Distinction between supraspinatus, infraspinatus and subscapularis tendon tears with ultrasound in 332 surgically confirmed cases

Abstract

The aim of this retrospective study was to determine the diagnostic value of preoperative ultrasonography for the characterisation of size and location of the involved tendons. in 332 consecutive patients who underwent surgery, all preoperative ultrasonographic reports were reviewed and ultrasound (US) and surgical findings were compared. Ultrasound criteria for cuff tears were complete nonvisualisation of the cuff tendons or localised absence and focal discontinuity. In all but 12 cases, US diagnoses corresponded with intraoperative findings (sensitivity 98%, confidence interval 95.1–99.3; specificity 93%, CI 85.7–97.1; accuracy 97%, CI 93.8–98.1). Size and location of the tear were correctly predicted in 69 of 96 cases (accuracy 87%, sensitivity 89%, specificity 87%). US demonstrated less extensive tears than observed at surgery in 18%. Ultrasonography was highly accurate and sensitive for detecting rotator cuff tears, but seems to be a method of limited value for evaluation of the size of cuff tears, in particular, for the detection of small tears. (E-mail: [email protected])

Introduction

Ultrasound(US) examination of the shoulder, as an adequate, noninvasive, frequently available and relatively inexpensive technique for the evaluation of pathologic conditions of the rotator cuff, was widely used during the last 10 years. However, in US studies, sensitivity varied between 57% and 100% and specificity between 50% and 98% (Hedtmann and Fett 1995; Hodler et al. 1991; Kurol et al. 1991; Middleton et al. 1985; Olive and Marsh 1992; Sperner et al. 1993; Turrin et al. 1997; van Moppes et al. 1995; Vick and Bell 1990). This enormous variability is attributed to a long learning curve in mastering this technique, and is mainly influenced by operating skills and equipment-dependent factors. Due to the improvement of technical equipment and scanning technique, recent studies (Bachmann et al. 1997; Sonnabend et al. 1997; Seibold et al. 1999; Turrin et al. 1997; van Holsbeeck et al. 1995) consistently demonstrate high sensitivity and specificity rates. In the international literature, only a few studies are reported that evaluated the accuracy of preoperative ultrasonography in larger series with more than 100 surgically confirmed cases (Hedtmann and Fett 1995; Sonnabend et al. 1997; Wiener and Seitz 1993), and even fewer correlated tear location, size and extent with the surgical findings (Teefey et al. 2000; Wiener and Seitz 1993).

The purpose of this retrospective study was to determine the diagnostic accuracy of preoperative high-resolution ultrasonography for the detection of rotator cuff tears and the characterization of size, location and extent of the involved tendons, because the choice of surgical procedure, either open surgery or athroscopy, is mainly based on this information.

Therefore, we decided to assess our own US technique by our surgical results in a sample of 332 cases, and to verify the hypothesis that preoperative US examination is an adequate and reliable instrument for the assessment of subacromial pathologies, especially for identification and distinction of the affected tendons of supraspinatus, infraspinatus and subscapularis in patients with cuff tears. Due to the large sample of 332 cases, higher statistical logical value than already published and small confidence intervals were expected.

Between January 1993 and July 1999, 332 consecutive patients (143 women, 189 men; mean age 53 yr, range 17–90 years) with shoulder pain for more than 3 months underwent standardised ultrasonography of the shoulder before arthroscopy or open shoulder surgery. Patients with shoulder instability were excluded from this study, as well as patients with massive osteoarthrosis or necrosis of the humeral head scheduled for shoulder arthroplasty. Patients with a “frozen” shoulder were also excluded because these patients did not undergo surgery.

The preoperative procedure included clinical examination using the constant shoulder score (Constant 1987). Clinical examination was performed by independent investigators, especially not by the person who performed sonographic evaluation. Additionally, standard X-rays in three planes (a.p., axial, outlet view) were taken.

The decision regarding operative treatment (arthroscopy or open surgery) was based on the US examinations, as well as on clinical and radiological findings. All patients with cuff tears underwent open surgery (n = 96), except three patients with an isolated small rupture of the supraspinatus tendon in whom an arthroscopically assisted reconstruction of the tendon was performed. Arthroscopy was also performed in patients with simple impingement and sonographically—excluded cuff tears, intratendineous calcifications or simple synovitis (n = 205). The arthroscopy included glenohumeral arthroscopy for visualisation of the articular side of the rotator cuff and bursal arthroscopy for inspection of the bursal tendon surface. In cases with small lesions of the supraspinatus tendon, an arthroscopically, assisted procedure with tendon reconstruction was performed. The final clinical diagnoses are shown in Table 1. Surgery was performed by two physicians (C. Wurnig, T. Katterschafka), and all US examinations were performed on the day before surgery and results were locked in the computer immediately after sonographic examination.

Sonographic examination was performed by three orthopaedic surgeons (C. Wurnig, T. Katterschafka, K. Schatz). Sonograms were obtained using an Acuson scanner (128×P/10c, Mountain View, CA) or a Siemens Sonoline scanner (SI-450, Erlangen, Germany) with a 7.5-MHz linear-array transducer in real-time.

The examination technique was standardised and performed as described by Hedtmann and Fett (1995)). The patients were seated on a stool with the arm hanging free, using armrests neither for the patient nor for the examiner. The examiner was standing behind the patient, facing the screen, with one hand leading the transducer. Each component of the rotator cuff and the biceps tendon in the bicipital groove were examined in both axial and sagittal planes. Evaluation of the supraspinatus, infraspinatus and subscapularis tendons included views in internal and external rotation of the shoulder. For evaluation of the supraspinatus tendon, the shoulder was extended downward and rotated internally. The rotator cuff tendons were displayed in two different transducer positions. During longitudinal scanning, the transducer was placed just in front of the lateral and anterior edges of the acromion, translated downward parallel to the tendon fibers, with depiction of the tapering-off supraspinatus and infraspinatus tendons. The Transverse view was at a right angle to the previous position, with scanning of the cuff tendons parallel to the anterior edge of the acromion, and depiction of the cuff as a cartwheel, the humeral head as a rim and the rotator cuff as a tube.

The tendons of the rotator cuff normally reveal as a relatively homogeneous structure covering the head of humerus. Criteria for an intact cuff were constant tendon thickness in axial and sagittal planes (Fig. 1). Lesions located within the tendon but not extending to the surface were not considered as tears.

Regarding the shape of the rotator cuff, the following criteria were used for the evaluation of full thickness cuff tears in both transverse and longitudinal planes:

• Complete nonvisualisation of the cuff tendons in case of large or massive full-thickness cuff tears (Fig. 2).

• Localised absence or focal discontinuity of the cuff in smaller tears with consecutive loss of the normal anterior arc of the subdeltoid bursa and deltoid muscle and development of focal concavity of the subdeltoid bursal contour (Fig. 3).

• Sandglass-like diameter shift of the cuff, focal thinning, hypoechogenic discontinuity or nonvisualisation during dynamic examination in small cuff tears (Fig. 4)

• Distinct echogenic foci in combination with distal attenuation were rated as a sign for intratendineal calcifications.

In transverse views, the intra-articular portion of the biceps tendon appears as an oval-shaped echogenic structure adjacent to the humeral head and is located between the supraspinatus tendon and the subscapularis tendon. Because of its relationship to the supraspinatus tendon, it was used as a valuable landmark for identifying the supraspinatus tendon, especially for the location of cuff tears. Extent and location of cuff tears were determined in the transverse plane and were recorded as involving the supraspinatus tendon if only the superior part of cuff was involved and the tear was located posteriorly of the intra-articular segment of the biceps-tendon. If a tear extended into the posterior parts of the cuff, it was recorded as involving both the infraspinatus and supraspinatus tendons and, if it extended also into the anterior parts, it was recorded as a combined supraspinatus and subscapularis tendon rupture. Complete nonvisualisation of the rotator cuff among the whole humeral head was recorded as a massive full-thickness cuff tear with combined supraspinatus, infraspinatus and subscapularis tendon rupture.

US findings were recorded preoperatively and were compared with intraoperative findings with regard to the location of the involved tendons. In the case of different findings, they were recorded as false for calculation of sensitivity and specificity. Sonographically described lesions of the rotator cuff that could not be seen at surgery were defined as false-positive results. All cuff tears not discovered preoperatively were regarded as false-negative results. Additionally, all cuff tears were analyzed with regard to their location and false findings were defined as erroneously-diagnosed locations of cuff tears. The physicians performing surgery were not blinded to the results of US examination because the decision of the surgical procedure was based on the results of clinical examination and US depiction.

In 320 of 332 cases, US diagnosis corresponded with the intraoperative findings. Five false-negative and seven false-positive findings were obtained. A high diagnostic value of preoperative US, with an accuracy of 97% (confidence interval 93.8 to 98.1), sensitivity of 98% (confidence interval 95.1 to 99.3) and a specificity of 93% (confidence interval 85.7 to 97.1) for subacromial pathologies was found. The overall results are summarised in Table 2.

In 91 of 96 cases, preoperative US examination correctly exhibited cuff tears. In 5 cases with false-negative results, tendon tears remained undetected by preoperative US but surgery revealed isolated small partial tears of the supraspinatus tendon located at the bursal side of the tendon, less than 1 cm in diameter.

In 182 of 189 cases of subacromial impingement, a cuff tear was correctly excluded by preoperative US examination. In these 182 patients, no pathologic changes were found by US examination within the rotator tendons and the long head of the biceps tendon; arthroscopy confirmed intact rotator cuff tendons and only signs for bursal irritation were found.

The remaining cases with false-positive results with ultrasonographically depicted full-thickness cuff tears did not show a cuff tear during surgery. In five of these seven cases, US erroneously depicted small tears of the supraspinatus tendon but, during surgery, an intact rotator cuff was found. In two cases with ultrasonographically, depicted full-thickness tears of the supraspinatus and infraspinatus tendons, only severe abrasive changes of the bursal side of these tendons were found, considered as a bursal-sided partial-thickness tear.

In all 51 cases of calcifying tendinitis or synovitis and in three cases of a suprascapularis nerve entrapment, pathological changes and cuff tears were correctly excluded by US examination; no false-positive or false-negative results were observed.

Location of the rotator cuff tear was correctly identified by US examination in 69 of 96 cases with cuff tears (accuracy 88%, sensitivity 80%, specificity 93%). Sensitivity and specificity of US for distinction between the supraspinatus, infraspinatus and subscapularis tendons are shown in Table 3.

No isolated tears of the infraspinatus or subscapularis tendons were found either during US examination or surgery; tears of these tendons only occurred in combination with tears of the supraspinatus tendon.

In 27 cases, preoperative US depiction was incorrect with respect to the location of the involved tendons when compared with the surgical findings (Table 4).

US demonstrated less extensive tears than observed at surgery in 18% and more extensive tears than verified at surgery in 11% of these cases.

In 10 cases, preoperative US overestimated the extent of involved tendons as verified at surgery. Of these, 7 combined supraspinatus and infraspinatus tears only involved the supraspinatus tendon and 3 ultrasonographically-estimated massive tears of all three rotator tendons (supraspinatus, infraspinatus and subscapularis tendon) intraoperatively turned out to be combined supraspinatus and infraspinatus tears.

In 15 cases, preoperative US examination did not identify the complete extent of the involved tendons. Nine cases with ultrasonographically-estimated isolated supraspinatus tears exceeded into the infraspinatus part of the rotator cuff, whereas one spread into the subscapularis part. One isolated supraspinatus tear, as well as one combined supraspinatus/infraspinatus tear and three combined supraspinatus/subscapularis tears, actually turned out to be massive tears involving all three tendons.

Location of the cuff tear was incorrectly identified in two cases. Preoperative US examination revealed a combined supraspinatus and subscapularis tear instead of a combined supraspinatus and infraspinatus cuff tear.

In the literature, shoulder US is described as a sufficient instrument for imaging of the rotator cuff tendons (Hedtmann 1988; Hedtmann and Fett 1995; Hodler et al. 1991; Kurol et al. 1991; Mack et al. 1985; Middleton 1992; Middleton et al. 1985, Middleton et al. 1986; Olive and Marsh 1992; Vick and Bell 1990; Wiener and Seitz 1993). Wide ranges of sensitivity and specificity have been reported in the past (Hedtmann and Fett 1995; Hodler et al. 1991; Kurol et al. 1991; Middleton et al. 1985; Olive and Marsh 1992; Sperner et al. 1993; Turrin et al. 1997; van Moppes et al. 1995; Vick and Bell 1990), due to invalid screening methods and poor technical equipment with inadequate resolution capabilities. Generally accepted ultrasonomorphologic criteria, as well as modern scanning equipment with high-resolution capabilities, lead to high accuracy (Hedtmann 1988; Hedtmann and Fett 1995; Middleton et al. 1985; Vick and Bell 1990). To reduce operator dependency, we prefer a standardized examination protocol with depiction of the supraspinatus, infraspinatus and subscapularis tendons in both axial and sagittal planes. Compared to magnetic resonance (MR) assessment, ultrasonography is more advantageous with respect to the possibility of a dynamic exploration of the rotator cuff (Brenneke and Morgan 1992; Farin and Jaroma 1996; Mack et al. 1988), with enhanced visualisation of the tendons during internal and external rotation.

We found that ultrasonography was highly accurate (accuracy 97%, confidence interval 93.8 to 98.1) and is, therefore, a sufficient instrument for detecting cuff tears, as well as for exclusion of tendon pathologies in cases caused by subacromial impingement, tendinitis calcarea, synovitis and suprascapular nerve entrapment. The high diagnostic value of US for detection of cuff tears in this study is comparable with the experiences of Hedtmann and Fett (1995), Wiener and Seitz (1993), Middleton et al. (1986) and Mack et al (1985). The sensitivity rate of 97.9% (confidence interval 93.8 to 98.1%) is comparable to the rates reported in recent ultrasonographic studies (Brenneke and Morgan 1992; Hedtmann and Fett 1995; Sonnabend et al. 1997; van Holsbeeck et al. 1995), but also with those reported in numerous MR imaging studies. Rates of sensitivity and specificity of these MR studies were only based on series with small numbers of cases ranging between 34 and 71 patients (Kanek et al. 1994; Needell and Zlatkin 1997; Torsten and Hollinsheaed 1999; Wnorowski and McAndrew 1997) and, therefore, are of lower statistical logical value.

All false-positive cases with erroneously depicted cuff tears (n = 7) only concerned the anterior part of the rotator cuff; in particular, the supraspinatus tendon. In the cases with these false-positive depicted tears of the supraspinatus tendon, the rotator interval, described by Neer (1990), and located between the supraspinatus and subscapularis tendons, may cause these erroneous results due to ineffective discrimination of US depiction.

No isolated tears of the infraspinatus or subscapularis tendon were observed in this series of 332 cases. A singular and isolated rupture of the infraspinatus or subscapularis tendon seems to be a very rare entity (Gerber et al. 1996).

The aim of this study was to test the accuracy of US examination for preoperative determination of size and location of rotator cuff tears. Because several arthroscopic methods for the repair of small rotator cuff tears were established, the location and size of rotator cuff tears affect the choice of the method of surgical repair (Neviaser and Neviaser 1984), arthroscopy or open procedure, and it is, therefore, important to classify the location of cuff tears preoperatively.

In this US study, the location and size of the tears was correctly determined by preoperative US with an accuracy of 88%, a sensitivity of 80% and a specificity of 93%.

US demonstrated less extensive tears than observed at surgery in 18% and more extensive tears than verified at surgery in 11% of these cases so that, in general, US seems to tend to underestimate the size of rotator cuff tears. For distinction of the involved tendons of the rotator cuff, shoulder US seems to be highly sufficient for the larger cuff tears, which is in conformity with the experience of Brenneke and Morgan (1992), who reported a higher accuracy of ultrasonography for predicting the size of large tears and a lower one for small tears. Apart from this, the diagnosis of isolated small tears in the anterior part of the cuff was difficult, especially if the supraspinatus tendon was involved, whereas this fact seems to be affected by an ineffective discrimination of US depiction (Farin and Jaroma 1996). High accuracy and specificity were achieved for larger cuff tears, such as combined supraspinatus and subscapularis tendon tears and massive tears with supraspinatus, infraspinatus and subscapularis tendon tears.

Because US tends to underestimate the size of the tear, an arthroscopic procedure could be attempted in cases with small depicted tears of the supraspinatus tendon; however, changing to an open procedure should be possible to facilitate reconstruction of cuff tears enlarged in the infraspinatus or subscapularis tendon.

According to our results, preoperative US depiction is a good screening method for differentiation of subacromial pathologies, especially for distinction between impingment syndrom and full-thickness rotator cuff tears, but seems to be a method of limited value for exact evaluation of size and location of cuff tears. For the characterisation of location and extent of rotator cuff tears, additional MR examination should be done (Fig. 5).