I presented a poster at the Lifespan Annual Research Celebration on the preliminary results from MUSTACHE. You can find the abstract published online here.
Research update: Not much happened in the last few weeks. I've been fixing bugs in the original MUSTACHE code and writing up the research proposal for my honour's thesis. I collected accuracy, sensitivity, and specificity of each feature to see if I can identify some that are not contributing. There do seem to be a few features that didn't improve accuracy at all, but when they're taken out, they still made the accuracy worse?? Another thing I'm trying is to standardize the sizes of the ROI's. That doesn't seem to have much of an effect right now.
We got some dual energy CT data for ablation to look at if having both energy levels can reduce the noise. That seems plausible, but I haven't been able to test my procedure, with only one set of data. We're going to try collecting pseudo dual energy CT data next time Garron ablates a cow liver by having two sets of data collected right after each other.
Siemens people finally came today, and they seemed pretty excited about the thyroid elastography project. It'd be cool to have perfectly registered elastography and b-mode ultrasound data.