1 From the Departments of Radiology (S.C.W., M.E.S., W.B.M.) and Orthopedic Surgery (J.M.F.), Thomas Jefferson University Hospital, 3390 Gibbon, 111 S 11th St, Philadelphia, PA 19107; and Department of Orthopedic Surgery, Pennsylvania Hospital, University of Pennsylvania Health System, Philadelphia (A.R.B.). From the 2000 RSNA scientific assembly. Received March 5, 2001; revision requested April 9; revision received June 11; accepted July 5.
PURPOSE: To analyze a series of postoperative magnetic resonance
(MR) images obtained in patients with recurrent signs or symptoms
of instability in whom subsequent surgical correlation was performed
to determine the accuracy in diagnosing recurrent injury.
MATERIALS AND METHODS: The authors identified 24 patients who underwent MR imaging after shoulder instability surgery and had recurrent instability requiring repeat surgery. Twelve nonenhanced MR images and six indirect and six direct MR arthrograms were retrospectively reviewed with consensus to determine the presence or absence of recurrent labral or rotator cuff tear. Operative reports were reviewed to confirm the presence or absence of labral or rotator cuff tear. The mean interval between initial surgery and MR imaging was 10 months. The mean interval until repeat surgery was 2 months.
RESULTS: Overall, the accuracy of postoperative MR imaging was 79% in depicting recurrent labral tear and 88% in depicting recurrent rotator cuff tear. Indirect MR arthrography had 100% accuracy for recurrent labral tear detection, whereas direct MR arthrography and nonenhanced MR imaging had accuracies of 67% and 75%, respectively. Direct MR arthrography was more sensitive, 100% versus 71%, but less specific, 60% versus 80%, than nonenhanced MR imaging in depicting recurrent labral tears. Direct MR arthrography had 100% accuracy in depicting rotator cuff tear, whereas both indirect MR arthrography and nonenhanced MR imaging had 83% accuracy.
CONCLUSION: MR imaging, indirect MR arthrography in particular, appears to be an accurate means of evaluating the shoulder following instability surgery.
Index terms: Magnetic resonance (MR), arthrography, 41.121411, 41.121412, 41.121415, 41.121416, 41.12143 • Shoulder, abnormalities, 41.4811, 48.4812, 41.4813, 41.4819 • Shoulder, arthrography, 41.121411, 41.121412, 41.121415, 41.121416, 41.12143 • Shoulder, MR, 41.121411, 41.121412, 41.121415, 41.121416, 41.12143
Shoulder instability is associated with abnormalities of the
shoulder capsule, labrum, glenohumeral ligaments, rotator cuff,
and/or surrounding musculature. Magnetic resonance (MR) imaging
is an accurate means of evaluating potential causes of shoulder
instability (
1–
4). Surgical repair for shoulder instability
is a common procedure, yet to our knowledge, there has been
little research dedicated to the evaluation of the postoperative
status of these patients. These patients may present with recurrent
dislocations, pain, clicking, or popping and are at risk for
recurrent labral injury and other shoulder abnormalities, such
as rotator cuff tears.
Other authors (5) have discussed the expected postoperative MR imaging findings in the shoulder after instability surgery. To our knowledge, the accuracy of postoperative MR imaging performed in patients with recurrent shoulder instability has not been previously reported. Therefore, the purpose of our study was to analyze a series of postoperative MR imaging studies in patients with recurrent signs or symptoms of instability in whom subsequent surgical correlation was performed to determine the accuracy in diagnosing recurrent injury.
Patient PopulationPatients were identified retrospectively by means of review
of the radiologic reports of shoulder MR imaging, direct MR
arthrography, and indirect MR arthrography performed from January
1992 to January 2000. We identified these reports by searching
with the keywords "post-op," "post-surgical," "instability,"
and "recurrent subluxation/dislocation" in our electronic report
database. We identified 84 patients by using the described keyword
criteria, and a chart review was performed (by S.C.W.) for all
patients after approval was obtained from the institutional
review board at Thomas Jefferson University Hospital. Informed
consent was not required. From this group, we identified 24
patients (18 men, six women; mean age, 30 years; age range,
22–53 years) with a history of shoulder instability and
prior capsulolabral repair who had presented with signs or symptoms
of recurrent instability and subsequently returned for repeat
surgery after MR imaging. The signs and symptoms of instability
included dislocation, subluxation, pain, clicking, and catching
(ie, locking sensation during movement). Subsequent operative
reports on each patient were reviewed (by S.C.W.) to confirm
the presence or absence of labral or rotator cuff tear at the
time of repeat surgery. The repeat surgeries included 15 arthroscopic
and nine open capsulolabral repairs.
Imaging Studies
The MR imaging studies included 12 nonenhanced images, six indirect arthrograms, and six direct arthrograms. All studies were obtained with a 1.5-T unit (Signa; GE Medical Systems, Milwaukee, Wis). Nonenhanced MR imaging was performed by using the following pulse sequences: coronal-oblique T1-weighted spin echo with 600–650/14–15 (repetition time msec/echo time msec), coronal-oblique fat-suppressed fast spin echo with 4,000–2,100/75–84 (effective), sagittal-oblique fat-suppressed fast spin echo with 4,000/85 (effective), and transverse three-dimensional gradient-recalled echo (GRASS; GE Medical Systems, Milwaukee, Wis) with 22.1–36.0/5.5–15.0 and a 20° flip angle.
Direct MR arthrography was performed by using a fluoroscopically guided intraarticular injection of 12–15 mL of a 3 mmol/L solution of gadopentetate dimeglumine (Magnevist; Berlex, Wayne, NJ), with MR imaging performed at completion of the injection. Indirect MR arthrography was performed 20 minutes after the intravenous injection of 0.1 mmol/kg of gadopentetate dimeglumine. The protocols for direct and indirect MR arthrography included the following pulse sequences: transverse fat-suppressed T1-weighted imaging with 433–766/8–16; coronal-oblique fast spin-echo imaging with 2,433–3,417/69–125 (effective); coronal-oblique fat-suppressed T1-weighted imaging with 400, 600, 767/9–15; transverse three-dimensional fast spoiled gradient-recalled-echo imaging with 22.8–26.3/9.8–11.3 and a 45° flip angle, and sagittal-oblique fat-suppressed T1-weighted imaging with 400–767/8–15.
The following parameters were used for all the MR modalities: 256 x 192 matrix, 13–16-cm field of view, 5-mm section thickness with 1-mm intersection gap for the sagittal plane, 3–4-mm section thickness with no intersection gap for the coronal-oblique plane, and 1.3-mm section thickness with no intersection gap for transverse three-dimensional gradient-recalled-echo and fast spoiled gradient-recalled–echo sequences.
Image Analysis
All studies were reviewed by two musculoskeletal radiologists (M.E.S., W.B.M.) who were blinded to the surgical reports and given only the clinical histories of recurrent instability after previous surgery. These reviewers were not involved in the prior patient identification and selection process. The reviewers used consensus agreement to determine the presence or absence of recurrent labral injury or rotator cuff tear. The criteria used to diagnose a labral tear were derived from the literature (3) and included the identification of fluid or contrast material, either within the labrum or dissecting under the labrum, that accounted for the potentially normal appearance of thin, linear fluid under the anterior and superior labra. The criteria used to diagnose a rotator cuff tear were derived from the literature (6) and included the identification of focal fluid signal intensity in the tendon on T2-weighted images or contrast material extending into or through the tendon, with acknowledgment that in postoperative patients, fluid normally may track into the subdeltoid and subacromial bursae.
The mean interval between initial surgery and MR imaging was 10 months (range, 1–24 months). The mean interval between MR imaging and repeat surgery was 2 months (range, 1 week to 6 months).
Data Analysis
We compared the presence or absence of labral tear or rotator cuff tear at MR imaging with the surgical findings in each patient and determined the number of true-positive, true-negative, false-positive, and false-negative imaging results. The sensitivity, specificity, and accuracy of each imaging modality were calculated. Our small sample size limited statistical analysis; however, we reviewed the cases of false-positive or false-negative results a second time to evaluate for confounding factors that may have led to the incorrect interpretations.
Recurrent Labral TearThe numbers of true-positive, true-negative, false-positive,
and false-negative cases, as well as the sensitivity, specificity,
and accuracy, with each MR imaging modality in the diagnosis
of recurrent labral tear are listed in
Table 1. Overall, MR
imaging yielded 80% sensitivity, 79% specificity, and 79% accuracy
in the diagnosis of recurrent labral tear. Specifically, nonenhanced
MR imaging yielded 71% sensitivity, 80% specificity, and 75%
accuracy. In the 12 patients who underwent nonenhanced MR imaging,
there were two false-negative readings and one false-positive
reading.
| fig.ommitted |
TABLE 1. Detection of Recurrent Labral Tears at MR Imaging and MR Arthrography
| |
There were seven surgically proved labral tears: five anterior,
one superior, and one involving both the anterior and posterior
labra. Nonenhanced MR imaging correctly depicted one superior
and four anterior labral tears, whereas a normal labrum was
correctly reported in four of five patients.
Figure 1 shows
an example of a true-positive recurrent anterior labral tear
at nonenhanced MR imaging. One of the two false-negative cases,
that of a patient with prior capsulolabral reconstruction in
whom nonenhanced contrast MR imaging failed to depict a recurrent
anterior labral tear, is illustrated in
Figure 2. The second
false-negative recurrent labral tear occurred in a patient in
whom extensive metallic artifact limited visualization. The
false-positive recurrent labral tear was in a patient who had
previously undergone labral debridement and had an apparent
inferior labral tear that was not noted at subsequent surgery
(
Fig 3).
| fig.ommitted |
Figure 1. Transverse nonenhanced three-dimensional gradient-recalled-echo MR image (36/15) of the shoulder of a 36-year-old man who had undergone anterior stabilization surgery shows a true-positive recurrent labral tear. Image shows that, compared with the signal intensity of the normal posterior labrum (open arrow), there is increased signal intensity in the region of the anterior labrum (solid arrow). Findings were proved at subsequent surgery and represented a recurrent anterior labral tear.
| |
fig.ommitted
|
Figure 2. Transverse nonenhanced three-dimensional gradient-recalled-echo MR image (36/15) of the shoulder of a 41-year-old man who had undergone instability surgery shows a false-negative recurrent labral tear. Image shows a round anterior labrum (arrow) without definite evidence of a tear. Even at second review, no definite tear was identified; however, an anterior labral tear was seen at subsequent surgery.
| |
| fig.ommitted |
Figure 3. Coronal-oblique nonenhanced fat-suppressed fast spin-echo MR image (4,000/75) of the shoulder of a 36-year-old man with an inferior labral tear who had undergone instability surgery shows a false-positive recurrent labral tear. Image demonstrates linear high signal intensity (arrow) within the inferior labrum that was interpreted as a recurrent inferior labral tear; however, no tear was identified at subsequent surgery.
| |
Indirect MR arthrography resulted in no false-positive or false-negative
findings in the six patients, yielding 100% sensitivity, specificity,
and accuracy in the diagnosis of recurrent labral tear. There
were two anterior labral tears and one superior labral tear,
all of which were surgically proved and correctly identified
at arthrography. A true-positive recurrent anterior labral tear
is shown in
Figure 4, and a true-negative labrum is shown in
Figure 5; both findings were depicted at indirect MR arthrography
performed in patients with prior Bankart repairs.
| fig.ommitted |
Figure 4. Transverse three-dimensional fast spoiled gradient-recalled-echo indirect MR arthrogram (22.9/9.8) of the shoulder of a 23-year-old man who had undergone Bankart repair shows a true-positive recurrent labral tear. Image shows a fixation device (solid arrow) within the glenoid and intermediate signal intensity (open arrow) in the expected location of the anterior labrum. This signal intensity was shown to represent a recurrent anterior labral tear at subsequent surgery.
| |
fig.ommitted
|
Figure 5. Transverse T1-weighted fat-suppressed spin-echo indirect MR arthrogram (717/16) of the shoulder of a 23-year-old man who had undergone Bankart repair shows a true-negative recurrent labral tear. Image shows postsurgical changes in the anterior labrum with intermediate signal intensity (arrow) at the junction of the labrum and glenoid; however, there is no fluid depicted within the labrum to suggest a recurrent tear. The anterior labrum was intact at subsequent surgery.
| |
Direct MR arthrography performed in six patients resulted in
two false-positive and no false-negative findings, yielding
100% sensitivity, 60% specificity, and 67% accuracy in the diagnosis
of recurrent labral tear. One surgically proved superior labral
tear was correctly identified; however, one anterior and one
superior labral tear were reported at imaging of a surgically
proved normal labrum.
Figure 6 shows a true-positive posterior
recurrent labral tear at direct MR arthrography. One of the
two false-positive cases, that in which metallic artifact mimicked
a recurrent anterior labral tear in a patient who had undergone
Bankart repair, is shown in
Figure 7, and the other, a case
of overreading a frayed, irregular labrum, is shown in
Figure 8.
fig.ommitted
|
Figure 6. Transverse three-dimensional fast spoiled gradient-recalled-echo direct MR arthrogram (26.3/11.3) of the shoulder of a 24-year-old man who had undergone instability surgery shows a true-positive recurrent labral tear. Image shows a small cyst adjacent to a rim of high signal intensity (arrow) surrounding the posterior labrum. The high signal intensity represents a recurrent posterior labral tear, which was confirmed at subsequent surgery.
| |
fig.ommitted
|
Figure 7. Transverse T1-weighted fat-suppressed spin-echo direct MR arthrogram (450/20) of the shoulder of a 30-year-old woman who had undergone Bankart repair shows a false-positive recurrent labral tear. Image shows a low-signal-intensity focus (arrow) that was thought to represent the anterior labrum separated from the glenoid; however, the anterior labrum was intact at subsequent surgery, and these findings represented metallic artifact mimicking a recurrent anterior labral tear.
| |
| fig.ommitted |
Figure 8. Transverse T1-weighted fat-suppressed spin-echo direct MR arthrogram (433/10) of the shoulder of a 34-year-old woman who had undergone instability surgery shows a false-positive recurrent labral tear. Image shows irregularity and increased signal intensity in the anterior labrum (arrow). The increased signal intensity was interpreted as a recurrent anterior labral tear; however, no recurrent tear was identified at subsequent surgery.
| |
Recurrent Rotator Cuff TearTable 2 shows the numbers of true-positive, true-negative, false-positive,
and false-negative cases, as well as the sensitivity, specificity,
and accuracy, with each MR imaging modality in the diagnosis
of a rotator cuff tear in the shoulder after instability surgery.
Overall, MR imaging yielded 50% sensitivity, 100% specificity,
and 88% accuracy. Nonenhanced MR imaging performed in 12 patients
resulted in no false-positive cases, two false-negative cases,
and one true-positive case, yielding 33% sensitivity, 100% specificity,
and 83% accuracy. Three tears were proved at surgery: one supraspinatus
full-thickness tear, one partial-thickness tear, and one subscapularis
partial-thickness tear. Nonenhanced MR imaging depicted only
the supraspinatus full-thickness tear, possibly missing the
partial-thickness tears because of the lack of intraarticular
contrast or joint distention (
Fig 9).
| fig.ommitted |
TABLE 2. Detection of Recurrent Rotator Cuff Tears at MR Imaging and MR Arthrography
| |
| fig.ommitted |
Figure 9. Coronal-oblique nonenhanced fat-suppressed fast spin-echo MR image (3,650/75) of the shoulder of a 35-year-old man who had undergone anterior stabilization surgery shows a false-negative recurrent rotator cuff tear. Image shows no fluid signal intensity within the supraspinatus tendon (arrow) to suggest a rotator cuff tear; however, at subsequent arthroscopy, a partial undersurface tear of the supraspinatus tendon was identified.
| |
Indirect MR arthrography performed in six patients resulted
in one false-negative and no false-positive cases, yielding
67% sensitivity, 100% specificity, and 83% accuracy. There were
two surgically proved rotator cuff tears: a supraspinatus undersurface
tear and a full-thickness subscapularis tear, both of which
were depicted at arthrography.
Figures 10 and
11 show examples
of true-positive and true-negative rotator cuff tears, respectively.
The false-negative case of a partial rotator cuff tear that
was not identified at indirect MR arthrography in a patient
with prior Bankart repair (
Fig 12) was possibly related to enhancement
surrounding the cuff and subacromial space.
fig.ommitted
|
Figure 10. Coronal-oblique T1-weighted fat-suppressed spin-echo indirect MR arthrogram (517/13) of the shoulder of a 53-year-old man who had undergone instability surgery shows a true-positive recurrent rotator cuff tear. Image shows contrast material extending through the supraspinatus tendon (curved arrow) and within the subacromial and subdeltoid bursae (straight arrow). These findings were consistent with the rotator cuff tear confirmed at subsequent surgery.
| |
| fig.ommitted |
Figure 11. Coronal-oblique fat-suppressed fast spin-echo indirect MR arthrogram (2,633/125) of the shoulder of a 39-year-old man who had undergone instability surgery shows a true-negative recurrent rotator cuff tear. Image shows the normal low signal intensity of the supraspinatus tendon (arrow) without discontinuity or fluid signal intensity within the tendon to suggest a tear. A normal rotator cuff was confirmed at subsequent surgery.
| |
fig.ommitted
|
Figure 12. Coronal-oblique T1-weighted fat-suppressed spin-echo indirect MR arthrogram (650/15) of the shoulder of a 23-year-old man who had undergone Bankart repair shows a false-negative recurrent rotator cuff tear. Image shows no discontinuity or high signal intensity within the supraspinatus tendon. There is slightly increased signal intensity within the subacromial bursae (arrow), which normally may enhance during indirect MR arthrography. This finding was interpreted as normal; however, at subsequent surgery, a partial rotator cuff tear was identified.
| |
Direct MR arthrography performed in six patients resulted in
six true-negative cases and yielded 100% accuracy and 100% specificity.
All rotator cuffs were normal at surgery, and because all of
the cases involved true-negative findings (
Fig 13), sensitivity
was not calculated.
| fig.ommitted |
Figure 13. Coronal-oblique T1-weighted fat-suppressed spin-echo direct MR arthrogram (400/13) of the shoulder of a 24-year-old man who had undergone instability surgery shows a true-negative recurrent rotator cuff tear. Image shows no discontinuity or contrast material within the supraspinatus tendon. The normal recess (curved arrow) on the undersurface of the supraspinatus tendon is shown to have good joint distention, as noted by the distended axillary recess. A recurrent labral tear (straight arrow) with paralabral cyst also is demonstrated. The rotator cuff was interpreted as normal, and this was confirmed at subsequent surgery.
| |
The labrum and surrounding tissues provide stability to the
shoulder, and instability is often related to labral injury
(
7–
9). Surgical management of instability may include
operative labral repair or reconstruction that can result in
loss of normal tissue planes, scar formation, alterations in
joint architecture, and foreign materials within the joint,
making interpretation of postoperative images difficult. Rand
et al (
10) described changes in the capsule without extension
of contrast material into or beneath the labrum, extracapsular
leaks, communication with adjacent bursae, or changes in axillary
recess size at postoperative MR arthrography performed in patients
who had undergone instability surgery. Thus, our standard criteria
for labral and rotator cuff tears are applicable to these patients.
Therefore, we evaluated the findings in a group of patients
who had undergone instability surgery and presented with recurrent
signs or symptoms of instability and correlated their MR imaging
results with subsequent surgical findings to determine the accuracy
of MR imaging in depicting recurrent labral tears and/or rotator
cuff tears.
Recurrent Labral Tear
Our initial experiences showed that despite postoperative alterations, MR imaging is an accurate modality for aiding in the diagnosis of recurrent labral tear with use of standard criteria. Of the MR imaging techniques evaluated, indirect MR arthrography appears to be the most accurate, with no false-positive or false-negative results among six patients. Our study results showed sensitivity and specificity that were comparable to those reported in the literature on diagnosing labral tears preoperatively (11).
Direct MR arthrography and nonenhanced MR imaging demonstrated 67% and 75% accuracy, respectively, in depicting recurrent labral tears. However, direct MR arthrography was more sensitive yet less specific than nonenhanced MR imaging. Further review of the false-positive cases revealed that a metallic anchor in the anterior glenoid from a prior Bankart repair produced an artifact that mimicked a recurrent anterior labral tear. Another false-positive case was likely that of an overreading of a frayed and irregular labrum since no labral tear was seen at second review. There were no false-negative cases of recurrent labral tear at direct MR arthrography, and although this modality was sensitive, there was poor specificity and decreased accuracy. We expected direct MR arthrography to be at least as accurate as indirect MR arthrography, and the decreased accuracy of this modality was likely the result of our small sample size. The two false-positive cases described had a substantial effect on specificity and accuracy, and we believe that a larger sample would reduce this bias. Compared with the reported 91% sensitivity and 93% specificity of direct MR arthrography in the diagnosis of labral tears preoperatively (12), our study results showed similar sensitivity with less specificity.
Nonenhanced MR imaging yielded the lowest sensitivity in depicting recurrent labral tears compared with both indirect and direct MR arthrography. Extensive metallic artifact complicated one of the false-negative cases, whereas joint distention and intraarticular contrast may have helped to delineate the detached labrum in the other case. The false-positive finding was reported in a patient in whom there appeared to be a recurrent labral tear with fluid signal intensity within the inferior labrum, even after second review; however, at surgery, there was no labral tear noted. This may represent another case of overreading of a labral tear. Our study results are within the ranges of 44%–95% sensitivity and 67%–97% specificity reported for the diagnosis of labral tears preoperatively (3,13,14).
On the basis of our initial findings, indirect MR arthrography appears to be the most accurate for the evaluation for recurrent labral tear after instability surgery. Direct MR arthrography and nonenhanced MR imaging have moderate accuracy; however, nonenhanced MR imaging is more likely to yield false-negative results. Thus, we speculate that the joint distention and intraarticular contrast afforded by direct MR arthrography may enable better delineation of labral structures. Metallic artifact appears to affect the accuracy in diagnosing recurrent labral tears. Knowledge of the presence and type of surgical hardware can help the radiologist and orthopedic surgeon predict the accuracy of postoperative imaging or potentially guide their choice of imaging modality. The use of sutures without metallic anchors or of bioabsorbable anchors during labral repair might improve the accuracy of MR imaging in evaluating recurrent symptoms.
Recurrent Rotator Cuff Tear
In our study population, all modalities had poor sensitivity in depicting rotator cuff tears; however, direct MR arthrography appeared to be the most accurate. All cases in the direct MR arthrography group were true-negative; thus, sensitivity data were not obtainable. Our specificity data are comparable to the specificity of 75%–100% in diagnosing rotator cuff tears preoperatively reported by Palmer et al (15).
Indirect MR arthrography and nonenhanced MR imaging had equal accuracy and specificity; however, indirect MR arthrography had higher sensitivity. Indirect MR arthrography resulted in one false-negative and no false-positive cases. Because the undersurface of the cuff and subacromial space enhances at indirect MR arthrography, the detection of subtle increased signal intensity within the rotator cuff itself may be difficult.
At nonenhanced MR imaging, only one of three rotator cuff tears was identified, possibly because of a lack of intraarticular contrast or joint distention. Indirect MR arthrography had lower sensitivity (67% vs 100%) and comparable specificity (100% vs 86%) compared with the values reported in diagnosing rotator cuff tears preoperatively (2). Nonenhanced MR imaging had even lower sensitivity (33% vs 93%) and similar specificity (100% vs 87%) compared with reported preoperative values (1).
Overall, when evaluating the rotator cuff after instability surgery, we found that all errors occurred as false-negative findings, with a tendency toward underreading of cuff tears. We speculate that the readers may have placed too much emphasis on the labrum in these patients, given that the clinical history was recurrent instability. On the basis of our findings, we recommend the use of either direct or indirect MR arthrography in the diagnosis of rotator cuff tears in this setting. Again, knowledge of the metallic appliances used at prior surgery and additional attention to the rotator cuff may be useful in optimizing accuracy.
The effect of imaging sequences and parameters on the appearance of metallic artifact has been previously described (16,17), and our study data showed the presence of metallic artifact to affect the accuracy of postoperative imaging. There are several techniques that one may use to minimize this artifact. For example, the use of inversion recovery instead of fat saturation, use of fast spin-echo instead of conventional spin-echo sequences, and avoiding the use of gradient-echo techniques are recommended. Increasing the bandwidth and decreasing the voxel size also may help minimize metallic artifact.
Our initial experience in evaluating the accuracy of MR imaging after instability surgery was limited by a small sample size, which limited statistical analysis of the data. The retrospective design of the study was another limitation since we could not randomly assign patients to indirect MR arthrography, direct MR arthrography, and nonenhanced MR imaging groups. Also, since our field of view ranged from 13 to 16 cm, depending on patient factors, a slight variation in spatial resolution among studies may have affected the detection of tears. In addition, an initial negative MR imaging report during patient evaluation may have dissuaded the surgeon from performing a second surgery and therefore led to selection bias. However, we believe our study is the first to focus on the evaluation of the accuracy of imaging in the shoulder after instability surgery and thus provides preliminary information that may assist radiologists and orthopedic surgeons in the examination of these patients.
On the basis of our initial experience, MR imaging—indirect MR arthrography in particular—appears to be an accurate means of evaluating the shoulder following instability surgery. Overall, MR arthrography—whether indirect or direct—yielded fewer errors and thus is recommended over nonenhanced MR imaging in this group of postoperative patients. Knowledge of the previously used surgical anchors, emphasis on evaluating the rotator cuff, and use of indirect or direct MR arthrography rather than nonenhanced MR imaging may optimize diagnostic accuracy. Further data collection in a larger number of subjects, with critical review of the incorrect results, may help in the development of improved criteria for evaluating the shoulder after instability surgery.
- Iannotti JP, Zlatkin MB, Esterhai JL, et al. Magnetic resonance imaging of the shoulder. J Bone Joint Surg Am 1991; 73:17-29.
- Shankman S, Bencardino J, Beltran J. Glenohumeral instability: evaluation using MR arthrography of the shoulder. Skeletal Radiol 1999; 28:365-382.
- Legan JM, Burkhard TK, Goff WB, et al. Tears of the glenoid labrum: MR imaging of 88 arthroscopically confirmed cases. Radiology 1991; 179:241-246.
- Seeger LL, Gold RH, Bassett LW. Shoulder instability: evaluation with MR imaging. Radiology 1988; 168:695-697.
- Vahnlensieck M, Lang P, Wagner U, et al. Shoulder MRI after surgical treatment of instability. Eur J Radiol 1998; 30:2-4.
- Kneeland JB, Middleton WD, Carrera GF, et al. MR imaging of the shoulder: diagnosis of rotator cuff tears. AJR Am J Roentgenol 1987; 149:333-337.
- Townley CO. The capsular mechanism in recurrent dislocation of the shoulder. J Bone Joint Surg Am 1950; 32:370-380.
- Mosley HF, Overgaard B. The anterior capsular mechanism in recurrent anterior dislocation of the shoulder. J Bone Joint Surg Am 1962; 44:913-927.
- Pappas AM, Goss TP, Kleinman PK. Symptomatic shoulder instability due to lesions of the glenoid labrum. Am J Sports Med 1983; 11:279-288.
- Rand T, Freilinger W, Breitenseher M, et al. Magnetic resonance arthrography (MRA) in the postoperative shoulder. Magn Reson Imaging 1999; 17:843-850.
- Vahlensieck M, Lang P, Sommer T, Genant HK, Schild HH. Indirect MR arthrography: techniques and applications. Semin Ultrasound CT MR 1997; 18:302-306.
- Palmer WE, Brown JH, Rosenthal DI. Labral-ligamentous complex of the shoulder: evaluation with MR arthrography. Radiology 1994; 190:645-651.
- Gusmer PB, Potter HG, Schatz JA, et al. Labral injuries: accuracy of detection with unenhanced MR imaging of the shoulder. Radiology 1996; 200:519-524.
- Garneau RA, Renfrew DL, Moore TE, et al. Glenoid labrum: evaluation with MR imaging. Radiology 1991; 179:519-522.
- Palmer WE, Brown JH, Rosenthal DI. Rotator cuff: evaluation with fat-suppressed MR arthrography. Radiology 1993; 188:683-687.
- White LM, Kim JK, Mehta M, et al. Complications of total hip arthroplasty: MR imaging—initial experience. Radiology 2000; 215:254-262.
- Tartaglino LM, Flanders AE, Vinitski S, Friedman DP. Metallic artifacts on MR images of the postoperative spine: reduction with fast spin-echo techniques. Radiology 1994; 190:565-569.
作者：
Steven C. Wagner MD Mark E. Schweitzer MD Will 2007-5-12