2025 ISAKOS Biennial Congress ePoster
Statistical Shape Model Approach for the Prediction of Premorbid Glenoid Shape in Complex Shoulder Arthroplasty Using Artificial Intelligence
Asma Salhi, PhD, Brisbane, QLD AUSTRALIA
Kristine R. Italia, MD, FPOA, Quezon City, Metro Manila PHILIPPINES
Marine Launay, MEng, Greenslopes, QLD AUSTRALIA
Ignacio Viedma, PhD, Brisbane, QLD AUSTRALIA
Luke Gilliland, BEng, Brisbane, QLD AUSTRALIA
Shanthan Pather, PhD, Brisbane, QLD AUSTRALIA
Jashint Maharaj, MBBS, FRSPH, Brisbane, QLD AUSTRALIA
Kenneth Cutbush, MBBS, FRACS, FAOrthA, Spring Hill, QLD AUSTRALIA
Ashish Gupta, MBBS, MSc, FRACS, FAORTHOA, Brisbane, QLD AUSTRALIA
Queensland Unit for Advanced Shoulder Research (QUASR), Brisbane, QLD, AUSTRALIA
FDA Status Not Applicable
Summary
This study aims to evaluate the accuracy of statistical shape model (SSM) in predicting the premorbid glenoid shape in complex shoulder arthroplasty cases with significant glenoid bone defects using preoperative planning software.
ePosters will be available shortly before Congress
Abstract
Introduction
Preoperative planning in shoulder arthroplasty has demonstrated its value in managing complex cases. Reconstruction of the premorbid glenoid morphology helps surgeons to accurately determine the extent of reconstruction required, with the aim of improving patient outcomes and reducing the risk associated with failure and subsequent revision surgeries. This study aims to evaluate the accuracy of statistical shape model (SSM) in predicting the premorbid glenoid shape in complex shoulder arthroplasty cases with significant glenoid bone defects using preoperative planning software.
Methods
Three-dimensional (3D) models of 113 morphologically healthy scapulae were obtained by manually segmenting computed tomography (CT) scans. Significant bone defects were created for each scapular shape to simulate the bone loss typically encountered in complex cases. SSM-based method was then employed to reconstruct the premorbid shape of the scapula with the glenoid bone loss. Quantitative and qualitative validations were conducted to assess the accuracy of this reconstruction method. Clinical measurements of glenoid centre, version, inclination, and root mean square (RMS) distance metric errors were calculated and compared between the original and SSM-reconstructed scapula. A shoulder subspecialist surgeon performed qualitative validation of each reconstructed scapula.
Results
A mean error of 3.3 mm +/- 2.2 mm was obtained for glenoid centre calculation using the planning software’s premorbid prediction of severe glenoid bone loss. For glenoid version, the mean error was 2.8 degrees +/- 2.4 degrees. For glenoid inclination, the mean error was 3.4 degrees +/- 2.4 degrees. An RMS distance mean error of less than 1 mm was obtained by comparing the reconstructed scapula shapes to the original ground truth scapulae. The qualitative validation demonstrated that all the reconstructed premorbid scapular shapes represented a morphologically accepted scapula shape for clinical use.
Conclusion
The findings of this study highlight the efficacy of preoperative planning techniques utilizing SSM to accurately predict the premorbid glenoid morphology in complex shoulder arthroplasty cases.