Abstracts

Diagnostic imaging of shoulder rotator cuff lesions

Diagnóstico por imagem nas rupturas do manguito rotador

Marcello Henrique Nogueira-Barbosa^I; José Batista Volpon^II; Jorge Elias Jr^III e Gerson Muccillo^IV

^IMaster and Assistant Doctor

^IIAssociate Professor.

^IIIMáster and Professor

^IVPhD Professor

Address for correspondence

SUMMARY

Shoulder rotator cuff tendon tears were evaluated with ultrasonography (US) and magnetic resonance imaging (MRI). Surgical or arthroscopical correlation were available in 25 cases. Overall costs were also considered. Shoulder impingement syndrome diagnosis was done on a clinical basis. Surgery or arthroscopy was considered when conservative treatment failure for 6 months, or when rotator cuff repair was indicated. Ultrasound was performed in 22 patients and MRI in 17 of the 25 patients. Sensitivity, specificity and accuracy were 80%, 100% and 90.9% for US and 90%, 100% and 94.12% for MRI, respectively. In 16 cases both US and MRI were obtained and in this subgroup statistical correlation was excellent (p< 0.001). We concluded that both methods are reliable for rotator cuff full thickness tear evaluation. Since US is less expensive, it could be considered as the screening method when rotator cuff integrity is the main question, and when well trained radiologists and high resolution equipment are available.

Key words: Ultrasound; magnetic resonance imaging; rotator cuff introdução

RESUMO

Foram avaliadas rupturas de espessura total nos tendões do manguito rotador do ombro por ultra-sonografia (US) e ressonância magnética (RM). A correlação cirúrgica ou artroscópica foi obtida em 25 casos. A US foi realizada em 22 pacientes e a RM em 17 ombros sintomáticos. Foram obtidas sensibilidade, especificidade e eficiência de 80%, 100% e 90,9% para a US e 90%, 100% e 94,12% para a RM, respectivamente. No subgrupo de pacientes submetidos a ambos os métodos diagnósticos (16 casos), a correlação estatística foi excelente (p< 0,001). Concluímos que US e RM são confiáveis na avaliação de rupturas de espessura total do manguito rotador. Quando a principal questão é a integridade do manguito rotador, a US pode ser considerada como o exame de investigação inicial, devido a sua disponibilidade e aos custos,na presença de radiologistas treinados e de equipamentos de alta resolução.

Descritores: Ultra-som; ressonância magnética; manguito rotador

INTRODUCTION

Rotator cuff, particularly supraspinous tendon slide at the space between humeral head and the choraco-acromial arc (choracoid, choraco-acromial ligament and acromion) and can have fiction to these structures during arm elevation. Microtraumas resulting from repetitive movements end causing inflammation of the tendons that belong to the rotator cuff and sub-acromial and sub-deltoidal bursae above located. A chronic and degenerative process of these tendons takes place during lifetime.

Diagnosis of impingement of rotator cuff (PMR) is got from medical history, physical examination including specific tests, and, when necessary, tests of symptoms relief by local anesthesia by subacromial injection. Diagnosis of PMR is not by radiograph. The role of radiograph is to try identifying the cause of PMR, the presence and extension of possible tendon injuries. Presence of a tear of the cuff is not always easy to be clinically identified and is an important point for decision on operative or conservative treatment⁽¹⁷⁾. The degree of retraction and the quality of the ruptured borders, and the quality of the muscle itself also affect the therapeutic decision.

Plain radiographs of the shoulder are important to evaluate anatomical features such as format and inclination of the acromion, the presence of acromial spines, os acromiale and acromio-clavicular ostearthrosis. Radiographs are as well useful for demonstration of secondary impingement changes, nevertheless they are non specific such as subcondral cysts, sclerosis, and irregularities of the humeral tuberculum. Calcification of tendons and surrounding them can also be identified.

Shoulder arthrograph is an efficient diagnosis method for detecting rotator cuff discontinuities by demonstrating a communication between gleno-umeral joint and subdeltoid-subacromial bursae, however it is an invasive method and subject to false-negative results, in cases of partial tear in superficial part of the cuff.

Currently, non invasive methods are available such as ultrasonography and magnetic resonance (Figures 1 and 2).

This work aimed to evaluate the reproducibility of results previously achieved in literature, the applicability of both non-invasive methods, and, particularly among us, considering the cost/benefit relationship.

MATERIALS AND METHOD

Were included adult patients, both male and female, clinically diagnosed PMR, surgically treated and followed-up at our outpatient clinic. This work was approved by Institutional Ethics Committee.

Clinical diagnosis was based on classical findings of pain at anterior and lateral shoulder aspects, worsening with arm elevation and night pain. At physical examination, there was generally pain when humeral tuberculae were palpated, some degree of movement restriction and pain on stressing maneuvers⁽⁹⁾.

Only patients with symptoms at one single side and with surgical treatment indication were included. This indication was based on conservative treatment failure during six months, in cases of suspected tear of rotator cuff or in recidivated cases when the patient had several times underwent conservative treatment without succeeding.

Once surgery was indicated, patients underwent ultrasound examination of both shoulders and/or MIR examination of symptomatic shoulder.

Were excluded from this analysis those patients who refused surgical treatment or those in whom it was not possible to perform at least one of the diagnosis methods.

Pre-operative examinations were performed with a ultra-sound ACUSON 128XP 10 equipment with a linear dynamic transductor of 7.0 MHz and a MRI equipment Vision SIEMENS (1.5 Tesla), with a surface reel.

Criteria used for diagnosing tear of one or both tendons of rotator cuff in US were: a) one or more cuff tendon(s) was not visible b) focal non visibility of one of the tendons and c) defect of well defined discontinuity of the tendon, reaching both tendon aspects (superficial and deep).

Criteria used for tear of rotator cuff in MRI were a) complete of focal absence of the cuff, b) hypersignal, similar to water signal weighted in T2, crossing all cuff substance, that is, well defined interruption of the tendon, whether or not associated to retraction.

During the surgery, open or by arthroscopy, the rotator cuff was evaluated aiming to identify fiber ruptures as well as degenerative changes. Surgical technique comprised subacromial decompression⁽⁹⁾ and repair of tendon lesions, when present.

Twenty-five patients were included over a two years period. Two cases were explored, exclusively, by arthroscopy (one of them with a previous exploration two years before) and two of them by arthroscopy and open surgery. All remaining were explored exclusively by open surgery.

Were obtained for each diagnosis method the sensitivity, specificity and efficacy in detecting rotator cuff tears. Were also calculated the sensitivity and specificity for each one of the ultrasonography deployed diagnosis criteria. An agreement test between the methods was performed using the group of patients in whom both exams were performed, with calculation of the kappa(K) coefficient⁽³⁾.

A cost assessment of both exams was performed, checking private radiology services from the city as well as the value reimbursed by two different health insurances and by the governmental health system in the same period (Table 1).

RESULTS

Over a period of two years comprised by the study, 25 patients underwent surgical treatment, 20 females and 5 males, 22 of them with US and 17 with MRI, being 16 patients preoperatively submitted to both exams.

The age of the patients at the time of the surgery ranged from 18 to 78 years (average 48.7 years; median 50 years).

The time between the exams and the surgery ranged from 1 day to 10.8 months for US (average 4 months) and for MRI 24 days to 8.6 months (average 3.8 months).

In total 22 cases were evaluated with subacromial decompression and US, with a sensitivity of 80% and specificity of 100%, with an efficiency of 90.9% (Figures 1, 2, 3, 4 and 5).

In the sub-group of the 17 patients who underwent MRI and surgery, it was found a sensitivity of 90%, specificity of 100% and efficiency of 94.12% in detecting tears.

In the sub-group of 16 patients who underwent both US and MRI it was found that there was an excellent correlation between both exams, with a K (kappa) coefficient of 0.87 and p=0.000221. In only one case ultrasonography was not able to identify a complete tear of the rotator cuff diagnosed by MRI. Both could correctly identify all intact cuffs.

Costs for each of the examinations are in (Table 1), including medical fees and operational costs.

DISCUSSION

It uses to be hard to get a consensus on the best option for study of rotator cuff impingement, even after several published studies, if we take into consideration both efficacy and costs of each method.

Development of high resolution equipments as well as high frequency transductors changed the efficacy of ultrasonography. For MIR the same took place, with the surge of better series and faster surface reels.

In recent reviews on this subject, the value of ultrasonography in evaluating the rotator cuff was well established. It was emphasized the need of experienced examiners for performing and interpreting this method, with suitable equipments⁽¹⁵⁾.

Nevertheless, this is not routinely used in many centers, particularly in United States. In the North-American concept, this method takes a time that is considered as an exaggeration from the radiologist both in learning as in performing the examination, as well as on the other hand there is a large availability of MRI equipments. Most of ultra-sound examinations are performed in that country by specialized non medical technicians (sonographers). The apprenticeship curve for performing musculoskeletal ultra-sound examinations is long since it is mandatory an specific knowledge of anatomy and physiopathology. There is a trend to less trust when the exams are not performed by a doctor, such as happens in that country. In European countries, ultrasonography is a first-line imagery exam for many soft tissue pathologies.

Mid eighties, three different groups performed studies stressing the usefulness of ultrasnongraphy in evaluation of the rotator cuff^(2,6,7).

Tears of rotator cuff can be of different shapes and sizes and this reflects in different features in ultrasonography⁽⁸⁾. In wide tears of the rotator cuff, the tendon will not be seen and sub-deltoide bursa will close to humerus head surface, with apposition of the deltoid, which partially replaces the cuff (Figures 3 and 4). Larger lesions may appear as a local absence of the cuff (Figure 5). A cuff discontinuity is observed when a smaller defect is filled with articular anechoic fluid, or by reactive granulation tissue, that is hypoechoic (Figure 6). Focal changes of the echo of these tendons should be very carefully interpreted, in general corresponding to a anisotropy feature or a degenerative change.

A literature review displays a wide variation of ultrasonography results in identification of rotator cuff tears, with reports of sensitivity superior to 90%, however with sensitivity reports of less than 70%^(1,2,6,7,10). Differences in equipments resolution and in studied populations as well as in experience and interest of the examiners, variability of the diagnosis criteria used, and difference in the diagnosis information desired are crucial factors for explaining the variation of published results.

Our series showed sensitivity and specificity of 80% and 100% respectively, slightly below the literature, supporting US as a method for evaluation of cuff integrity.

During the period of data collection examiners were under training, even though with previous experience in musculoskeletal exams, and this can explain the observed difference.

Several papers demonstrated MRI to be useful in evaluating rotator cuff. The sensitivity and specificity of this method in full thickness tears, that is, those in that fiber discontinuity reaches both deep and superficial aspects of the tendon, range respectively from 82% to 100% and from 79% to 100%^{(1,4,10,11,12,18).} In the present paper similar results to literature were found, with 90% and 100% of sensitivity and specificity, respectively (figures 7 and 8). Changes in sign from rotator cuff observed in short time and echo sequences (T1 and proton density) corresponding to an intermediate signal area, lower than water signal, in sequences weighted in T2 represented degenerative changes of the tendons and not a fibers rupture (Figures 9 and 10).

MRI not only is able to demonstrate how much tendons are retracted, but also allows to evaluate muscle atrophy for its multi-layer capacity and due to the intense sign of the fat tissue, in these cases in T1 scans (Figure 11).

The observed high specificity of MRI and ultrasonography observed in this work is probably exaggerated. The explanation for this is probably the absence of partial tears in our series, more difficult to be identified. Partial tears are those that are restricted to one of the tendon aspects. In the literature few studies included specific analysis of these injuries. A work willing to adequately evaluate partial tears should include evaluation of both cuff aspects, meaning an open or arthroscopic surgery in each individual case.

Partial tears importance, and imagery methods efficacy in this evaluation are still under discussion ^{(10,11,12,13,16)}.

Nevertheless arthroscopy and surgery are frequently a pattern in this kind of study, they also have limitations. Small healed injuries may not be seen. Arthroscopy vision has limitations specially in tendon insertion, where smaller lesions are not always found, since they can be faded by sinovial changes⁽¹¹⁾.

In our work, literature data were well reproduced for both diagnostic methods. The fact it was better for MRI is accordingly to a longer learning curve for US. Correlation between both, in the subset of patients who underwent both exams, was excellent (p< 0.001).

The issue cost/benefit of each test can not be avoided nowadays. Choice of diagnosis method for rotator cuff tears, was, in practice, being guided according to individual experience and personal preference of each radiologist, without dogmatic recommendations in the literature⁽¹⁴⁾.

STILES & OTTE ⁽¹⁴⁾ and JACOBSON & van HOLSBEECK ⁽¹⁶⁾ stressed this question and, as well as we did in our cost evaluation, found a better relation between cost and benefit for ultrasonography.

SEIBOLD et al. ⁽¹³⁾, in a review paper on this subject conclude by suggesting that in their opinion US is more limited in regard of examiner's experience while MRI is more limited in regard of access restrictions.

Cost relation between MRI/US ranged from 5.42 to 15.76, averaging 7.85.

CONCLUSION

It was possible to reproduce international literature results in our evaluation of rotator cuff tears. The correlation between both methods was statistically excellent. We conclude that both are trustful in evaluation of full thickness tears of shoulder rotator cuff.

Due to a better cost/benefit relationship for ultrasonography, we suggest that, given experienced radiologists and high resolution equipment are available, this could be the method of choice for an initial evaluation.

Address for correspondence

Av.Bandeirantes, 3900

CEP 14048-900 Ribeirão Preto, SP

E-mail: marcellonog@hotmail.com

Trabalho recebido em 27/07/2002. Aprovado em 28/07/2002

Work performed at Centro de Ciências da Imagem e Física Médica da Faculdade de

Medicina de Ribeirão Preto - USP.

Publication Dates

History