Key result
Statistical shape models (SSMs) and statistical intensity models (SIMs) outperformed combined SSIMs (e.g., shape generalization error <1 mm vs 2.2 mm) for biomechanical surgical planning.
Why the study?
Shoulder arthroplasty patients with bone erosion are challenging to treat and would benefit from improved surgical planning.
Using statistical shape models and statistical intensity models in conjunction generates synthetic bone models with realistic characteristics for biomechanical surgical planning in shoulder arthroplasty.
May enhance synthetic bone modeling for shoulder arthroplasty planning; extends statistical approaches but leaves open clinical validation.
This work developed, validated, and compared statistical shape, statistical intensity, and statistical shape and intensity models (SSMs, SIMs, SSIMs) of scapulae from a clinical population. SSMs efficiently describe bone shape variation while SIMs describe bone material property variation, and SSIM's combine description of both variables. This work establishes these models' efficacy and whether they can be used in surgical planning. Models were developed using shoulder arthroplasty data of patients with bone erosion, which is challenging to treat and would benefit from improved surgical planning. Models were created using previously validated nonrigid registration and material property assignment processes that were optimized for scapula characteristics. The models were assessed using standard metrics, anatomical measurements, and correlation analyses. The SSM and SIM specificity and generalization error metrics were 3.4 mm and <1 mm and 184 HU and 156 HU, respectively. The SSIM did not achieve the same level of performance as the SSM and SIM in this study (e.g., shape generalization: SSIM-2.2 mm versus SSM-<1 mm). Anatomical correlation analysis showed that the SSM more effectively and efficiently described shape variation compared to the SSIM. The SSM and SIM modes of variation were not strongly correlated (e.g., rmax = 0.56 for modes explaining ≤2.1% of variance). The SSIM is outperformed by the SSM and SIM and the latter two are not strongly correlated; therefore, using the SSM and SIM in conjunction will generate synthetic bone models with realistic characteristics and thus can be used for biomechanical surgical planning applications.
No takes yet. Share an insight, caveat, or question.
Sharif-Ahmadian et al. (2023) studied Bone erosion requiring shoulder arthroplasty. Statistical shape models (SSMs) and statistical intensity models (SIMs) vs. Statistical shape and intensity models (SSIMs) was evaluated on Model specificity and generalization error. Statistical shape models (SSMs) and statistical intensity models (SIMs) outperformed combined SSIMs (e.g., shape generalization error <1 mm vs 2.2 mm) for biomechanical surgical planning.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: