Planning of mandibular reconstructions based on statistical shape models

The reconstruction of large continuity defects of the mandible is a challenging task, especially when the shape of the missing part is not known prior to operation. Today, the surgical planning is based mainly on visual judgment and the individual skills and experience of the surgeons. The objective of the current study was to develop a computer-based method that is capable of proposing a reconstruction shape from a known residual mandible part.

The volumetric data derived from 60 CT scans of mandibles were used as the basis for the novel numerical procedure. To find a standardized representation of the mandible shapes, a mesh was elaborated that follows the course of anatomical structures with a specially developed topology of quadrilaterals. These standard meshes were transformed with defined mesh modifications toward each individual mandible surface to allow for further statistical evaluations. The data were used to capture the inter-individual shape variations that were considered as random field variations and mathematically evaluated with principal component analysis. With this information of the mandibular shape variations, an algorithm was developed that proposes shapes for reconstruction planning based on given residual mandible geometry parts.

The accuracy of the novel method was evaluated on six different virtually defined continuity defects that were each created on three mandibles that were not part of the initial database. Virtual reconstructions showed sufficient accuracy of the algorithm for the planning of surgical reconstructions, with average deviations toward the actual geometry of \(1.82 \pm 0.11\) mm for small missing parts and 5 mm for large hemi-lateral defects.

The presented algorithm may be a valuable tool for the planning of mandibular reconstructions. The proposed shapes can be used as templates for computer-aided manufacturing, e.g., with 3D printing devices that use biocompatible materials.