger patients with irreparable symptomatic
massive rotator cuff tears without glenohumeral osteoarthrosis, most commonly utilizing
the latissimus dorsi or pectoralis major tendons.[39] When the latissimus dorsi is transferred to the greater tuberosity, it can simulate
posterosuperior rotator cuff function (external rotation and humeral head depression[39]
[40]) that helps improve the biomechanics of the glenohumeral motion.[36] Pectoralis major transfers are more frequently performed to restore anterosuperior
cuff deficiencies.[39]
The interpreting radiologist should be aware of the postoperative appearance of each
of these techniques and the typical locations for graft failure. Normal intact grafts
are low to intermediate in signal with minor heterogeneity.[36] Grafts may fail at three major sites: (1) at the proximal or distal attachments,
(2) midsubstance, or (3) at the side-to-side anastomosis with the native tendon.[41] The presence of a fluid cleft or defect in any of these locations is in keeping
with a graft tear. [Table 1] summarizes the expected postsurgical findings for the different surgical techniques,
with further discussion on latissimus dorsi tendon transfer and superior capsular
reconstruction here (which are the preferred treatments for massive rotator cuff tears
at our institution).
With latissimus dorsi tendon transfer, the tendon is removed from its humeral attachment
at the base of the intertubercular groove and passed through a soft tissue plane fashioned
deep to the deltoid and posterior to the teres minor muscle.[34] The tendon graft is sutured to the lateral aspect of the greater tuberosity to assist
with external rotation.[34] On MR imaging, the normal tendon transfer is low to intermediate in signal with
suture anchors at the greater tuberosity attachment[36] ([Fig. 8]). Complications can occur in up to 9.5% of latissimus dorsi transfer procedures
and include infection, hematoma, injury to the latissimus muscle neurovascular pedicle,
and tears of the tendon graft.[42] The latissimus dorsi muscle should be evaluated for denervation edema and fatty
infiltration that could result from nerve damage during surgery.[36]
Fig. 8 (a) Coronal proton-density (PD) and (b) axial PD-weighted sequences showing the normal postoperative appearance of a latissimus
dorsi tendon transfer. The latissimus dorsi tendon (arrows) is detached from its humeral
attachment and passed through a soft tissue plane between the deltoid and teres minor,
then sutured to the lateral aspect of the greater tuberosity.
In superior capsular reconstruction, suture anchors are typically placed at the 10
and 2 o'clock position of the glenoid with additional anchors at the articular margin
of the greater tuberosity and lateral footprint[36] ([Fig. 9]). The distance between these anchors is measured to determine the appropriate graft
size for adequate humeral head coverage. The graft is advanced into the subacromial
space and fastened to the suture anchor. Additional side-to-side anastomoses with
the residual native cuff anteriorly and posteriorly may be performed in some cases.[36] The graft can tear at the glenoid or humeral anchors or in the midsubstance ([Fig. 10]). Care should be taken not to interpret small suture holes in the graft as tears;
these are an expected postsurgical finding. The acromiohumeral interval is also a
useful predictor for graft integrity following superior capsular reconstruction, with
progressive narrowing of the interval and eccentric humeral head positioning suggestive
of graft failure.[41]
Fig. 9 (a) Coronal T2 fat-sat and (b) axial T2 fat-sat sequences of the left shoulder showing the normal postoperative
appearance of a superior capsular reconstruction. The graft (hashed arrows) is homogeneously
low in signal and uniform thickness, with the medial margin secured at the superior
glenoid tubercle and the lateral margin attached to the greater tuberosity.
Fig. 10 (a) Anteroposterior radiograph and (b) coronal T2 fat-sat sequence of the left shoulder demonstrating failed superior capsular
reconstruction. The acromiohumeral interval (a, open arrow) showed progressive narrowing
postoperatively, a finding that may suggest graft failure. The magnetic resonance
image in (b) shows a focal defect (arrow) in the midsubstance of the graft with fluid
traversing through this defect.
Other modes of failure in rotator cuff repair include anchor loosening and suture
cutout from the tendon.[29] When a double-row suture repair technique is used, failure may occur at the medial
anchors, exposing the articular margin while the anchors at the lateral footprint
remain intact. Progressive fatty infiltration of the rotator cuff musculature may
be another helpful ancillary feature, suggesting graft failure when direct signs of
retear are absent.[29]
Conclusion
Rotator cuff tears are one of the most commonly encountered pathologies in musculoskeletal
radiology. First-line investigations include radiography, followed by MRI or US that
show similar sensitivities for detecting pathology in the native rotator cuff. MRA
should be reserved for postoperative cases to delineate granulation tissue and fibrosis
from retear. Several features should be described in the report that will help guide
management, most notably tear size, delamination and tendon quality, retraction, and
degree of fatty atrophy and infiltration. Radiologists should be aware of the various
surgical techniques used for partial- and full-thickness and massive rotator cuff
tears and the expected postoperative imaging appearances, to improve diagnostic accuracy
for the detection of recurrent tears.
