Detection of Anterior and Posterior Glenoid Bone Defects After Reverse Shoulder Arthroplasty With Tomosynthesis in a Pig Model.

Background Scapular notching is a well-recognized complication of reverse shoulder arthroplasty (RSA) and is typically identified at the inferior scapular neck on conventional radiographic evaluation. However, anterior and posterior glenoid bone defects around the baseplate may be difficult to detect using standard imaging modalities, including radiographic evaluation or computed tomography (CT). Tomosynthesis provides sectional imaging with reduced metal artifact and may improve the detection of region-specific glenoid bone defects after RSA. Methods The scapula of a slaughter pig was used in this study. Delta XTEND® implants (DePuy, Warsaw, IN, USA) were used to create four models: anterior, posterior, and inferior glenoid bone defect models, and an intact control model. Images obtained by radiographic evaluation, CT, and tomosynthesis were independently reviewed by 12 orthopedic surgeons to assess the presence or absence of glenoid bone defects. Detection sensitivity and specificity were calculated for each imaging modality. Results For anterior and posterior glenoid bone defect models, the sensitivities and specificities of radiographic evaluation were 25%/100% and 8.3%/94.4%, respectively. CT demonstrated sensitivities and specificities of 25%/86.1% and 33.3%/86.1%, respectively. In contrast, tomosynthesis demonstrated a sensitivity of 83.3% and a specificity of 94.4% for both anterior and posterior glenoid bone defects. The sensitivity of tomosynthesis was markedly higher than that of radiographic evaluation and CT. Conclusions In this controlled ex vivo model, tomosynthesis enabled improved detection of anterior and posterior glenoid bone defects compared with conventional radiography and CT. Although these findings are limited to an experimental setting, tomosynthesis may serve as a useful adjunctive imaging modality for characterizing region-specific glenoid bone defects after RSA.