Towards structure-aware surrogate modeling: explicit region interaction improves knee contact stress prediction in young men.

Abstract

Knee contact-stress hotspots are closely linked to meniscal/cartilage injury risk. Still, high-fidelity subject-specific FEA is too computationally expensive for large-cohort, multi-condition, near-real-time use. Existing MeshGraphNet (MGN)-style surrogates mainly rely on stacked local message passing, which is often insufficient for modeling long-range dependencies and limits interpretability. This study benchmarked a deep-stacked MGN baseline against three explicit region-interaction MGN architectures using data from nine young male soccer players performing a 90° change-of-direction task. Under a strict cross-subject evaluation framework, we assessed whole-field error, peak stress fidelity, and hotspot spatial consistency under matched computational budgets. Region-interaction models significantly reduced whole-field nodal stress errors compared to the purely stacked baseline. Crucially, they achieved markedly higher accuracy in reconstructing the high-stress tail and demonstrated superior spatial consistency and temporal robustness in localizing high-risk stress hotspots. These findings from a young-male cohort suggest that explicit region-level interaction may provide a more structure-aligned surrogate modeling paradigm for knee contact mechanics, yielding stronger risk-relevant stress phenotype recovery under comparable computational budgets while supporting more interpretable injury-risk assessment.