PT  - JOURNAL ARTICLE
AU  - Yu-Lung Hsieh
AU  - Sharon Wang
AU  - Jinghan Ye
AU  - Jun Zhang
AU  - Katherine Binzel
AU  - Chadwick Wright
AU  - Michael Knopp
AU  - Zhiqiang Hu
TI  - Optimizing Y-90 imaging for a digital PET/CT system
DP  - 2015 May 01
TA  - Journal of Nuclear Medicine
PG  - 1282--1282
VI  - 56
IP  - supplement 3
4099  - http://jnm.snmjournals.org/content/56/supplement_3/1282.short
4100  - http://jnm.snmjournals.org/content/56/supplement_3/1282.full
SO  - J Nucl Med2015 May 01; 56
AB  - 1282 Objectives Y-90 radioembolization becomes a common treatment for inoperable hepatocellular carcinoma (HCC). However, Y-90 has very low positron branching ratio of 3.187 x 10-5 along with high cascade gamma rays at energy of 628 keV. The cascade gamma rays can reduce the prompts sensitivity (prompts/(prompts+delays)) to below 30%. Low branching ratio and cascade gamma rays make it challenging to generate quantitative PET images from Y-90. This study aims at a method for optimizing the Y-90 images.Methods Instead of correcting the cascade gamma rays with modeling [1], a new method is proposed to reduce the cascade gamma rays and improve the prompts sensitivity by selecting an optimized upper energy window at 552 keV, which represents the valley between the 511 keV gamma rays and the cascade gamma rays in the energy spectrum. In addition, due to the low count statistics, a system modeling of random and scatter events is applied to the reconstruction algorithm to improve the intermediate image quality for estimating the scatter distribution. A time-of-flight based PET reconstruction algorithm has been implemented in the Philips digital PET/CT system (Vereos) using blobs as basis function. The selection of blob parameters, number of iterations, and number of subsets were optimized to improve the Y-90 image quality.Results With the optimized energy window the prompts sensitivity for Y-90 was raised from 26.4% to 52.5% and improved the NAC (with no attenuation or scatter correction) images with much cleaner background than the images from the original energy window. The optimization in reconstruction process and parameter tuning yielded better CTAC (with attenuation and scatter correction) images showing more realistic Y-90 distribution.Conclusions A new method is proposed to optimize the data and reconstruction processes for improving the Y-90 image quality on the Philips digital PET/CT system (Vereos). It is demonstrated that this new method shows promise to advance clinically feasible image quality and quantification accuracy.