In vivo articular cartilage contact at the glenohumeral joint: preliminary report

J Orthop Sci (2008) 13:359–365 DOI 10.1007/s00776-008-1237-3

Original article In vivo articular cartilage contact at the glenohumeral joint: preliminary report PATRICK J. BOYER1, DANIEL F. MASSIMINI1,2, THOMAS J. GILL1, RAMPRASAD PAPANNAGARI1, SUSAN L. STEWART1, JON P. WARNER1, and GUOAN LI1 1 2

Bioengineering Laboratory, Massachusetts General Hospital, Harvard Medical School, GRJ-1215, 55 Fruit Street, Boston, MA 02114, USA Massachusetts Institute of Technology, Cambridge, MA, USA

Abstract Background. Little is known about normal in vivo mechanics of the glenohumeral joint. Such an understanding would have significant implications for treating disease conditions that disrupt shoulder function. The objective of this study was to determine articular contact locations between the glenoid and humeral articular surfaces in normal subjects during shoulder abduction with neutral, internal, and external rotations. We hypothesized that glenohumeral articular contact is not perfectly centered and is variable in normal subjects tested under physiological loading conditions. Methods. Orthogonal fluoroscopic images and magnetic resonance image-based computer models were used to characterize the centroids of articular cartilage contact of the glenohumeral joint at various static, actively stabilized abduction and rotation positions in five healthy shoulders. The shoulder was investigated at 0°, 45°, and 90° abduction with neutral rotation and then at 90° abduction combined with active maximal external rotation and active maximal internal rotation. Results. For all the investigated positions, the centroid of contact on the glenoid surface for each individual, on average, was more than 5 mm away from the geometric center of the glenoid articular surface. Intersubject variation of the centroid of articular contact on the glenoid surface was observed with each investigated position, and 90° abduction with maximal internal rotation showed the least variability. On the humeral head surface, the centroids of contact were located at the superomedial quarter for all investigated positions, except in two subjects’ positions at 0° abduction, neutral rotation. Conclusions. The data showed that the in vivo glenohumeral contact locations were variable among subjects, but in all individuals they were not at the center of the glenoid and humeral head surfaces. This confirms that “ball-in-socket” kinematics do not govern normal shoulder function. These insights into glenohumeral articular contact may be relevant to an appreciation of the consequences of pathology such as rotator cuff disease and instability.

Offprint requests to: G. Li Received: December 25, 2007 / Accepted: March 20, 2008

Introduction The shoulder not only has the widest range of motion of any joint in the body but also the least intrinsic stability1 of any major joint. Thus, joint mechanics have been shown to be significantly affected by ligamentous injury as well as rotator cuff disruption.2 Yet little is known about normal joint mechanics. Because the goal of surgical intervention is to restore function as well as relieve pain, it is important to have an understanding about normal joint mechanics. Accurate knowledge might be important to select the best method of treatment to restore the shoulder to a normal state. Glenohumeral joint biomechanics has been extensively described in the literature. The cartilage morphology, including the thickness and volume, of the glenoid surface and humeral head has been reported in various studies. In general, the glenohumeral joint has been described as a “ball in socket.” Various methods, including plain radiography, computed tomography (CT), and magnetic resonance imaging (MRI), have been used to characterize the relative congruency of the glenohumeral joint.3,4 Shoulder kinematics have been analyzed by measuring humeral head translation and rotation relative to a fixed glenoid surface using various techniques.5–10 Glenohumeral articular contact patterns6,7,11–15 and glenohumeral joint contact pressures16,17 have been examined at different positions of the shoulder in a few in vitro studies using cadaveric specimens. Because of the complicated anatomy of the shoulder complex, which consists of the intercalated joints of the scapulothoracic articulation and the glenohumeral joint, combined motion patterns occur during active shoulder motion; thus in vitro experiments can simulate only a limited approximation of shoulder motion under controlled laboratory studies that attempt to create ideal conditions.7,10,15,18,19 Even though muscle forces have already been applied to simulate physiological shoulder

360

functions in various in vitro studies, the physiological loads of the glenohumeral joint during in vivo shoulder function are still unknown. Moreover, no in vitro studies have accounted for the normal scapulothoracic contribution to overall shoulder motion. This is why normal and pathological glenohumeral articular contact kinematics during in vivo shoulder motion has not been reported in the literature. Recently, we have developed a combined dualorthogonal fluoroscopic imaging and MRI-based modeling technique that can be used to investigate in vivo articular cartilage contact kinematics of musculoskeletal joints.20 The objective of this study was to analyze six degrees-of-freedom (DOF) in vivo kinematics to clarify glenohumeral contact in the shoulders of normal subjects using this unique measurement tool. Specifically, the glenohumeral articular contact kinematics of normal shoulders was quantified during active abduction combined with neutral, internal, and external rotations. We hypothesized that normal glenohumeral articular contact is not centered during shoulder motion under physiological loading conditions, and that there is variation among normal subjects.

Materials and methods Five normal shoulders (two left and three right) from five male subjects (mean ± SD age 26 ± 4 years) were recruited for this study by an orthopedic surgeon under the guidance of our institutional review board. Informed consent was obtained from each participant. All subjects had healthy shoulder function with a complete range of motion and no previous history of trauma. The choice of left or right shoulder was randomly made by the subjects themselves. The shoulder was scanned with a 1.5-T magnet (GE, Milwaukee, WI, USA) using a FIESTA (fast image employing steady-state acquisition) sequence. The MRI scan created a cubic viewing volume of approximately 16 cm on each side, which included a complete view of the humeral head and scapula. Parallel sagittal plane images of the shoulder at 1.0-mm intervals were acquired with a resolution of 512 × 512 pixels. The MR images were used to construct three-dimensional (3D) models of the humeral head and scapula as well as the cartilage surface models of the glenoid and the humeral head in a 3D solid modeling software (Rhinoceros; Robert McNeel & Associates, Seattle, WA, USA). The bony surfaces of the humerus and the scapula together with the cartilage surfaces of the humeral head and the glenoid were digitized within each image. The digitized spatial data (X, Y, Z coordinates) were linked using B-Spline curves to reproduce the contours of the humerus, the scapula, and the carti-

P.J. Boyer et al.: In vivo shoulder biomechanics

lage of the humeral head and the glenoid. Bicubic BSpline surfaces were then created using the contour lines to construct geometric models of the shoulder. After MRI scanning, the subject was positioned inside a dual-orthogonal fluoroscopic system to capture the active shoulder motion during abduction in neutral rotation and then with internal or external rotation. The subject was protected using lead aprons in the whole body except the shoulder being tested. A goniometer was used to control the shoulder abduction angles. Two fluoroscopes (OEC 9800; GE) were positioned in an orthogonal manner so the subject could stand upright and the shoulder could be positioned in both imaging zones of the fluoroscopes. The GE fluoroscope has a clearance of about 800 mm between the X-ray source and the intensifier, which allows the subject to be imaged by the fluoroscopes simultaneously as the subject performs abduction and rotation activities of the shoulder. The shoulder was first imaged at 0° abduction (neutral rotation) of the forearm (with the forearm vertical to the ground) while the subject stood in a relaxed position (Fig. 1A). The subject then lifted the forearm to 45° of shoulder abduction combined with neutral rotation (no axial rotation of the forearm with respect to the longitudinal axis of the humerus), and the shoulder position was imaged by the dual-orthogonal fluoroscopic system. In this study, the abduction angle was measured in the body plane and with respect to the vertical direction to the ground. Similarly, the shoulder was imaged at 90° of abduction, neutral rotation (Fig. 1B). In these three shoulder positions, the elbow angle was kept at 0°. With the shoulder being kept at 90° abduction, the shoulder was then actively externally rotated maximally about the longitudinal axis of the humerus with a 90° elbow angle and imaged by the two fluoroscopes (Fig. 1C). This position was similar to the cocking phase of the throwing motion of the forearm, and it is the position in which the shoulder is at risk for anterior instability. Finally, the shoulder was actively rotated internally to the maximum position about the humeral longitudinal axis, while the shoulder was kept at 90° abduction with a 90° elbow angle, to represent the opposite extreme position to the cocking phase of the throwing motion. At each position, the shoulder was scanned from proximal-anterior and proximal-posterior directions; therefore, two orthogonal images were obtained. These positions were chosen for study because they simulate not only positioning of the shoulder for a throwing motion but also functional daily living activities such as combing one’s hair. To reproduce in vivo shoulder positions using 3D shoulder models, a virtual dual-fluoroscope system20 was created in the solid modeling software (Rhinoceros; Robert McNeel & Associates). The orthogonal images

P.J. Boyer et al.: In vivo shoulder biomechanics

361

Fig. l. Dual fluoroscopic imaging system shown with a subject. A 0° abduction neutral rotation. B 90° abduction neutral rotation. C 90° abduction with maximum external rotation

Fig. 2. Virtual dual fluoroscopic imaging system with kinematically accurate placement of reconstructed shoulder and scapula models from magnetic resonance imaging

were placed in the software to reproduce the positions of the image intensifiers of the two fluoroscopes during image acquisition. Two virtual perspective cameras, corresponding to the positions of the two X-ray sources of the fluoroscopes during image acquisition, were positioned to represent two virtual X-ray sources. Next, the 3D shoulder model was imported into the “virtual” dual fluoroscopic system and was viewed from two orthogonal directions by the two virtual cameras. The positions

of the humerus and scapula were manually adjusted inside the software in six DOF until their projections matched the outlines of the fluoroscopic images (Fig. 2). Thus, the in vivo shoulder position was reproduced. This procedure was repeated to reproduce each in vivo shoulder position tested. Therefore, glenohumeral kinematics during the abduction motion was determined from the models matched at each shoulder position. The accuracy of this image matching method to deter-

362

P.J. Boyer et al.: In vivo shoulder biomechanics

Fig. 3. A Polar coordinate system on the left glenoid surface. B Spherical coordinate system superimposed on the left humeral head relative to the native anatomy

mine 3D joint positions in space has been reported in our previous work,20 where the method was shown to have an accuracy of 0.1 mm in determining translation and 0.1° in determining rotation in space using standard geometries. Similar accuracy was obtained using human cadaveric knees. This series of models that reproduced the motion of the shoulder was then used to determine the motion of the glenohumeral cartilage contact as a function of shoulder positions. When the glenohumeral position was determined, the positions of the cartilage layers of the glenoid and humerus were also determined. The overlap of the cartilage surfaces of the glenoid and humeral head defined the overlap region that was used to calculate the centroid of contact. This centroid of the glenohumeral contact area was defined as the contact location in this study. To quantify the location of the contact point on the glenoid surface, a polar coordinate system was created on the glenoid surface (Fig. 3A). The glenoid surface was vertically positioned, and a superoinferior (SI) axis was defined by direct visualization as the Y-axis collinear with the lateral border of the scapula. The origin of the coordinate system was placed at the midpoint of this axis. An anteroposterior (AP) axis perpendicular to the SI axis was defined as the X-axis. The location of a contact point was defined by its distance from the origin (R, radius) and the angle (α, alpha) subtended with the X-axis. Similarly, the articular contact location on the humeral head was defined by two angles in a spherical coordinate system (Fig. 3B). A sphere of idealized

radius was fit to the humeral head, and the origin of the coordinate system (noted as “O”) was chosen at the center of the humeral head. The Z-axis was defined along the long axis of the humeral shaft; the X-Y plane was perpendicular to the long axis of the humerus shaft, and the X-Z plane was perpendicular to the glenoid surface at the neutral position (0° abduction) of the shoulder. Therefore, for a contact point (noted as “P”) on the humeral head, its azimuthal angle θ was defined as the angle between the projection of the line O–P on the X-Y plane and the positive X-axis. The colatitude angle ϕ was defined as the angle between the line O–P and the positive Z-axis. In this way, the contact points on the humeral head surface can be uniquely defined by an azimuthal angle and the colatitude angle. To present the left and right shoulder data in a consistent manner, all right shoulders were mirrored onto left glenoids and humeral heads. We reported the articular contact location on the glenoid surface using the distance from the glenoid center and the angle subtended with the posterior axis (X-axis) for each patient. The contact point on the humeral head was reported using an azimuthal angle and a colatitude angle.

Results Contact on the glenoid fossa Patient-specific glenoid contact locations of the glenohumeral joint at various shoulder positions are shown

P.J. Boyer et al.: In vivo shoulder biomechanics

363

Fig. 4. Patient-specific glenohumeral articular cartilage centroid locations on the humeral surface and individual glenoid surfaces. Inset Histogram of discrete quadrant contact on the glenoid surface for all positions tested

Table 1. Radial distances and angles Radial distance for subjects 1–5 Position Glenoid Radius (mm) 0° 45° 90° 90° MaxER 90° Max IR Alpha α (°) 0° 45° 90° 90° MaxER 90° Max IR Humerus Azimuth θ (°) 45° 90° 90° MaxER 90° Max IR Colatitude ϕ (°) 0° 45° 90° 90° MaxER 90° Max IR

1

2

3

4

5

7.8 6.4 12.3 7.0 12.5

3.8 5.9 3.1 7.4 7.9

3.6 4.1 4.1 4.2 10.9

7.5 8.7 10.0 7.0 8.8

8.5 4.9 5.8 8.1 3.1

4.6 147.3 271.5 97.4 277.4

34.7 46.7 331.3 103.3 348.2

8.7 82.2 139.2 124.5 294.3

279.8 295.4 299.2 307.8 317.8

91.8 127.4 115.9 258.7 340.5

0.6 −32.4 −24.9 83.1 18.9

−5.9 −61.8 −10.2 60.7 17.7

−23.2 51.5 31.3 22.7 −1.5

−32.2 −45.5 −54.1 −67.1 79.3

−6.1 82.1 56.3 62.2 18.5

72.4 44.1 49.2 5.6 52.8

84.3 37.3 44.9 22.7 59.8

104.6 49.3 38.8 31.9 63.4

106.0 82.6 45.5 47.1 14.3

73.2 11.8 20.0 57.6 49.0

MaxER, maximal external rotation; Max IR, maximal internal rotation

in Fig. 4, and radial distances and angles are presented in Table 1. All shoulders had glenohumeral contact locations away from the glenoid center. The average radial position away from the center of the glenoid for

all positions was more than 5 mm away from the origin. Abduction from 0° to 90° neutral rotation generally exhibited an inferior translation on the glenoid surface. High interpatient variability was observed. Contact was almost exclusively in an arc across the superior AP quadrant and inferoposterior (IP) quadrant except for one contact point in the inferoanterior (IA) quadrant. All patients during 90° abduction with combined maximal internal rotation exhibited contact in the IP quadrant, representing 20% of the total contact. Contact on the humeral head For all abduction positions (except two 0° abduction contact points) and the combined internal and external rotation positions, the glenohumeral contact locations were located at the superomedial quarter sphere of the humeral head surface (Fig. 4). Abduction of the humerus from 0° to 45° neutral rotation exhibited a superior translation of contact on the humeral surface corresponding to an average change in colatitude of 43° consistent with the abduction observed in the body plane. Continued abduction from 45° to 90° neutral rotation showed no appreciable change in average colatitude. Internal rotation showed tighter subject grouping of contact location on the superoanterior portion of the humeral surface compared to external rotation, which had more variability across the superior AP surface. Individual patient contact locations on the humeral surface represented by a colatitude and azimuthal angle are presented in Table 1.

364

Discussion This study utilized a newly developed dual-orthogonal fluoroscopic system20 to investigate the in vivo articular cartilage contact kinematics of the glenohumeral joint during shoulder abduction and rotation motion. In doing so we proved our hypothesis that the normal glenohumeral joint contact centroid is not centered during any movement of the arm into abduction and internalexternal rotation positions. Furthermore, there is a variation among normal subjects. This confirms that “ball-in-socket” kinematics do not seem to be the norm. It also raises the question of what is normal joint mechanics. The variation in contact points may reflect differences in joint congruity with a variable radius of curvature, or it may reflect differences in ligamentous anatomy, with some individuals having tighter anterior structures and others having more lax structures; thus, the effect on humeral translation may be different in each individual. To our knowledge, this is the first in vivo clarification of the range of normal glenohumeral articular joint contact, demonstrating that kinematics are not “ball-and socket” and to some degree are variable among normal subjects. It also confirms that there is a coupled translation-rotation that occurs as the shoulder is moved into progressive abduction and external rotation. Contemporary knowledge of shoulder biomechanics is mostly based on cadaveric experiments.5,6,9,10,14–18,21 Owing to its small articular contact surface and wide range of coupled multiplanar motion,22,23 investigation of shoulder biomechanics, especially under in vivo loading conditions, presents a significant challenge in the biomedical engineering field. Many in vitro and in vivo radiographic studies have considered that the humeral head was perfectly centered in the glenoid fossa when the arm was abducted,3,4,23,24 and in vitro glenohumeral contact measurements has positioned the joint in such way.17 The cartilage at the glenoid center has been shown to be the thinnest25,26 and is associated with the thickest subcortical trabecula.17,27 In an in vitro study, Conzen et al. found different types of articular contact patterns, with some joints exhibiting a central and some a bicentric (superoinferior) distribution at abduction.16 Soslowsky et al., investigating cadaveric shoulders, noted that the contact area moved on the glenoid surface when the arm was abducted with combined external rotation.15 This demonstrated that the glenohumeral joint contact locations moved during isolated abduction across the glenoid surface and more frequently in the anterior portion of the glenoid surface. The glenohumeral articular contact kinematics we observed in this in vivo study showed different patterns from those reported from cadaveric in vitro investi-

P.J. Boyer et al.: In vivo shoulder biomechanics

gations. This may be explained by the differences between in vivo and in vitro loading conditions as well as the young, living shoulders versus old, cadaveric shoulder specimens. Most of the in vitro studies have utilized shoulders from old adult donors, some of which might have had arthritis or rotator cuff deficiency, thus representing altered articular kinematics compared to the healthy shoulders of our study. In addition, in many in vitro investigations, all soft tissues had to be removed down to the capsule or major muscles during specimen preparation.10,15,17 Even in the cases where muscles and capsule were preserved, the specimens were tested under various simplified loading conditions without simulation of scapulothoracic motion and other extrinsic shoulder muscles.9 The data gained from this study represent a physiological analysis of normal shoulder biomechanics and may have important implications for some of the assumptions made about normal glenohumeral joint biomechanics. These observations raise some interesting questions about how surgeons affect joint mechanics when treating rotator cuff disease, instability, and arthritis. For example, it is not clear if repair of damaged ligaments or tendons restores normal kinematics to the glenohumeral joint or if restoration of such normal mechanics correlates with functional recovery. Moreover, little is known about the consequence of joint reconstruction with shoulder arthroplasty and the change in contact between joint surfaces with such a procedure. In this study, the glenohumeral articular contact kinematics were demonstrated using a polar coordinate system. As illustrated by the data, this type of presentation may better show the contact motion along the glenoid surface. Similarly, a spherical coordinate system can clearly delineate the contact motion on the humeral head surface. However, it should be noted that we studied only normal shoulder kinematics without distinguishing sex and dominance differences. Moreover, no extra load-bearing conditions other than forearm weight were considered in this study. All shoulders were investigated under a static condition. With further development of the image acquisition and processing techniques, the articular cartilage contact of the glenohumeral joint under dynamic loading conditions will be investigated, such as lifting a weight, pushing or pulling an object, and throwing motion. More subjects should also be included so the sex and dominance effects can be delineated. It should also be noted that the contact location reported in this article represents the centroid of the cartilage overlap area and not that of the actual contact area. Nevertheless, the data obtained in this study may help design future research of the shoulder joint with or without pathology.

P.J. Boyer et al.: In vivo shoulder biomechanics

Conclusion This study examined the glenohumeral articular contact kinematics in normal subjects who performed abduction and internal-external rotation of the shoulder. We demonstrated that “ball-in-socket” kinematics do not seem to be the norm, and the pattern of coupled translation with rotation is variable among normal individuals. This work provides a basis for future study to clarify the consequences of loading the shoulder during activities such as throwing. Furthermore, we may be able to clarify the effect of rotator cuff disease and instability on joint mechanics, thereby allowing us to analyze the consequences of surgical treatment. Acknowledgments. This work was supported by a research grant from the NFL Charity Foundation and the Department of Orthopaedic Surgery, Massachusetts General Hospital. Technical assistance by Lou “Dawg” DeFrate and Elizabeth “Betsy” Desouza are greatly appreciated.

References 1. Lippitt S, Matsen F. Mechanisms of glenohumeral joint stability. Clin Orthop 1993;291:20–8. 2. Cole BJ, Warner JJP. Anatomy, biomechanics, and pathophysiology of glenohumeral instability. In: Iannotti JP, Williams GR, editors. Disorders of the shoulder: diagnosis and management. Philadelphia: Lippincott Williams & Wilkins; 1999. p. 207–32. 3. Engin AE, Peindl RD. On the biomechanics of human shoulder complex. I. Kinematics for determination of the shoulder complex sinus. J Biomech 1987;20:103–17. 4. Högfors C, Sigholm G, Herberts P. Biomechanical model of the human shoulder: I. Elements. J Biomech 1987;20:157–66. 5. Harryman DT 2nd, Sidles JA, Clark JM, McQuade KJ, Gibb TD, Matsen FA 3rd. Translation of the humeral head on the glenoid with passive glenohumeral motion. J Bone Joint Surg Am 1990; 72:1334–43. 6. Howell SM, Galinat BJ. The glenoid-labral socket: a constrained articular surface. Clin Orthop 1989;243:122–5. 7. Kelkar R, Wang VM, Flatow EL, Newton PM, Ateshian GA, Bigliani LU, et al. Glenohumeral mechanics: a study of articular geometry, contact, and kinematics. J Shoulder Elbow Surg 2001; 10:73–84. 8. Lösch A, Eckstein F, Haubner M, Englmeier KH. A non-invasive technique for 3-dimensional assessment of articular cartilage thickness based on MRI. Part 1. Development of a computational method. Magn Reson Imaging 1997;15:795–804. 9. Poppen NK, Walker PS. Forces at the glenohumeral joint in abduction. Clin Orthop1978;Sept(135):165–70. 10. Wuelker N, Schmotzer H, Thren K, Korell M. Translation of the glenohumeral joint with simulated active elevation. Clin Orthop 1994;309:193–200.

365 11. Greis PE, Scuderi MG, Mohr RA, Bachus KN, Burks RT. Glenohumeral articular contact areas and pressures following labral and osseous injury to the anteroinferior quadrant of the glenoid. J Shoulder Elbow Surg 2002;11:442–51. 12. Iannotti JP, Gabriel JP, Schneck SL, Evans BG, Misra S. The normal glenohumeral relationships: an anatomical study of one hundred and forty shoulders. J Bone Joint Surg Am 1992;74: 491–500. 13. Scherrer PK, Hillberry BM. Piecewise mathematical representation of articular surfaces. J Biomech 1979;12:301–11. 14. Soslowsky LJ, Flatow EL, Bigliani LU, Mow VC. Articular geometry of the glenohumeral joint. Clin Orthop 1992;285:181–90. 15. Soslowsky LJ, Flatow EL, Bigliani LU, Pawluk RJ, Ateshian GA, Mow VC. Quantitation of in situ contact areas at the glenohumeral joint: a biomechanical study. J Orthop Res 1992;10: 524–34. 16. Conzen A, Eckstein F. Quantitative determination of articular pressure in the human shoulder joint. J Shoulder Elbow Surg 2000;9:196–204. 17. Warner JJ, Bowen MK, Deng X-H, Hannafin JA, Arnoczky SP, Warren RF. Articular contact patterns of the normal glenohumeral joint. J Shoulder Elbow Surg 1998;7:381–8. 18. Apreleva M, Hasselman CT, Debski RE, Fu FH, Woo SL-Y, Warner JJP. A dynamic analysis of glenohumeral motion after simulated capsulolabral injury: a cadaver model. J Bone Joint Surg Am 1998;80:474–80. 19. Debski RE, McMahon PJ, Thompson WO, Woo SL-Y, Warner JJP, Fu FH. A new dynamic testing apparatus to study glenohumeral joint motion. J Biomech 1995;28:869–74. 20. Li G, Wuerz TH, DeFrate LE. Feasibility of using orthogonal fluoroscopic images to measure in vivo joint kinematics. J Biomech Eng 2004;126:314–8. 21. Edelson JG. Patterns of degenerative change in the glenohumeral joint. J Bone Joint Surg Br 1995;77:288–92. 22. McQuade KJ, Smidt GL. Dynamic scapulohumeral rhythm: the effects of external resistance during elevation of the arm in the scapular plane. J Orthop Sports Phys Ther 1998;27:125–33. 23. Poppen NK, Walker PS. Normal and abnormal motion of the shoulder. J Bone Joint Surg Am 1976;58:195–201. 24. Howell SM, Galinat BJ, Renzi AJ, Marone PJ. Normal and abnormal mechanics of the glenohumeral joint in the horizontal plane. J Bone Joint Surg Am 1988;70:227–32. 25. Graichen H, Jakob J, von Eisenhart-Rothe R, Englmeier K-H, Reiser M, Eckstein F. Validation of cartilage volume and thickness measurements in the human shoulder with quantitative magnetic resonance imaging. Osteoarthritis Cartilage 2003;11: 475–82. 26. Yeh LR, Kwak S, Kim Y-S, Chou DSW, Muhle C, Skaf A, et al. Evaluation of articular cartilage thickness of the humeral head and the glenoid fossa by MR arthrography: anatomic correlation in cadavers. Skeletal Radiol 1998;27:500–4. 27. Schulz CU, Pfahler M, Anetzberger HM, Becker CR, MüllerGerbl M, Refior HJ. The mineralization patterns at the subchondral bone plate of the glenoid cavity in healthy shoulders. J Shoulder Elbow Surg 2002;11:174–81.