Introduction: Successful shoulder arthroplasty is based on restoration of the individual’s proximal humeral morphology with a precise osteotomy of the humeral head at the level of the anatomical neck. The objective of this study was to determine the geometry of the articular portion of the humeral head in contact with the glenoid in the neutral position and compare the orientation to the geometry of the humeral head determined using the cartilage/calcar interface of the anatomical neck.

Methods: An intact rotator cuff and joint capsule were exposed for six cadaveric full arms. Precision perspex reference cubes were attached to the greater tuberosity of the humerus and to the coracoid process of the scapula on each specimen. Each shoulder was mounted in a custom built jig with the arm fixed in the neutral position and a Microscribe 3D-X digitizer used to digitize three faces of each precision cube. The shoulder joint was then disarticulated and both the humerus and scapula re-mounted on the same jig, independently. The cube faces were re-digitized and relevant points, lines and surfaces were identified and digitized on each humerus and scapula. The humeri were then scanned using a high precision surface laser scanner.

The data collected from both digitizing tools were merged into the same coordinate system and graphically represented. Paired Student’s t-tests were used to compare the inclination and retroversion angles for the two techniques.

Results and discussion: The study found a significant difference in inclination (p less than 0.02) and no difference in retroversion (p equal to 0.75) when the glenoid position was used to calculate humeral head orientation (Inclination: Mean 11.5 deg., StD. 11.2 deg.; Retroversion: Mean 20.5 deg., StD. 6.6 deg.) when compared to using the cartilage/calcar interface (Inclination: Mean 134.1 deg., StD. 1.9 deg.; Retroversion Mean 21.7 deg., StD. 13.9 deg.).

Small deviations in the recovery of head orientation in shoulder arthroplasty may impact on the longevity of an implant. The differences in inclination and retroversion noted in this study may alter the load on the glenoid and/or rotator cuff mechanism in joint replacement. Further research is necessary.

Introduction: Accurate recovery of humeral head geometry in shoulder arthroplasty is an important requirement for a good functional outcome. Despite this, spherical prosthetic components are implanted when the total proximal humerus is described as ovoid¹. However, 60 to 80 % of the head is spherical¹. If, in the normal glenohumeral joint, only the spherical portion is in contact with the glenoid then recovery of normal mechanics is likely with a spherical prosthetic component.

Contact patterns have been examined ex vivo² under static conditions but do not reflect the likely in vivo contact pattern under dynamic loading and have not been correlated to changes in sphericity of the articular surface. A recent study of the distal femur found that thickness of normal articular cartilage is positively correlated with loading³ and, thus, contact.

The objective of this study was to determine the feasibility of using a surface laser scanner to determine cartilage thickness and, therefore, likely contact area and to correlate changes in thickness to changes in sphericity of the articular surface.

Methods: A cadaveric arm without bony deformity or evidence of rotator cuff disease was dissected free of soft tissue and mounted on a rigid block within the frame of a surface laser scanner (Kestrel3D Ltd., UK). The articular surface of the humerus was scanned at a resolution of 200 μm. The cartilage was then dissolved away and the humerus re-scanned. The x,y,z coordinate data of the re-scanned bone were used to match the sub-chondral bone with the cartilage from the previous scan using Pointstream™ software (Kestrel3D Ltd., UK).

The cloud point data for the cartilage and bony surfaces were exported into modelling software (McNeal and Assoc., Seattle, WA) and the surface area of the head divided into ten equal sections. For each slice of both the cartilage and bony surface, the radius of curvature was calculated using a least square fit optimisation technique⁴. The differences in radius of curvature between the cartilage surface and subchondral bone surface were used to calculate the cartilage thickness for each slice. The standard deviation from the radius of curvature was used to calculate the degree of deviation from sphericity.

Results: For the first 60 % of the surface area, the deviation from sphericity was 0.5% of the radius with a cartilage thickness of 0.74 mm. The deviation from sphericity and cartilage thickness for 100% of the articular surface was > 1% and 0.63 mm, respectively.

Conclusions: The experiment proved that the surface laser scanner can be used to elucidate the relationship between contact patterns and articular curvature of the proximal humerus. The changes in sphericity concur with results from previous studies¹. Assuming cartilage thickness correlates to contact patterns at the normal glenohumeral joint, the change in cartilage thickness suggests that contact may occur only at the spherical portion of the head. Knowledge of this relationship may aid in future prosthetic design considerations or in modification of the osteotomy technique. To further support these findings, a 50μm laser scanner is being developed and will be used on a larger sample size.