Mimics Innovation Finalist 2021: Michael Boelstoft Holte

To develop an automated method for virtual analysis of maxillofacial surgery in order to assess and optimize personalized virtual surgical planning and the surgical outcome.

Today, surgical planning in clinical practice is unidirectional, meaning that the outcome of transferring the virtual treatment plan to the patient is not assessed nor used as feedback to optimize surgery. M. B. Holte believes that developing a way to efficiently and reliably evaluate the surgical outcome can improve personalized treatment planning and thereby the surgery. Furthermore, more clinical studies are needed to clearly determine the accuracy of virtual surgical planning (VSP) and its effectiveness over traditional surgical planning.

Most of the methods for evaluating the postoperative outcome compared to the preoperative planning are based on cephalometric analysis, where anatomical landmarks are manually placed and re-identified to measure bony movements. This analysis, however, is time-consuming and accumulates manual landmark identification errors, which have been shown to exceed the clinical relevant error margin of 0.5 mm and possibly exceed the clinical acceptable accuracy of 2 mm.

M.B. Holte aims to propose and validate an automated and reliable approach for virtual surgical analysis of maxillofacial surgery, which meets the following criteria suggested by Gaber et al.:

1. Voxel-based registration, on the cranial base, of planned and postoperative images should be performed to decrease the possibility of human error.
2. Automated or semiautomated evaluation of the outcome should be performed. The outcome should be indicative of changes in 3D; whether it is translational or rotational based on the different axes (x, y, z).
3. Interobserver and intraobserver reliability should be used to validate the results.

M. B. Holte introduced the concept virtual surgical analysis (VSA) — the next frontier in computer-assisted surgical assessment. Holte’s VSA demonstrated excellent reliability. It enables automated 3D assessment of the postoperative outcome of bimaxillary surgery and the accuracy of VSP.

The automated 3D measurements of all bone segments are performed using voxel-based registration, and the user is provided with the option to use no cephalometric landmarks (landmark-free) or a preferred set of landmarks, without the need for landmark re-identification. The virtual bone segments from the virtual planning are used directly in the virtual analysis without the need for any image segmentation of the follow-up scans.

The automated VSA makes a complete skeletal positional assessment of bimaxillary surgery in daily clinical practice feasible. Additionally, it enables evaluation and comparison of surgical techniques, such as the benefit of personalized implants, to optimize maxillofacial surgery.

After submission to the Mimics Innovation Awards 2021, the VSA has been expanded for automated analysis at multiple post-surgical time-points, enabling 3D assessment of skeletal stability in addition to surgical accuracy for better understanding of possible treatment relapse. The VSA expansion has been validated on scans acquired two years after surgery. Long-term postoperative assessment is notoriously more difficult than short-term assessment due to bone remodeling and resorption as response to the surgery. Again, the VSA demonstrated excellent reliability.