Optimization of S2-alar-iliac screw (S2AI) fixation in adult spine deformity using a comprehensive genetic algorithm and finite element model personalized to patient geometry and bone mechanical properties

Purpose: To optimize the biomechanical performance of S2AI screw fixation using a genetic algorithm (GA) and patient-specific finite element analysis integrating bone mechanical properties.

Methods: Patient-specific pelvic finite element models (FEM), including one normal and one osteoporotic model, were created from bi-planar multi-energy X-rays (BMEXs). The genetic algorithm (GA) optimized screw parameters based on bone mass quality (BM method) while a comparative optimization method maximized the screw corridor radius (GEO method). Biomechanical performance was evaluated through simulations, comparing both methods using pullout and toggle tests.

Results: The optimal screw trajectory using the BM method was more lateral and caudal with insertion angles ranging from 49° to 66° (sagittal plane) and 29° to 35° (transverse plane). In comparison, the GEO method had ranges of 44° to 54° and 24° to 30° respectively. Pullout forces (PF) using the BM method ranged from 5 to 18.4 kN, which were 2.4 times higher than the GEO method (2.1-7.7 kN). Toggle loading generated failure forces between 0.8 and 10.1 kN (BM method) and 0.9-2.9 kN (GEO method). The bone mass surrounding the screw representing the fitness score and PF of the osteoporotic case were correlated (R² > 0.8).

Conclusion: Our study proposed a patient-specific FEM to optimize the S2AI screw size and trajectory using a robust BM approach with GA. This approach considers surgical constraints and consistently improves fixation performance.