NA-MIC Project Weeks

Back to Projects List

Evaluate the fit of preformed plates in orbital surgery

Key Investigators

Chi Zhang (Texas A&M College of Dentistry)
Braedon Gunn (Texas A&M College of Dentistry)
Andrew Read-Fuller (Texas A&M College of Dentistry)

Project Description

Orbital fracture repair commonly used preformed/contoured plates. ProprietaryVirtual Surgical Planning (VSP) soft focused on customized plate manufacture and placement guidance but does not allow comparing the fit of plates across vendors. The goal of this project is to allow surgeons to interactively try different ways of placement or different plates and compare the fit to find the more suitable implant. I created a customized extension orbiSurgerySim.

Current procedures:

Alignment using the Interaction Transform Handle: rotate around the “posterior stop” landmark until the plate sits just above the. Real-time heatmap using Probe Model with Volume module and VTK collision detector can mark intersections to aid plate position adjustment.

Reconstructing the fractured orbit using the mirror of the contralateral side: use DentalSegmentator to segment the bone. I created a module to streamline reconstruction by using the mirrored contralateral side to rigidly register to fracture side followed by a warping.

Measuring fit by projecting fiducial points on plate edges to the reconstructed orbit. Fit is based on overall distances. Overall distance map can also be used. These distances can also be used to highlight largest deviation.

Objective

The current workflow is tedious. My goal is to simplify the process.

Design UI and workflow to streamline the processes
Alternative methods for plate registrastion to ensure the plate sits at the surface of the bone (e.g., methods from SlicerHeart).
Improve and design additional metrics for measuring the fit of different plates and their ways of placement.
Explore additional or improved methods for orbit reconstruction (e.g., using SlicerAntsPy for image deformable registration).

Approach and Plan

Same as objectives
Simplify segmentation workflows for thin orbital bones. In some cases, the orbital floor and medial wall are too thin, and only a boundary between the sinus and orbit is visible. I segmented maxillary sinus, expand it, and create a shell to fix the orbit. Perhaps models used in DentalSegmentator or TotalSegmentator can be tuned to depict the boundary between sinus & orbit.