2013 Project Week:CARMA PractialLASeg
Key Investigators
Salma Bengali, Greg Gardner, Alan Morris, Josh Cates, Rob MacLeod
Automatic segmenation of the left atrium (LA) from LGE-MRI images continues to be challenging to automate. Recent work by Ross and Gopal looks promising, but will require further validation and some significant work to implement in Slicer in a useful format for CARMA technicians. In the meantime, we propose to explore some new approaches to a semi-automated segmentation, with the goal of reducing the necessary time for the manual process (the process is described in File:CarmaWorkflowV2.pdf).
CARMA has recently begun to acquire fully gated MRA images, along with the LGE-MRI protocol. Gated MRA can produce much sharper blood pool boundaries (LA endocardium) than standard, 1-pass MRA and may prove valuable for localizing endocardial contours.
Objective
The objective of this project is to collaboratively brainstorm and (ideally) implement a semi-automated LA segmentation workflow (wizard?) within slicer. We plan to take advantage of available Slicer experts and provide a wish list to Slicer engineers of any missing pieces -- or nice to have features -- necessary for our proposed workflow.
Approach, Plan
- Explore the use of Slicer to perform semi-automated segmentation of the LA Endocardium using gated MRA aligned with LGE-MRI
- Identify missing segmentation tools that may be useful in a proposed segmentation workflow (e.g. 3D editing?)
- Plan (or implement if time permits) a workflow (wizard) within CARMA based on what we learn from objectives 1-2
Progress
We decided that the first step will be to create a Slicer tutorial and then map this tutorial into a workflow module. Work is underway to port some existing CARMA tutorial materials to a Slicer-based tutorial.
