A methodology for constructing and assessing a combined experimental and computational validation domain for patient-specific mandibular reconstruction.

Abstract

Preclinical assessment of patient-specific mandibular reconstruction plates relies on bench testing and computational modeling, yet existing setups are typically designed on practical grounds without a structured method to ensure their mechanical relevance to the intended clinical context of use (COU). This study proposes and demonstrates a methodology for constructing and assessing a validation domain with experimental and computational components for patient-specific fibula free flap mandibular reconstruction. A clinical reference finite element model (M-COU) representing the reconstructed mandible under physiological clenching conditions was used as an active design tool to derive the validation setup. Parameters were identified through analytical calibration of M-COU reaction forces and refined via a design-of-experiments procedure. The resulting setup was implemented as two coordinated components: a physical validation platform (R-VAL) enabling bench testing of the actual reconstruction plate, and a validation computational model (M-VAL) reproducing the same setting numerically. The resulting setup preserved the dominant mechanical features of the clinical scenario. In consistency analysis against M-COU, M-VAL achieved R² = 0.75, substantially exceeding literature-based benchmark configurations (R² = 0.27 and 0.40), and was the only configuration to preserve the clinically relevant stress distribution at the mandibular angle. Comparison between M-VAL and R-VAL showed close agreement in initial structural stiffness (0.6% difference) and local strain distribution (R² = 0.93; RMSE = 10.9%). The proposed methodology provides a transferable framework for constructing validation domains that are simultaneously experimentally feasible, mechanically interpretable, and grounded in a defined clinical COU.