Radiol Oncol 2000; 34(2): 234-13.
Magnetic resonance arthrography of the glenohumeral joint
Vladimir Jevtic, Anton Praprotnik
Clinical Radiology Institute, University Medical Centre, Ljubljana, Slovenia
Background. The purpose of this study was to estimate the value of native magnetic resonance imaging (MRI) and direct gadopentate (Gd-DTPA) magnetic resonance artrography (MRA) in diagnosing posttraumatic pathological changes of the glenohumeral joint.
Patients and methods. In 27 consecutive patients with clinically significant trauma in whom native MRI of the glenohumeral joint was not diagnostically conclusive direct MRA with paramagnetic contrast agent was accomplished. Following intraarticular injection of Gd-DTPA diluted in saline (concentration 0.2 mmol/l) T1W spin echo (SE) and gradient echo (GE) images were performed. Diagnostic results of native MRI and MRA were compared.
Results. MRA revealed 5 rotator cuff lesions, 7 labral lesions, 4 osteochondral injuries and a loose body which were not definitely proved by native MRI. AH the pathological findings of MRA were confirmed operatively ar arthroscopically.
Conclusion. Direct MRA with Gd-DTPA showed to be a more sensitive technique far the demonstration of clinically significant intraarticular posttraumatic changes than noncontrast MRI.
Key words: magnetic resonance imaging; shoulder joint - injuries; rotator cuff
Introduction
Owing to multiplanar imaging capabilities, to superb contrast and to high spatial resolution MRI is able to demonstrate noninvasively joint anatomy. Consequently, this modality has become the method of choice for joint imaging. However, MRI diagnosis of different intraarticular pathologic conditions in anatomically complex joints, such as the gleno-
Received 15 May 2000 Accepted 30 May 2000
Correspondence to: Prof. VladimirJevtič, M.D., Ph.D., Clinical Radiology Institute, University Medica! Centre Ljubljana, Zaloška 7, SI-1525 Ljubljana, Slovenia. Phone: +386 1 143 15 30; Fax: +386 1 133 31 0444
humeral joint, may be difficult. Therefore several techniques of MRA were evaluated. These include intraarticular use of saline or Gd-DTPA (direct arthrography) and the intravenous application of Gd-DTPA (indirect arthrography). An important advantage of direct ar-thrography is the joint capsule distension which enables the better visualisation of intra-articular anatomic structures especially in tight joints. According to some authors MRA of the shoulder joint may provide some information of diagnostic importance about the labrum, the joint capsule and the rotator cuff which are not possible without the use of Gd-DTPA.1^ Vast majority of centres still use non-contrast MRI as a routine diagnostic method.
146
Jevtic Vet nl. / Magnetic resonance arthrography
The purpose of this study was to estimate the value of native MRI and direct Gd-DTPA MRA in diagnosing posttraumatic pathological changes of the glenohumeral joint and eventually to include MRA in a routine diagnostic work up.
Patients and methods
On the basis of history and physical examination the clinical diagnosis of rotator cuff tears (n=18) and the lesions of the labum and the capsule (n=9) was made following significant trauma. In all of these patients native MRI was performed 0.5 -2 months following trauma. Magnetic resonance images were obtained using dedicated shoulder coil on 1.5 T scanner (Magnetom SP 63, Siemens, Erlangen, Germany). A SE sequence with Tl-weighted (TR 450 ms, TE 15 ms), PD-weight-ed (TR 1900 ms, TE 20 ms) and T2-weighted (TR 1900 ms, TE 80 ms) images was performed using angled coronal plane, followed by a GE Tl-weighted FISP 2D images (TR 40 ms, TE 10 ms, flip angle 40°) in an axial plane. In all the patients the MRI diagnosis was inconclusive (3 possible rotator cuff lesions, 2 suspect labral lesions and 2 possible osteochondral injuries). Due to persisting clinical symptoms MRA was indicated and accomplished 0,5 - 1 months after the first examination. The joint was punctured under fluoroscopic control using local anaesthesia. The proper position of the needle was controlled with 1-2 ml of non-ionic iodinated contrast material (Ultravist, Schering, Germany) and 10-15 ml of 2 mmol/1 Gd-DTPA ¡solution in saline (Magnevist, Schering, Germany) was then injected until the capsule became distended. After the injection passive and active movements of the shoulder joint were carried out for several minutes to achieve the uniform distribution of Gd-DTPA solution. Coronal oblique and axial 5 mm thick consecutive slices were obtained using
Radiol Oncol 2000; 34(2): 145-50.
a SE Tl-wieghted images (TR 450 ms, TE 15 ms), a GE FISP 2D images (TR 40 ms, TE 10 ms, flip angle 40°). A field of view was 16 cm with data acquisition matrix of 256 x 256. (MRA has been approved by the Republic's Ethics Committee).
The images were analysed in a qualitative fashion and the diagnostic results of native MRI and MRA were compared.
Results
Out of 27 cases MRI revealed equivocal pathological findings in 7 patients (26%). A diagnosis of possible full thickness rotator cuff lesion was made in 3 cases, of suspect labral injury in 2 cases and of osteochondral lesions in 2 cases. In the remainder of 21 (78 %) patients no pathological findings could be demonstrated (Tablel).
MRA confirmed rotator cuff lesions in all 3 cases with equivocal findings on MRI and proved 2 additional tears. In 3 of these patients the diagnosis of complete (Figure 1) and in 2 cases of partial rotator cuff rupture was made (Figure 2). Typical Bankart lesions were demonstrated in 7 patients using MRA, 2 of these were the cases with suspect labral lesions on native MRI (Figure 3). In patients with labral injuries 3 osteochondral (Sach -Hill) lesions of the posterolateral aspect of the humeral heads were revealed. MRA showed a loose body within the joint (Figure 4) which was not seen on native MRI (Tablel). In all the patients pathological MRA findings were confirmed by the operative procedure or the shoulder artroscopy.
Discussion
MRI replaced the conventional arthrography in most institutions and has become a method of choice for depicting the posttraumatic joint pathological changes. Using different imaging
Jevtic Vet al. / Magnetic resonance arthrography
147
Table l. MRI and MRA findings
Diagnostic modality		MRI		MRA
Diagnosis	possible		definite	possible definite
	N	%	n %	n % n %
normal findings			20 74	8 30
pathological findings	7	26		19 70
rotator cuff lesions	3	11		5 19
labral lesions	2	7,5		7 26
osteochondral lesions	2	7,5		3 11
loose bodies			o	1 3
frozen shoulder			o	3 11
Fi^we l. MRA, T1W spin echo image in an oblique coronal plane. Complete rotator cuff rupture is demonstrated with accumulation of paramagnetic contrast agent at the insertion of the m. supraspinatus terdon (big arrow heads). Leakage of contrast material into the subacromial-subdeltoid bursa is also revealed (small arrow heads).
protocols intra and paraarticular soft tissue structures can be demonstrated noninvasively with high contrast and spatial resolution in various planes. It has been proved that MRI shows high sensitivity and specificity for diagnosing the most of the clinically relevant pathological changes of the knee joint. Unfortunately the diagnostic results in some
other anatomically tight joints, including glenohumeral, were not so good. In an attempt to increase the diagnostic accuracy of MRI several MRA methods were evaluated changing noninvasive native MRI in an invasive diagnostic procedure. The best results were achieved by the intraarticular application of gadopentate diluted in saline.1-5
Radiol Oncol 2000; 34(2): 145-50.
148
Jevtic Vet al. / Magnetic resonance arthrography
Figure 2. MRA, TIW spin echo image in an oblique coronal plane. Partial rotator cuff rupture. Small defect of the m. supraspinatus tendon is filled with paramagnetic contrast material (arrow head).
In our study native MRI revealed equivocal pathological changes only in 7 joints (26 %) out of 27 patients with significant clinical symptoms. In none of these MRI diagnosis was definite. There was a number of cases (n=20; 74 %) with normal findings in which the clinical symptoms could not be explained by MRI. On the other hand MRA proved to be much more accurate and revealed definite, clinically relevant pathological changes in 19 (70 %) of the shoulder joints resulting in an operative procedure which confirmed the MRA diagnosis.
Although the diagnostic results of MRA were much better than that of MRI in 8 (30 %) out of 27 cases with normal MRA findings clinical symptoms could not be substantiated by imaging features. It may be postulated that
at least some of minor posttraumatic changes were spontaneously resolved in a period of several months following the initial trauma before MRA was accomplished, while some of them affected extraarticular anatomic structures not shown by MRA.
In all 3 cases with MRI diagnosis of possible full thickness rotator cuff tear MRA clearly demonstrated typical pathological findings, and proved 2 additional partial lesions not shown by MRI. MRI diagnosis was mainly based on T2-weighted SE images which revealed high signal intensity synovial fluid within the tendon of the supraspinatus muscle and an effusion within the subacromial -subdeltoid bursae. Using MRA in cases with complete tear the defect as well as the quality of neighbouring ends of the tendon were
Figure 4. MRA, TIW spin echo image in an oblique coronal plane. Low signal intensity loose body is shown (arrow) surrounded by high signal intensity contrast material.
Radiol Oncol 2000; 34(2): 145-50.
Jevtic Vetal./ Magnetic resonance arthrography
149
clearly outlined by the high signal intensity contrast agents on Tl-weighted SE images (Figure 1). Leaking of the contrast agent into the subacromial and subdeltoid bursae was also demonstrated. In two joints with partial tear a small defect of the tendon was filled with high signal intensity gadopentate (Figure 2). There are several reasons for the superior diagnostic results of MRA. Firstly, the joint capsule distension resulted in better filling of the thorn tendon and in leaking of contrast agent out of the joint cavity. 3A'6 Secondly, higher contrast resolution was achieved with the solution of gadopentate in comparison to synovial fluid in native MRI.1'2'7 Thirdly, the spatial resolution on Tl-weighted SE images in MRA was higher than on T2-weighted images which were diagnostic in MRI.8'9
Even greater differences between both examinations were documented in labral lesions. Only 2 possible labral lesions were
shown by MRI, mainly using GE Tl-weighted FISP 2D images. MRA proved each of these lesions and confirmed additional 5 Bankart lesions. Capsular distension with contrast accumulation within the capsulolabral lesions was the main reason for the better demonstration with MRA (Figure 3). MRI revealed 2 Sach-Hill lesions as high signal intensity areas on T2-weighted SE images and low signal intensity on Tl-weighted SE images. The diagnosis of 3 osteochondral lesions with MRA was mainly based on the contrast accumulation within the defect of the humeral head.
Low signal intensity loose body representing cortical bone fragment was clearly demonstrated within a distended joint due to the surrounding high signal intensity gado-pentate only using MRA (Figure 4).
In 3 cases the diminished capacity of the joint cavity with reduced bursae and the same
Radiol Oncol 2000; 34(2): 145-50.
150
Jevtic Vet al. / Magnetic resonance arthrography
leakage of the contrast agent out of the gleno-humeral joint indicated the frozen shoulder. Native MRI was not able to demonstrate these typical findings
There were no cases with pathological MRI findings which were not confirmed by MRA.
No adverse reactions to contrast agents or a complication due to intraarticular injection could be registrated.
In conclusion, in this study MRA proved to be more accurate than native MRI in diagnosing clinically relevant posttraumatic intraar-ticular changes of the glenohumeral joint. The results indicate that MRA may be used as a safe and reliable routine diagnostic procedure.
References
1.	Zanetti M, Hodler J. Contrast media in MR artrog-raphy of the glenohumeral joint: intra-articular gadopentetate vs saline; preliminary results. Eur Radiol 1997; 35: 498-502.
2.	Kopka L, Funke M, Fischer U, Keating D, Oestmann J, Grabbe E. MR arthrography of the shoulder with gadopentetate dimeglumine: influence of concentration, iodinated contrast material, and time on signal intensity. AJR 1994; 163: 621-3.
3.	Vahlensieck M, Peterfy CG, Wischer T, Sommer T, Lang P, Schlippert U, et al. Indirect MR Arthrography: Optimisation and clinical applications. Radiology 1996; 200: 249-54.
4.	Helgason JW, Chandnani VP, Yu JS. MR Arthrography: A review of current technique and applications. AJR 1997; 168: 1473-80.
5.	Palmer WE, Caslowitz PL, Chew FS. MR Arthrography of the shoulder: Normal intraarticular structures and common abnormalities. AJR 1995; 164: 141-6.
6.	Flannigan B, Kursunoglu-Brahme S, Snyder S, Karzel R, Del Pizzo W, Resnick D, MR Arthro-graphy of the shoulder: comparison with conventional MR imaging. AJR 1990; 155: 829-32.
7.	Hodler J, Kusunoglu-Brahme S, Snyder SJ, Cervilla V, Karzel RP, Schweitzer ME, et al. Rotator cuff disease: assessment with MR arthrography versus standard MR imaging in 36 patients with arthro-scopic confirmation. Radiology 1992; 182: 431-6.
8.	Palmer WE, Brown JH, Rosenthal DI. Rotator cuff: evaluation with fat-suppressed MR arthrography. Radiology 1993; 188: 683-7.
9.	Palmer WE, Brown JH, Rosenthal DI. Labral-liga-mentous complex of the shoulder: evaluation with MR arthrography. Radiology 1994; 190: 645-51.
10.	Chandnani VP, Yeager TD, DeBerardino T, Christensen K, Gagliardi JA, Heitz DR,et al. Glenoid labral tears: prospective evaluation with MR imaging, MR arthrography, and CT arthrogra-phy. Am J Roentgenol 1993; 161: 1229-35.
Radiol Oncol 2000; 34(2): 145-50.