Locking Plates With Computationally Enhanced Screw Trajectories Provide Superior Biomechanical Fixation Stability of Complex Proximal Humerus Fractures

Joint-preserving surgical treatment of complex unstable proximal humerus fractures remains challenging, with high failure rates even following state-of-the-art locked plating. Enhancement of implants could help improve outcomes. By overcoming limitations of conventional biomechanical testing, finite element (FE) analysis enables design optimization but requires stringent validation. This study aimed to computationally enhance the design of an existing locking plate to provide superior fixation stability and evaluate the benefit experimentally in a matched-pair fashion. Further aims were the evaluation of instrumentation accuracy and its potential influence on the specimen-specific predictive ability of FE. Screw trajectories of an existing commercial plate were adjusted to reduce the predicted cyclic cut-out failure risk and define the enhanced (EH) implant design based on results of a previous parametric FE study using 19 left proximal humerus models (Set A). Superiority of EH versus the original (OG) design was tested using nine pairs of human proximal humeri (N = 18, Set B). Specimen-specific CT-based virtual preoperative planning defined osteotomies replicating a complex 3-part fracture and fixation with a locking plate using six screws. Bone specimens were prepared, osteotomized and instrumented according to the preoperative plan via a standardized procedure utilizing 3D-printed guides. Cut-out failure of OG and EH implant designs was compared in paired groups with both FE analysis and cyclic biomechanical testing. The computationally enhanced implant configuration achieved significantly more cycles to cut-out failure compared to the standard OG design (p < 0.01), confirming the significantly lower peri-implant bone strain predicted by FE for the EH versus OG groups (p < 0.001). The magnitude of instrumentation inaccuracies was small but had a significant effect on the predicted failure risk (p < 0.01). The sample-specific FE predictions strongly correlated with the experimental results (R(2) = 0.70) when incorporating instrumentation inaccuracies. These findings demonstrate the power and validity of FE simulations in improving implant designs towards superior fixation stability of proximal humerus fractures. Computational optimization could be performed involving further implant features and help decrease failure rates. The results underline the importance of accurate surgical execution of implant fixations and the need for high consistency in validation studies.