Automated Global Positioning Layout (GPL) for accuracy assessment in CAD-CAM mandibular reconstruction - Method validation.

The lack of a standardized methodology complicates accuracy assessment in computer-assisted mandibular reconstruction. Existing landmark-based methods are susceptible to operator variability, while surface-based comparisons can mask local deviations. This study validates a novel, automated protocol, the Global Positioning Layout (GPL), to quantify the 3D discrepancy between the virtual surgical plan and the postoperative outcome, by comparing its performance and reliability against Methods A and B. A retrospective cohort of 17 patients was analysed, with three operators performing all measurements on two occasions. The GPL method demonstrated complete reproducibility, with no inter- or intra-operator variability, providing a detailed, spatially-oriented assessment of deviations. In contrast, the landmark-based method showed poor reproducibility and systematic bias and was often inapplicable due to the absence of landmarks after resection. The surface-based method, while objective for its mean error metric, was operator-dependent for initial alignment and its non-directional output masked significant localized deviations. This study validates GPL as a robust and fully reproducible tool that overcomes the critical limitations of established techniques. The GPL method provides a strong foundation for a standardized protocol, essential for the reliable comparison of surgical outcomes, refinement of surgical techniques, and improvement of long-term patient outcomes.