Soft Tissue Standardization Improves Humerus Tissue Thickness Precision and Lowers BMD in Total Shoulder Arthroplasty Patients

Abstract

DXA technology requires soft tissue attenuation measurement. This sampling is confined within a standard soft tissue region of interest (ROI) defined by the blue squares on Figure 1, which is automatically placed and used to compare with radiation absorption over bone. Variable placement of the soft tissue ROI likely alters BMD measurement and reproducibility. The aim of this study was to measure humeral BMD in total shoulder arthroplasty (TSA) patients. As no shoulder feature is available in GE Lunar enCORE v18, hip atypical femur fracture (AFF) software was utilized. In preliminary evaluation, we observed that soft tissue ROI placement and amount of tissue sampled were inconsistent (Figure 2). We hypothesized that use of a consistent soft tissue ROI would change BMD and improve precision. The purpose of this analysis was to compare humerus BMD and precision prior to and following soft tissue ROI standardization. After manufacturer consultation, all scans were reanalyzed using a soft tissue ROI of a fixed size (100 pixels tall, 50 pixels wide) manually placed on every scan distal and lateral to the humeral head (Figure 3). This study of TSA patients, 1-5 years post-surgery, included full humerus scans obtained on a GE Lunar iDXA with v17 or v18 software and analyzed with v18. Each participant had bilateral shoulder scans performed twice to assess precision and humerus BMD was measured at custom ROIs. BMD and tissue thickness obtained with automated and manual approaches were compared using T-test. Precision was determined using the ISCD calculator and compared between methods with F-test. Thirty subjects were included (20M/10F) with mean age and years’ post-surgery of 70.6 and 2.7 respectively. Average tissue thickness was reduced 13.5% (p < 0.01) with manual tissue sampling. Mean BMD among the nine custom ROIs was 2.2-8.5% lower (p < 0.01) with manual tissue sampling (Table 1). Additionally, tissue thickness %CV improved (p < 0.01) from 47.6 to 14.8% and 25.9 to 9.6%. Similarly, BMD precision improved (p < 0.01) at the 1/3 humerus and distal implant ROIs on the TSA side (Table 2). When measuring skeletal sites without optimized software available, it is necessary to critically evaluate software performance. In this study, soft tissue ROI standardization markedly improved tissue thickness sampling and reduced BMD at custom humerus ROIs. Optimization and automation of humerus DXA software would likely further improve BMD precision and measurement.