Phantomless calibration of glenoid trabecular bone density: improved accuracy using age-stratified reference tissue densities.

Evaluation of glenoid trabecular bone quality is important for patients undergoing shoulder arthroplasty, as the density of the glenoid bone is likely to influence implant stability. Routine CT scans show promise for quantitative bone density assessments, but CT-based calibration typically involves scanning a phantom, which limits its application in clinical practice and in retrospective research studies. Here, we developed and validated a phantomless BMD measurement technique to evaluate glenoid trabecular bone density (mg/cc) using clinical shoulder CT scans of 88 patients (48 males, 40 females; mean age: 53 years). Air, fat, and muscle tissue were used as internal density references to develop patient-specific calibration functions for glenoid BMD estimation. The phantomless calibration was validated by comparison with an asynchronous phantom-based calibration using the European Spine Phantom. A Bland-Altman analysis yielded a mean difference of 0.37 mg/cc (95% CI: [-1.39, 2.14] mg/cc) between the two methods. Further improvement in internal calibration accuracy was achieved by implementing an age-stratified approach (two age groups; cut-off 55 years) to account for age-related declines in muscle density. With age stratification of reference tissue densities, the mean difference between internal and phantom-based calibration was reduced to 0.20 mg/cc (95% CI: [-1.38, 1.78] mg/cc). The proposed phantomless calibration approach provides a robust and clinically feasible method to evaluate glenoid trabecular bone density in both research and clinical applications. Moreover, the results highlight the importance of considering patient-specific variations in shoulder muscle density to ensure accurate internal calibration.