Jia Wang PhD, Presenter: Nothing to Disclose

Shuai Leng PhD, Abstract Co-Author: Nothing to Disclose

Xinhui Duan PhD, Abstract Co-Author: Nothing to Disclose

Lifeng Yu PhD, Abstract Co-Author: Nothing to Disclose

Cynthia H. McCollough PhD, Abstract Co-Author: Research grant, Siemens AG

In multi-energy photon counting CT, the choice of energy bins can be optimized with Cramer-Rao lower bound (CRLB) for best performance of basis material decomposition. The purpose of this work is to apply a local highly constrained backprojection reconstruction (HYPR-LR) algorithm to further reduce the noise in basis material images.

The optimization of basis material decomposition with photon counting measurements was performed in three steps: 1. Optimized energy bin choice was determined based on CRLB results from incident spectrum and assumed material area densities; 2. Basis material images were computed by minimizing the negative log-likelihood function of data of multiple bins; 3. Basis images were processed with HYPR-LR algorithm using the composite image generated with data of all bins. Two phantoms were simulated: 1. A 30-cm water phantom included 4 objects of different iodine (I) concentrations (Fig.1a): 10,20,30 and 40 mg/ml. 2. A 30-cm water phantom included three inserts of I (20mg/ml), calcium (Ca) (200mg/ml) and gadolinium (Gd) (25mg/ml), respectively. For phantom 1, CRLB of 2-bin, 3-bin, 4-bin and 5-bin was calculated and density maps of I were then computed with optimized bin settings. For phantom 2, CRLB of 4-bin and 5-bin was obtained and I, Ca and Gd density maps were computed. HYPR-LR was applied on density map of each contrast material of the two phantoms, respectively. Average density and its standard deviation were measured on resultant density images.

As predicated by CRLB calculation (Fig.1d), increasing number of bins reduced the noise in basis material images. For phantom 1, noise was decreased by 24.9%, 22.4%, 26.9%, 26.5% within iodine objects 10,20,30 and 40mg/ml by increasing from 2-bin to 5-bin (Fig.1a,b,e). Applying HYPR-LR to 5-bin image further reduced noise by 50.7%,53.6%,54.7%,51.6% (Fig.1b,c,f). For the multi-contrast phantom 2, image noise was slightly decreased by 1.7%,1.3%,6.4% for I,Ca,Gd from 4-bin to 5-bin, while noise was further reduced by 56.9%,53.6%,55.1% with HYPR-LR.

Material decomposition in multi-energy photon counting x-ray imaging was optimized with CRLB and image noise was further decreased with HYPR-LR by taking advantage of the low-noise composite image.

Material specific information provided by multi-energy CT has promising clinical applications such as multi-contrast imaging.

Wang, J, Leng, S, Duan, X, Yu, L, McCollough, C, Noise Reduction on Basis Material Images with HYPR-LR in Photon Counting X-ray Imaging.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11009505.html