2011 Summer Project Week Spine Segmentation And Osteoporosis Screening in CT
With the increased volume of CT studies and the number of slices in each, diagnostic radiologists need to focus their attention on the clinical situation at stake, at the expense of potential under diagnosis of incidental findings. One commonly underdiagnosed domain is skeletal conditions in the region of the spine. Our goal is to create a fully automated screening tool to detect suspect incidental findings in the skeletal system in CT studies, while requiring the attention of a radiologist only in cases of positive findings.
Our first mission is fully automated segmentation of the spine and detection of osteoporosis symptoms.
Key Investigators
- Radnostics: Anthony Blumfield
- Albert Einstein College of Medicine: Einat Blumfield, MD
Objective
Primary goal: Slicer integration
Stretch goals (subject to resources):
P1 Improvements in segmentation. E.g. removal of artifacts related to contrast in the circulatory and digestive systems, calcifications etc.
P1 Add platform support for Mac & Linux
P2 Develop segmentation algorithms for posterior processes
P2 Noise reduction in extremely noisy CT studies
P2 Add Support for upper thoracic vertebra T1-6
Approach, Plan
Write code in consultation with Slicer experts
Collaborate with NA-MIC researchers to increase project participation
Collaborate with NA-MIC leadership on IP licensing
Progress
References
http://wiki.na-mic.org/Wiki/index.php/2011-Challenge:Segment-a-spine
http://radnostics.com/default.aspx
To join or collaborate with this project, please contact Anthony dot Blumfield At Radnostics dot Com.
Delivery Mechanism
This work will be delivered to the NAMIC Kit as a
- NITRIC distribution
- Slicer Module
- Built-in:
- Extension -- commandline:
- Extension -- loadable: