| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Mallinckrodt Institute of Radiology, Washington University School of Medicine,
510 S Kingshighway Blvd., St. Louis, MO 63110.
2 Department of Orthopedic Surgery, Washington University School of Medicine,
St. Louis, MO 63110.
Received January 19, 2004;
accepted after revision March 8, 2004.
Address correspondence to W. D. Middleton.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Two radiologists independently scanned 61 patients with shoulder pain. Each radiologist categorized the rotator cuff as normal, partially torn, or fully torn. When a tear was present, the tendons involved were specified. All diagnoses were made prospectively without knowledge of the findings of the other radiologist.
RESULTS. The radiologists were in full agreement in the categorization of 92% (56/61) of the patients. In four of the five discrepant cases, the disagreement was whether there was a full-thickness or a partial-thickness tear. The radiologists were in agreement concerning which tendons were involved in 80% (41/51) of the patients in whom a tear was detected by both observers. In all 10 discrepant cases, the disagreement was whether a tear involved both the supraspinatus and infraspinatus tendons or was isolated to one or the other of these tendons.
CONCLUSION. The level of interobserver variability in the sonographic detection and characterization of rotator cuff tears is low.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Both observers used similar scanning protocols. Patients were scanned while seated. The subscapularis tendon was examined in long and short axis with the shoulder externally rotated. The supraspinatus and infraspinatus tendons were scanned in long and short axis with the hand positioned on the posterior hip and the elbow flexed and directed posteriorly. All examinations were performed with a Siemens Elegra unit using a 7.5-MHz linear array transducer. Imaging parameters such as scanning frequency, focal zone number and placement, gain, field of view, and fundamental versus harmonic imaging mode were not standardized but left to the discretion of the observer. Most patients were scanned in the harmonic mode with a transmit frequency of 4.5 MHz.
After the examinations, each observer filled out a data sheet. The rotator cuff was graded as normal or torn. Torn cuffs were further graded as full-thickness, partial-thickness, or indeterminate tear (i.e., the observer can see that the cuff is torn but is unable to distinguish between a full- or partial-thickness tear). In addition, tears were localized to the subscapularis, supraspinatus, infraspinatus, or a combined tear of more than one tendon. The supraspinatus tendon was defined as the portion of the superior cuff within 1.5 cm of the biceps tendon. The remainder of the posterior cuff that was more than 1.5 cm posterior to the biceps tendon was considered the infraspinatus tendon. For the purpose of this study, the teres minor and the infraspinatus tendon were considered a single unit.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
The two observers were in full agreement about the status of the rotator cuff in the remaining 56 patients (92%). The sonograms of 38 of these patients were interpreted as showing full-thickness tears (Fig. 2A, 2B); nine, as showing partial-thickness tears (Fig. 3A, 3B); and nine, as showing normal rotator cuffs. Thirty-five of these 56 patients underwent surgery. All 23 full-thickness tears detected surgically were correctly classified sonographically. All eight partial-thickness tears detected surgically were detected sonographically; however, four were classified as full-thickness tears and four were classified as partial-thickness tears. Of the four rotator cuffs found to be normal at surgery, three were classified as normal on sonography and one was classified as a partial-thickness tear.
|
|
|
|
Both observers agreed about the location of the tear and the tendons involved in 41 of the 51 patients who had a tear (either partial- or full-thickness) that was detected by both observers. In 27, the tear involved both the supraspinatus and the infraspinatus tendons. In 13, the tear was isolated to the supraspinatus tendon. In one patient, the tear was isolated to the infraspinatus tendon. In 10 patients, the observers disagreed concerning which tendons were involved. In nine patients, one of the observers thought that the tear was isolated to the supraspinatus tendon, whereas the other observer thought that the tear involved both the supraspinatus and the infraspinatus tendons. In one patient, one observer thought that the tear was isolated to the infraspinatus tendon, whereas the other observer thought that the tear involved both the supraspinatus and the infraspinatus tendons.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
On the other hand, MRI is equally accurate [11–14], it gives a more global evaluation of the shoulder, and it has a shorter learning curve; consequently many radiologists are trained to interpret shoulder MR images. Perhaps as important as any of the advantages mentioned above, MRI has been shown in several studies to be a reproducible technique with low interobserver variability for detection of full-thickness tears [3, 4].
Because sonography is rightfully viewed as one of the most operator-dependent imaging examinations, and examination of the shoulder is considered one of the more challenging aspects of sonography, one might expect that there would be high interobserver variability for this examination. However, to our knowledge, there has been no systematic study specifically analyzing the interobserver variability for the examination. Undoubtedly, one of the impediments of investigating interobserver variability in shoulder sonography is the intrinsic difference in the way in which sonographic images are acquired compared with the way in which images are acquired with other techniques, such as MRI. With sonography, the recorded images largely display the findings observed by the individual performing the examination. If the examiner detects a rotator cuff tear, the recorded images show a tear. If the examiner misses a rotator cuff tear, then the recorded images do not show a tear. Because of these factors, it is not realistic to measure interobserver variability by having different observers interpret the same set of images. It actually requires having two independent examiners perform two completely separate examinations and interpret the results on the basis of real-time examinations and their own set of stored images.
The latter approach was adopted for this study, and the results confirm that interobserver variability is low. The two observers agreed on the classification of the cuff status in 92% of patients. The observers were in disagreement for only five patients. Three disagreements occurred when both observers detected a tear, but one of the observers thought that it was not possible to confidently categorize the tear as a full- or partial-thickness tear and therefore placed it in the indeterminate tear category. These were the only three patients with tears for which the indeterminate tear category was used. In the other two discrepant cases, the difference in categorization was partial-thickness tear versus normal and partial-thickness tear versus full-thickness tear. Thus, there was only a single case in which the difference in categorization was between a normal versus torn rotator cuff, and there were no cases in which one observer diagnosed a full-thickness tear and the other diagnosed a normal rotator cuff. Therefore, even when there was a disagreement, the discrepancy was typically minor.
As part of this study, the observers were also required to determine the rotator cuff tendon or tendons that were involved in the tear. In 80% (41/51) of the patients in whom both observers detected a tear, the observers agreed on the tendons involved. In the remaining 10 patients, the disagreements all occurred when one of the two observers thought that the tear involved both the supraspinatus and the infraspinatus tendons, whereas the other observer thought that the tear was isolated to either the supraspinatus (n = 9) or the infraspinatus (n = 1) tendon. We believe that these disagreements were relatively minor and are unlikely to have a significant impact in the treatment of these patients.
There are limitations to our study. The mix of patients reflects the type of population seen by a subspecialized shoulder surgeon who primarily sees referred patients who have a high likelihood of having significant structural abnormalities. For this reason, most patients enrolled had a torn rotator cuff, and only 15% of patients had normal cuffs. This high percentage of patients with torn cuffs is unlikely to match most patient populations referred for shoulder sonography. However, if anything, we believe that the high percentage of torn cuffs is more likely to make it more difficult for observers to agree. As our study showed, it is quite uncommon for a normal cuff to be confused with a torn cuff. However, distinguishing an extensive partial-thickness tear from a nonretracted full-thickness tear can be problematic [8]. In fact, this situation accounted for all but one of the discrepancies in our study. If there were more normal rotator cuffs and fewer torn cuffs, we believe that our rate of agreement would have improved.
A second limitation is the lack of surgical proof in all patients. Although such proof would have been desirable, many factors influence the choice of whether to perform surgery in patients with a rotator cuff tear, and some patients or their physicians opted for conservative treatment. As mentioned previously, many studies have documented the sensitivity of sonography for the detection of rotator cuff tears. In a recent technology assessment investigation that included analysis of 38 cohort studies for sonography and 29 for MRI, Dinnes et al. [14] found that either test could be used equally for detection of full-thickness rotator cuff tears. Accordingly, the focus of this study was to evaluate interobserver variability, not to reconfirm the accuracy. We believe that we were justified to include the patients who lacked surgical confirmation.
The final limitation relates to the observers. Both were experienced operators, each having more than 5 years of experience scanning shoulders. If a less experienced operator had participated in the study, it is likely that the variability would have been greater [13]. At our institution, only two radiologists perform shoulder sonography; a third less experienced operator was not an option. This limitation emphasizes one of the major problems with shoulder sonography. Only a limited number of individuals have the expertise to scan shoulders, and this fact has retarded growth of the technique. Hopefully, our results dealing with the reproducibility of the examination will add to the existing body of literature that shows that sonography is an accurate method and will stimulate more radiologists to become trained in shoulder sonography.
In summary, we have shown that in experienced hands, sonography has a low level of interobserver variability for the detection, classification, and localization of rotator cuff tears.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. L. Rees The pathogenesis and surgical treatment of tears of the rotator cuff J Bone Joint Surg Br, July 1, 2008; 90-B(7): 827 - 832. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Al-Shawi, R. Badge, and T. Bunker The detection of full thickness rotator cuff tears using ultrasound J Bone Joint Surg Br, July 1, 2008; 90-B(7): 889 - 892. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J M Koski, S Saarakkala, M Helle, U Hakulinen, J O Heikkinen, H Hermunen, P Balint, G A Bruyn, E Filippucci, W Grassi, et al. Assessing the intra- and inter-reader reliability of dynamic ultrasound images in power Doppler ultrasonography Ann Rheum Dis, December 1, 2006; 65(12): 1658 - 1660. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E Naredo, I Moller, C Moragues, J J de Agustin, A K Scheel, W Grassi, E de Miguel, M Backhaus, P Balint, G A W Bruyn, et al. Interobserver reliability in musculoskeletal ultrasonography: results from a "Teach the Teachers" rheumatologist course Ann Rheum Dis, January 1, 2006; 65(1): 14 - 19. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Teefey, W. D. Middleton, W. T. Payne, and K. Yamaguchi Detection and Measurement of Rotator Cuff Tears with Sonography: Analysis of Diagnostic Errors Am. J. Roentgenol., June 1, 2005; 184(6): 1768 - 1773. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
