View source for Projects:RiskMassEstimation

Back to [[Algorithm:BU|Boston University Algorithms]]
__NOTOC__

= Description =
The proposed method has three major components: First, extract the coronary arteries from CTA image and identify stenoses with severe occlusions. Second, segment the left ventricular surfaces from CTA image. Third, estimate the affected mass based on the segmentations of the arteries and left ventricle. These steps are illustrated in the figure below.

*[[File:RiskMassSteps.png|600px]]
Volume at risk estimation process. Left: Extract the area at risk (purple) on the epicardial surface given the stenosis location (greed dot) on the coronary arteries (red lines). Middle: Trace out the risk contour (yellow) on the endocardial surface. Right: Construct the volume at risk (purple).

= Results =
In experiments, 11 human cardiac CTA images were used to validate the proposed computational framework. In addition, a comparison was made between the percentage of the myocardial mass at risk estimated from CTA images with their corresponding values determined from SPECT MPI. 

*[[File:RiskMassPlots.png|600px]]
Left: Correlation analysis. Right: Bland-Altman plot for %LV of CTA vs. SPECT

= Key Investigators =
= Key Investigators =
*Georgia Tech: Liangjia Zhu and Anthony Yezzi
*BWH: Yi Gao
*Boston University: Allen Tannenbaum

= Publications =