NA-MIC Project Weeks

Back to Projects List

Virtual plate placement for orbital surgery

Key Investigators

• Chi Zhang (Texas A&M University School of Dentistry, USA)
• Kyle Sunderland (Queen's University, Canada)
• Rafael Palomar (Oslo University Hospital / NTNU, Norway)
• Braedon Gunn (Texas A&M University School of Dentistry, USA)
• Andrew Read-Fuller (Texas A&M University School of Dentistry, USA)

Presenter location: In-person

Project Description

We are developing a Slicer module to simulating registering titanium plate to orbital fracture sites to aid surgical planning and investigating the fitness of different commercial preformed plates across a large sample of patients.

Objective

Make a module to allow the plate to properly sits above the bone at the orbital fracture site

Approach and Plan

See below.

Progress and Next Steps

I have complied existing methods in Slicer and VTK into one preliminary module https://github.com/chz31/surgical_plate_registration to do the semi-automatic registration:

1. An initial fiducial registration and a refined registration based on only allowing the plate to rotate while pivoting on the posterior stop, an important landmark for place the nail.

2. Use VTKCollisionDetector() and intersect marker to detect collision and mark the intersection

3. Use Probe Volume with Model to paint both the orbit and the plate to mark the overlapping areas.

4. Use the new Interaction Transform Handle to manually fine tune the position

7. Update new intersection and overlapping areas until they ar minimized.

Next steps:

1. Improving the use of the Interaction Transform Handle and intersection marker to more efficiently adjust plate position so that it sits just above the bone. Converting transform matrix into standard descriptor: yaw, roll, and pitch. (from https://doi.org/10.1371/journal.pone.0150162)

2. Design measurements to quantify plate fitness. Comparing the shape of the plate with the unfractured orbit.

3. Automated segmentation for segmenting fractured orbit.

4. Explore methods such as reinforcement learning and SOFA for actual simulating how the plate is placed.

PW Progress

Being able to use interaction handle to rotate the plate and real-time update colomap to highlight intersection and report intersection. This can provide visual aid for manual adjustment of plate position

Video demo:

https://youtu.be/4no8vEyKo5s

Next steps

1. Refine color-map and details for visual aid
2. Working with Rafael for using Slicer-SOFA:
   • Deforming the plate:
   • Simulating plate failure (e.g. the bone area for screwing the plate damaged and plate became loose; adapting/bending poor fitted plate introduces metal fatigue)
   • Including soft-tissue
3. Segmentation of fractured orbit?

Illustrations

No response

Background and References

The repo for the current module: surgical_plate_registration Other studies using commercial software iPlan from BrainLab, which should still be based on manual adjustment: Schreur et al. 2017 and Schreur et al. 2021