The position and orientation of the glenoid component in total shoulder arthroplasty affect glenohumeral joint kinematics, soft-tissue forces, contact pressures, and stresses in the components and bone cement1-6. Over time, abnormal joint mechanics may influence joint stability, component wear, and the propensity for component loosening. Improper positioning of the fixation pegs within the glenoid vault may lead to perforation of the cortical vault wall and may compromise initial fixation7,8.
Surgically, the critical step in determining the final position and orientation of the glenoid implant occurs when the central peg is first drilled. Two different approaches described in the literature for orienting this central peg are (1) perpendicular to the glenoid face (termed the face plane direction in the present study)9-11; and (2) parallel to the scapular body (the neutralization direction), resulting in an implant version angle of 0°8,12-14. The face plane direction is based on replicating the orientation of the glenoid face, whereas the neutralization direction is based on aligning the implant with the overall scapular body.
The glenoid vault geometry beneath the glenoid face, into which the pegs or keel are placed, is likely an important factor to consider in glenoid component implantation. The vault becomes especially narrow in its anterior-posterior dimension, with <10 mm seen in some individuals at a depth of 15 mm beneath the face8,15. Hoenecke et al.7 predicted the rate of perforation of the vault wall for two commercially available, pegged and keeled glenoid component designs, when they are implanted in neutral version, to be 18% and 13%, respectively. Researchers have recently attempted to quantify the complex three-dimensional vault geometry15,16. Codsi et al.16 described a geometric model of the endosteal vault surface consisting of a series of connected transverse triangles, equally spaced in the axial direction. The model was to be used as a basis for alternative implant design.
The purpose of this study was to introduce a method for finding a position and orientation of optimal containment of the central peg within the normal glenoid cortical vault, with use of computer optimization. We also sought to compare, in a series of normal scapular models, this vault centerline with the central drill lines obtained from two existing surgical methods, the face plane and neutralization techniques. We hypothesized that (1) the glenoid vault centerline would be oriented at an angle posterior to the plane of the scapula and (2) the centerline would be closer to the direction obtained by the face plane technique than that obtained by the neutralization technique.
Materials and Methods
Thirty-four patient-specific three-dimensional computer models of full scapulae were created from computed tomography scans. The deidentified scans were obtained from patients from our institution with proximal humeral fractures and no visible glenoid damage or osteoarthritis. The scans had an in-plane pixel dimension of 0.39 to 0.67 mm and slice spacing of 0.5 or 0.6 mm; twenty-seven were from female patients and seven were from male patients. The images were imported into Mimics software (version 13; Materialise, Ann Arbor, Michigan) to generate the models17.
Ten anatomical landmarks were selected on each scapular computer model (Fig. 1) with use of the Mimics graphical interface. P1 through P4 were the four extreme glenoid poles of the glenoid articular surface: P1 was at the junction of the superior glenoid rim and coracoid, P2 was at the junction of the inferior rim and the lateral scapular body as described by Nguyen et al.13, P3 was the most anterior point of the glenoid rim, and P4 was on the posterior rim along a line passing through P3 and approximately perpendicular to the line connecting P1 and P2. Points P5, P6, and P7 defined the glenoid face plane and were points of three different regions of the glenoid rim: P5 was identical to the superior pole P1, P6 was from the inferior-anterior quadrant and was equidistant to both P2 and P3, and P7 was from the inferior-posterior quadrant and was equidistant to both P2 and P4.
Lastly, P8, P9, and P10 were chosen according to Kwon et al.18 and defined the coronal scapular plane. Point P8 was the inferior tip of the scapula, P9 was the center of the glenoid face, and P10 was the intersection of the spine and the medial border of the scapula. These points were selected by two analysts independently, and the reproducibility in key results was determined, as described later.
The glenoid face center point was determined from the four glenoid poles. Each pole point was first projected onto the glenoid face plane. The superior-inferior direction was defined by the line connecting the projected superior and inferior poles, and the anterior-posterior direction was orthogonal to this and parallel to the glenoid face plane. An anterior-posterior horizontal line was placed at the midpoint of the line connecting the superior and inferior poles, and a superior-inferior line was similarly placed at the midpoint of the line connecting the anterior and posterior poles. The intersection of these two lines was the glenoid face center point.
Glenoid Vault Centerline
The original computed tomography scan data were resliced in Mimics with use of planes that were parallel to the glenoid face plane, resulting in sagittal image slices of the glenoid vault spaced 1 mm apart. Slices were obtained for the region spanning from the central glenoid articular surface to 15 mm medial to that location (Fig. 2). This 15-mm vault depth was chosen to include all of the bone that encapsulates the central peg of a standard pegged glenoid component.
Outlining contours of the glenoid cortical vault were determined for each resliced sagittal image with use of a custom semiautomated algorithm implemented in MATLAB software (version 7.6; The MathWorks, Natick, Massachusetts). Initial contours were generated by an edge detection algorithm19, and these contours were then adjusted in some locations to obtain a visibly accurate outline of the vault. For each resulting outlining contour, an anterior-posterior line segment that connected the anterior and posterior edges at the height of the glenoid face center point was calculated (Fig. 2). The midpoint of each line segment was determined. A glenoid vault centerline was fit to all of the midpoints such that the sum of the squares of the distances from the line to each point was minimized.
Glenoid Face Plane and Neutralization Techniques
The three-dimensional directions representing the surgeon’s drill line for the central peg of the glenoid component were found for both the face plane and the neutralization techniques. The face plane direction was perpendicular to the glenoid face plane (Fig. 3). The neutralization direction was parallel to the coronal scapular plane. It was inclined such that it was perpendicular to the line connecting the superior and inferior poles of the glenoid.
Orientation Angles and Comparisons
For each subject, transverse and coronal plane orientation angles of the vault centerline, the face plane direction, and the neutralization direction were determined on the basis of conventions for measuring version and inclination18,20,21. The scapular axis passed through points P9 and P10 (Fig. 1). The transverse scapular plane passed through the scapular axis and was perpendicular to the coronal scapular plane. The angle of the drill direction (for example, the angle of the face plane direction) in the transverse plane was the angle between the drill direction and the scapular axis, measured in the transverse scapular plane. A drill direction located “posterior” to the scapular axis in the transverse plane corresponded to retroversion of the glenoid, i.e., the orientation of the starting drill position was located posteriorly aiming anteriorly (as shown in Figure 3). The angle in the coronal plane was similarly measured.
Shortest Distances to Vault Wall
For each subject, for the vault centerline, face plane, and neutralization drill lines, the shortest distance between the line and the cortical wall contours throughout the 15-mm vault depth was calculated. The face plane and neutralization drill lines were both assumed to initiate at the glenoid face center point.
Source of Funding
There was no external funding for this study.
The glenoid vault centerline, determined from the anterior-posterior midpoints, intersected the glenoid articular surface at an intersubject average (and standard deviation) of 1.1 ± 0.8 mm posterior to the glenoid face center point. The glenoid vault centerline was oriented at an intersubject average of 9.4° ± 5.1° posterior to the scapular axis in the transverse plane (Figs. 3 and 4). The face plane (perpendicular) direction was oriented at an average of 7.3° ± 4.0° posteriorly in the transverse plane, and the neutralization direction was oriented at 0° ± 0° in the transverse plane (by definition). The absolute difference in transverse plane angle between the face plane direction and the vault centerline (Fig. 4), averaged across subjects, was 4.9° ± 3.6° (with >10° in four subjects), whereas this angle between the neutralization direction and the vault centerline averaged 9.6° ± 4.7° (with >10° in seventeen subjects). Coronal plane orientation angles averaged 2.2° ± 4.1° superiorly for the face plane direction and 2.4° ± 4.2° superiorly for the neutralization direction.
The minimum distance between the face plane drill line (throughout the 15-mm depth) and the vault wall averaged 4.4 ± 1.5 mm (minimum, 1.0 mm) across subjects, and the minimum distance between the neutralization direction and the vault wall averaged 4.4 ± 1.5 mm (minimum, 1.4 mm). The minimum distance between the vault centerline and the vault wall averaged 5.1 ± 1.3 mm (minimum, 2.6 mm). Note that these distances do not take into account pegs, other than the center peg, or the radius of the drill-hole.
The sensitivities of results to both choice and interobserver variability in obtaining anatomical landmark coordinates were investigated as follows. First, calculations were rerun with the glenoid face center point set equal to point P9 instead of calculating it from the four glenoid poles. The transverse angle of the vault centerline decreased by an intersubject average of 1.3° ± 1.5°. Second, calculations were rerun for fifteen of the models with all points P1 through P10 identified by a second observer and the vaults reoutlined. The three-dimensional distances between the point coordinates obtained by Observer 1 compared with Observer 2, averaged across all models, were 1.1 mm for P1, 1.3 mm for P2, 1.2 mm for P3, 1.0 mm for P4, 1.1 mm for P5 (same as P1), 1.2 mm for P6, 0.9 mm for P7, 2.6 mm for P8, 0.9 mm for P9, and 3.1 mm for P10. With use of the points of Observer 2 instead of Observer 1, the transverse angle of the centerline decreased by an intersubject average of 0.7° ± 1.5°, whereas the transverse angle of the face plane direction decreased by 0.1° ± 1.5°, and the centerline crossing with respect to the glenoid face center point shifted 0.1 ± 0.6 mm posteriorly compared with the results of Observer 1.
These results show the three-dimensional geometric relationships between (1) the glenoid face plane, (2) the scapular body plane, and (3) the glenoid vault geometry. When orientation angles were averaged across all subjects, the glenoid vault centerline and face plane (perpendicular) direction were oriented at 9° and 7°, respectively, posteriorly from the neutralization direction in the transverse plane. On a subject-by-subject basis, the face plane and neutralization directions were an average of 5° and 10°, respectively, separated from the vault centerline in the transverse plane. That is, the face plane direction and vault centerline had a similar average orientation, which was different from the neutralization direction, but the relationships between these directions varied across subjects. The minimum distances between the vault centerline (throughout the 15-mm depth) and the vault wall were, on the average, larger than those between the center peg drill lines of both surgical techniques and the vault wall, indicating better containment within the vault for the vault centerline.
The vault centerline represents optimal containment of the glenoid central peg within the vault. This study provides the location of this centerline relative to scapular landmarks and relative to the drill directions from the two surgical techniques. Because a centerline or center point is not strictly defined for an amorphous shape such as the glenoid vault, an ad hoc approach was developed to calculate the vault centerline. The centerline was defined as a best-fit line through the midpoints of vault cross-sections, in which these cross sections are parallel to the glenoid face (Fig. 2).
The anterior-posterior width of the cortical vault, measured at a sagittal plane 15 mm medial to the glenoid face, was reported to average only 11.5 mm (range, 5.6 to 15.6 mm) in a series of twenty arthritic patients8. Because of this narrow vault width, consideration of peg placement can be important to avoid vault penetration. Calculating the vault centerline routinely on a patient-by-patient basis unfortunately is not currently feasible because it requires specialized software and three-dimensional scapular modeling. We aimed to quantify the normal glenoid-to-peg relationships as an initial step in understanding these relationships in the diseased glenoid.
There is no clear superior border of the glenoid vault because of its junction with the coracoid process. Because of the ambiguity in the geometric definition of the vault superiorly, and because the thinner anterior-posterior geometry is likely more critical, the vault centerline was calculated only from the anterior-posterior midpoints. In a preliminary analysis, an alternative vault centerline, in which the vault was defined by the heights of the superior and inferior poles or by the vault contours if located within the span of those poles, was calculated. In all but five of thirty-four subjects, the centerline obtained from anterior-posterior midpoints was oriented within 1° of this alternative centerline, leading to our focus on the anterior-posterior midpoints approach and results for transverse plane orientation.
The face plane technique in the present study aligns the implant with the anatomical plane formed by the glenoid face9-11 (Fig. 3). We defined a systematic method for estimating the glenoid face plane that relies on landmarks from the accessible, normal glenoid face. A limitation of the present study is that, in arthritic shoulders, the cartilage and subchondral bone of the glenoid—especially of the posterior aspect—are often eroded or deformed by posterior osteophytes. Landmarks on the posterior aspect of the glenoid are not reliable in such cases. Instead of relying on the glenoid face itself, Matsen et al.9,10 suggested intraoperatively locating a so-called centering point on the glenoid neck in order to estimate the line perpendicular to the glenoid face. The face plane technique described in the present study represents an idealized centering point technique.
The neutralization technique theoretically aligns the glenoid implant in a consistent manner with respect to the scapula, even in cases of glenoid face erosion. However, the mid-glenoid version angle of the native glenoid is known to vary across normal shoulders by >20°20, meaning that, in many shoulders, normal glenoid anatomy is not restored by the neutralization technique. The technique is assisted by appreciating glenoid version preoperatively on plain radiographs or a computed tomography scan, a process susceptible to error22, although recent efforts in computer-assisted surgery have sought to address this limitation13,18,23. Furthermore, mid-glenoid version measurement does not take into account the three-dimensional geometry of the glenoid face, such as the slight superior-inferior spiral twist of the glenoid evidenced by changes in version measured at different axial planes24.
In addition to orientation of the drill lines for glenoid implantation, the surgeon decides the location of the central peg on the glenoid face. The vault centerline intersected the face close to the glenoid center point when averaged across all subjects. This was expected since the centerline was a best-fit line through the vault midpoints, including midpoints at depths near the glenoid face. It should be noted that in the osteoarthritic glenoid, the presence of posterior osteophytes can distort what is the apparent face center. Rispoli et al.8 analyzed computed tomography scans of osteoarthritic glenoids. They calculated the center of the glenoid vault at a depth of 15 mm, and then projected this point to the glenoid face along a direction corresponding to the neutralization direction. This projected point was, on the average, slightly anterior and inferior to the center point of the glenoid face determined externally by the poles of the glenoid.
The primary limitations of this study include the use of computed tomography scans from patients without osteoarthritis. There can be errors associated with the process of creating the computer models of the subjects and obtaining the anatomical landmark positions from those models17,18. The vault centerline concept does not apply to alternative fixation approaches, such as placing the screws or pegs on the basis of cancellous bone density or material strength25, or fixation in the scapula beyond the vault region26. Centerlines for the endosteal surfaces may be calculated in a similar way as described above for the cortical surfaces. Factors associated with surgical error and clinical implementation of the face plane and neutralization techniques were generally not considered and are affected by surgical experience, technique, and available technology.
In conclusion, we described a novel method for calculating a glenoid vault centerline, and two approaches for orienting the glenoid component were compared with this centerline on the basis of the geometries of normal scapulae. The vault centerline representing optimal containment of the glenoid central peg was, on the average, oriented posteriorly 9.4° ± 5.1° from the plane of the scapula. The face plane perpendicular direction was oriented at 7.3° ± 4.0° posteriorly. On the average, the neutralization direction for drilling the glenoid central peg anteverts the glenoid component relative to the vault centerline and the normal glenoid face. Future aims include the assessment of three-dimensional vault containment of various glenoid peg and keel designs and the development of systematic means to estimate the face plane and vault centerline from accessible landmarks on osteoarthritic glenoids.
Investigation performed at the Department of Orthopaedics and Rehabilitation, Penn State Milton S. Hershey Medical Center, Penn State College of Medicine, Hershey, Pennsylvania
Disclosure: The authors did not receive any outside funding or grants in support of their research for or preparation of this work. Neither they nor a member of their immediate families received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity.
1. Nyffeler RW Sheikh R Atkinson TS Jacob HA Favre P Gerber C. Effects of glenoid component version on humeral head displacement and joint reaction forces: an experimental study. J Shoulder Elbow Surg. 2006;15:625–9.
2. Wong AS Gallo L Kuhn JE Carpenter JE Hughes RE. The effect of glenoid inclination on superior humeral head migration. J Shoulder Elbow Surg. 2003;12:360–4.
3. Shapiro TA McGarry MH Gupta R Lee YS Lee TQ. Biomechanical effects of glenoid retroversion in total shoulder arthroplasty. J Shoulder Elbow Surg. 2007;16(3 Suppl):S90–5.
4. Oosterom R Rozing PM Bersee HE. Effect of glenoid component inclination on its fixation and humeral head subluxation in total shoulder arthroplasty. Clin Biomech (Bristol, Avon). 2004;19:1000–8.
5. Farron A Terrier A Büchler P. Risks of loosening of a prosthetic glenoid implanted in retroversion. J Shoulder Elbow Surg. 2006;15:521–6.
6. Hopkins AR Hansen UN Amis AA Emery R. The effects of glenoid component alignment variations on cement mantle stresses in total shoulder arthroplasty. J Shoulder Elbow Surg. 2004;13:668–75.
7. Hoenecke HR Jr Hermida JC Dembitsky N Patil S D’Lima DD. Optimizing glenoid component position using three-dimensional computed tomography reconstruction. J Shoulder Elbow Surg. 2008;17:637–41.
8. Rispoli DM Sperling JW Athwal GS Wenger DE Cofield RH. Projection of the glenoid center point within the glenoid vault. Clin Orthop Relat Res. 2008;466:573–8.
9. Matsen FA 3rd Lippitt SB Sidles JA Harryman DT 2nd. Practical evaluation and management of the shoulder. Philadelphia: WB Saunders; 1994.
10. Matsen FA Rockwood CA Jr Wirth MA Lippitt SB. Glenohumeral arthritis and its management. In: Rockwood CA Jr Matsen FA 3rd, editors. The shoulder. 2nd ed. Philadelphia: WB Saunders; 1998. p 840–964.
11. Meyer NJ Pennington WT Ziegler DW. The glenoid center point: a magnetic resonance imaging study of normal scapular anatomy. Am J Orthop. 2007;36:200–2.
12. Kelly JD Jr Norris TR. Decision making in glenohumeral arthroplasty. J Arthroplasty. 2003;18:75–82.
13. Nguyen D Ferreira LM Brownhill JR Faber KJ Johnson JA. Design and development of a computer assisted glenoid implantation technique for shoulder replacement surgery. Comput Aided Surg. 2007;12:152–9.
14. Iannotti JP Spencer EE Winter U Deffenbaugh D Williams G. Prosthetic positioning in total shoulder arthroplasty. J Shoulder Elbow Surg. 2005;14(1 Suppl S):111S–21S.
15. Bicknell RT Patterson SD King GJ Chess DG Johnson JA. Glenoid vault endosteal dimensions: an anthropometric study with special interest in implant design. J Shoulder Elbow Surg. 2007;16(3 Suppl):S96–101.
16. Codsi MJ Bennetts C Gordiev K Boeck DM Kwon Y Brems J Powell K Iannotti JP. Normal glenoid vault anatomy and validation of a novel glenoid implant shape. J Shoulder Elbow Surg. 2008;17:471–8.
17. Bryce CD Pennypacker JL Kulkarni N Paul EM Hollenbeak CS Mosher TJ Armstrong AD. Validation of three-dimensional models of in situ scapulae. J Shoulder Elbow Surg. 2008;17:825–32.
18. Kwon YW Powell KA Yum JK Brems JJ Iannotti JP. Use of three-dimensional computed tomography for the analysis of the glenoid anatomy. J Shoulder Elbow Surg. 2005;14:85–90.
19. Canny J. A computational approach to edge detection. IEEE Trans Pattern Anal Mach Intell. 1986;8:679–98.
20. Churchill RS Brems JJ Kotschi H. Glenoid size, inclination, and version: an anatomic study. J Shoulder Elbow Surg. 2001;10:327–32.
21. Friedman RJ Hawthorne KB Genez BM. The use of computerized tomography in the measurement of glenoid version. J Bone Joint Surg Am. 1992;74:1032–7.
22. Bokor DJ O’Sullivan MD Hazan GJ. Variability of measurement of glenoid version on computed tomography scan. J Shoulder Elbow Surg. 1999;8:595–8.
23. Kircher J Wiedemann M Magosch P Lichtenberg S Habermeyer P. Improved accuracy of glenoid positioning in total shoulder arthroplasty with intraoperative navigation: a prospective-randomized clinical study. J Shoulder Elbow Surg. 2009;18:515–20.
24. Inui H Sugamoto K Miyamoto T Machida A Hashimoto J Nobuhara K. Evaluation of three-dimensional glenoid structure using MRI. J Anat. 2001;199(Pt 3):323–8.
25. Anglin C Tolhurst P Wyss UP Pichora DR. Glenoid cancellous bone strength and modulus. J Biomech. 1999;32:1091–7.
26. Codsi MJ Bennetts C Powell K Iannotti JP. Locations for screw fixation beyond the glenoid vault for fixation of glenoid implants into the scapula: an anatomic study. J Shoulder Elbow Surg. 2007;16(3 Suppl):S84–9.