Zhihua Qi PhD, Presenter: Nothing to Disclose

Brian Edward Nett PhD, Abstract Co-Author: Employee, General Electric Company

There are a number of emerging clinical applications for computed tomography from systems which collect axial scan data with a large cone angle. The clinical use of these systems, faces challenges from cone beam artifacts caused by: data insufficiency, data mishandling and data truncation associated with an axial cone beam CT scan. This study aims to evaluate the performance of a newly developed analytical algorithm, referred to here as the improved algorithm, with a focus on cone beam artifact evaluation, for axial cone beam CT in cardiac imaging.

Numerical simulations of an anthropomorphic phantom (i.e XCAT) were performed with a total coverage of 14.6 deg. Images were reconstructed by both the traditional FDK algorithm and the improved algorithm for both full-scan and cardiac modes. Then three types of tissue: cardiac muscle, soft tissue, and lung, were segmented and separately analyzed for the effect of cone beam artifacts. The severity of cone beam artifacts for a certain type of tissue in different slices is evaluated using the metric of mean CT# deviation from the ground truth.

For both full-scan and cardiac acquisitions, the mean CT# deviation in the images reconstructed with the improved algorithm remains significantly lower than the FDK type algorithm for all slices and all tissues. For a full-scan acquisition the images reconstructed with the improved algorithm, compared with the FDK reconstruction, show a reduction of the mean CT# deviation of 52%±16%, 39%±21%, and 53%±19%, for cardiac muscles, soft tissues, and lungs, respectively, over the studied cone angle coverage. For a cardiac acquisition, the reductions of the mean CT# deviation from FDK to WC reconstructed images for the three tissue types are 79%±24%, 77%±20%, and 79%±17%, respectively.

Compared to the traditional FDK-type algorithm, the newly developed Wide-Cone algorithm substantially reduces the cone-beam artifacts without degrading the other image quality aspects.

A new analytic reconstruction algorithm for large cone-angles can enable quantitative imaging of whole organs such as the heart.

Qi, Z, Nett, B, Performance Evaluation of a New Analytic Reconstruction Algorithm for Axial CT with Large Cone-angle Coverage.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13016770.html