137Cs transmission imaging and segmented attenuation corrections in a small animal PET scanner.

Abstract

Attenuation correction (AC) is required for accurate quantitative evaluation of small animal PET data. Our objective was to compare three AC methods in the small animal Clairvivo-PET scanner. The three AC methods involve applying attenuation coefficient maps generated by simulating a cylindrical map (SAC), segmenting the emission data (ESAC), and segmenting the transmission data (TSAC), imaged using a 137Cs single-photon source. Investigation was carried out using a 65mm uniform cylinder and an NEMA NU4 2008 mouse phantom, filled with water or tungsten liquid, to mimic bone. Evaluation was carried out using the difference of the segmented map volume from the known cylindrical phantom volume, the recovery of the radioactivity concentration, and the line profiles. The optimal transmission scan time for achieving accurate AC using TSAC was determined using 5, 10, 15, 20, and 25min transmission scan time. The effects of scatter correction and reconstruction algorithms on ESAC were investigated. SAC showed the best performance but was unable to correct for different tissues and the scanner bed, and faced difficulty with correct positioning of the attenuation coefficient map. ESAC was affected by scatter correction and reconstruction algorithm, and may result in poor boundary delineation, and hence was unreliable. TSAC showed reasonable performance but required further optimization of the default segmentation setting. A minimum transmission scan time of 20min is recommended for Clairvivo-PET using 137Cs source to ensure that sufficient transmission counts are obtained to generate accurate attenuation coefficient map.