2012 Winter Project Week:PVRegistration

• 

  Projects List

Key Investigators

Josh Cates, Yi Gao, Liang-Jia Zhu, Danny Perry, Alan Morris, Greg Gardner, Rob MacLeod, Sylvain Bouix, Allen Tannenbaum

Automatically selected correspondences

User-chosen fiducial-based approach

• We need a more accurate registration pipeline for the pulmonary veins.

Objective

• To explore ideas for establishing correspondence (via registration or other means) between points on the surface of the pulmonary veins (PV) and PV antrum near the left atrium in pre-RF-ablation DEMRI images with post-RF-ablation DEMRI images to within about 2 mm precision. This may be an impossible task (or impossible to validate the results, anyway), but there are many important application areas for AF research and clinical treatment if we can do it. For the Project Week, our main goal is to explore possibilities with available NAMIC expertise and Slicer tools during the meeting. If a likely algorithm can be identified, we will plan a Slicer module to implement it.
• Related to the idea of registration and correspondence in the PV, we have some specific ideas for how to effectively visualize and interact with off-axis, 2D slices through registered (in the sense of the aforementioned surface correspondence to within 2mm) pre- and post- RF ablation DEMRI images. Our specific needs are to present a physician with side-by-side views of corresponding anatomy on which to make manual measurements and qualitative analyses. The visualization tools we have in mind do not seem to easily fit within the constraints of a current Slicer module, so we may need to talk about additional Slicer core functionality.
• See also this related [2012_Winter_Project_Week:LAWallRegistration|registration project]. These are listed as separate projects because they may involve different registration methods and visualization tools.

Approach, Plan

• Manual landmarks
• Establish a rough registration
• Visualization tools?

Progress

• We have designed a tentative workflow. We need to explore how to implement this in Slicer.