Abstract
The purpose of this study is to quantify the distribution of bone density in the scapulae of patients undergoing reverse shoulder arthroplasty (RSA) to guide optimal screw placement. To achieve this aim, we compared bone density in regions around the glenoid that are targeted for screw placement, as well as bone density variations medial to lateral within the glenoid.
Specimen included twelve scapula in 12 patients with a mean age of 74 years (standard deviation = 9.2 years). Each scapula underwent a computed tomography (CT) scan with a Lightspeed+ XCR 16-Slice CT scanner (General Electric, Milwaukee, USA). Three-dimensional (three-D) surface mesh models and masks of the scapulae containing three-D voxel locations along with the relative Hounsfield Units (HU) were created. Regions of interest (ROI) were selected based on their potential glenoid baseplate screw positioning in RSA surgery. These included the base of coracoid inferior and lateral to the suprascapular notch, an anterior and posterior portion of the scapular spine, and an anterosuperior and inferior portion of the lateral border. Five additional regions resembling a clock face, on the glenoid articular surface were then selected to analyze medial to lateral variations in bone density including twelve, three, six, and nine-o'clock positions as well as a central region. Analysis of Variance (ANOVA) tests were used to examine statistical differences in bone density between each region of interest (p < 0 .05).
For the regional evaluation, the coracoid lateral to the suprascapular notch was significantly less dense than the inferior portion of the lateral border (mean difference = 85.6 HU, p=0.03), anterosuperior portion of the lateral border (mean difference = 82.7 HU, p=0.04), posterior spine (mean difference = 97.6 HU, p=0.007), and anterior spine (mean difference = 99.3 HU, p=0.006).
For the medial to lateral evaluation, preliminary findings indicate a “U” pattern with the densest regions of bone in the glenoid most medially and most laterally with a region of less dense bone in-between.
The results from this study utilizing clinical patient CT scans, showed similar results to those found in our previous cadaveric study where the coracoid region was significantly less dense than regions around the lateral scapular border and scapular spine. We also have found for medial to lateral bone density, a “U” distribution with the densest regions of bone most medially and most laterally in the glenoid, with a region of less dense bone between most medial and most lateral. Clinical applications for our results include a carefully planned trajectory when placing screws in the scapula, potentially avoiding the base of coracoid. Additionally, surgeons may choose variable screw lengths depending on the region of bone and its variation of density medial to lateral, and that screws that pass beyond the most lateral (subchondral) bone, will only achieve further purchase if they enter the denser bone more medially. We suspect that if surgeons strategically aim screw placement for the regions of higher bone density, they may be able to decrease micromotion in baseplate fixation and increase the longevity of RSA.
