Background/Purpose: Patients with inflammatory rheumatic and musculoskeletal diseases (iRMDs) are at elevated risk for osteoporotic fragility fractures. Standard assessment via dual-energy X-ray absorptiometry (DXA) does not fully capture fracture risk. Cortical Backscatter (CortBS) ultrasound is a radiation-free technique assessing viscoelastic and microstructural cortical bone properties and has shown high accuracy in detecting prior fractures in postmenopausal women. Bone material strength index (BMSi), derived from reference-point indentation, reflects mechanical bone properties. However, their comparative performance in fracture discrimination in iRMDs is unclear. This study aimed to evaluate the performance of CortBS in detecting prevalent fragility fractures in iRMD patients, compared to DXA and BMSi.

Methods: Consecutive iRMD patients from our outpatient osteoporosis clinic, with and without prior fragility fractures and no prior anti-osteoporotic treatment, were analyzed. CortBS was measured at the anteromedial tibia using a custom ultrasound scanner. DXA-based T-scores and BMSi (via 8 indentations) were obtained at the same visit. Fracture discrimination models for vertebral, non-vertebral, and any fragility fracture were developed using multivariate partial least squares linear discrimination analysis with Leave-One-Out Cross-Validation (PLS-LDA LOOCV). For each method, the area under the curve (AUC) of the receiver operating characteristics (ROC), sensitivity, and accuracy were calculated, with and without anthropometric data (weight, height, BMI). CortBS-BMSi correlation was assessed using PLS regression with threefold cross-validation.

Results: 97 iRMD patients (mean age 64 ± 10 years) were included: 42% with rheumatoid arthritis, 23% spondyloarthritides, and 19% connective tissue diseases. 44 patients (45%) had fragility fractures (19 vertebral, 34 non-vertebral, 9 both). Complete CortBS data were available in 40 patients (15 with fractures). CortBS showed comparable discrimination to DXA for any fracture (AUC 0.61 vs. 0.60) and non-vertebral fractures (AUC 0.68 vs. 0.69). For vertebral fractures, CortBS (AUC 0.66) outperformed both DXA (0.43) and BMSi (0.45), though small sample size limits interpretation. Adding anthropometric data improved prediction for all tools (CortBS AUC 0.73, DXA 0.75, BMSi 0.62). No significant correlation was found between CortBS and BMSi (R² = 0.05, p = 0.18).

Conclusion: CortBS demonstrated fracture discrimination comparable to DXA and even superior performance for vertebral fractures. In contrast, BMSi showed limited discriminatory ability. As a radiation-free and accessible technique, CortBS holds promise for improving fracture risk assessment in iRMDs. Larger and longitudinal studies are needed to evaluate whether CortBS can enhance fracture prediction by capturing cortical porosity and structural bone properties.

Figure 1: ROC-AUC with 5-fold cross-validation for any (left), non-vertebral (center) and vertebral (right) fragility fractures for DXA, BMSi and CortBS. AP (anthropometric data: weight, height, body mass index)

Figure 2: Predictions of aBMD at spine (A), femur neck (B), and total femur (C) based on ultrasound backscatter and anthropometric parameters using multivariate PLS regression Spearman correlation

« Back to ACR Convergence 2025

ACR Meeting Abstracts - https://acrabstracts.org/abstract/comparison-of-cortical-backscatter-ultrasound-reference-point-indentation-and-bone-mineral-density-in-discriminating-prevalent-fragility-fractures-in-inflammatory-rheumatic-diseases-a-pilot-study/