Implementation and evaluation of a CT-based attenuation correction method for a dedicated small-animal PET system

Abstract

1718

Objectives: The purpose of this work is to develop a practical CT-based attenuation correction (AC) method for the GE eXplore VISTA small-animal PET system, and quantitatively evaluate its accuracy and reliability when combined with the VISTA scatter correction (SC). Methods: Two cylindrical phantoms with inside diameters of 2.8 cm and 4.5cm containing 4mm diameter small sphere source inserts filled with [18]F-FDG were scanned both on the VISTA PET and the CT subunit of Gamma Medica XSPECT/CT system. Six 0.4~0.6mm diameter beads soaked in the [18]F-FDG for 30min were taped separately on each phantom’s surface to guide the PET/CT image co-registration. An attenuation calibration phantom (ACP) with inserts mimicking cortical and spongiosa bone, adipose and lung tissue was used to derive the relationship between CT number and linear attenuation coefficient (mu-value) at 511 keV (Radiology Support Devices Inc.). PET images were reconstructed with and without SC+AC to assess quantitative accuracy of the corrections. For comparison with the CT based results, we generated an AC based on PET transmission scans obtained with an annular [68]Ge source. PET images of rats and mice were also analyzed with and without SC+AC to characterize the extent of these corrections on different organs. Results: For volumes of interest (VOIs) covering the small sources in the 2.8 and 4.5 cm diameter phantoms, the activity was underestimated by 18.1% and 34.1% when no correction is applied. Meanwhile, images reconstructed with SC+AC overestimated point source activity by 11.3% and 10.9%, respectively. For VOIs placed over the brains and hearts of a 33g mouse, SC+AC increased image counts by 26.8% and 27.1%, respectively. The case for a 270g rat showed 45.6% and 45.3% increase. CT-based and transmission-based AC results for the ACP phantom agreed within 3.2%. Conclusions: This straightforward CT-based AC method achieves equivalent performance over the quite time-consuming transmission-based AC on the eXplore VISTA PET system. The overestimation of small point source activities by the combined SC+AC may be caused by an insufficient VISTA scatter correction and is subject to further investigation.

Research Support (if any): NIH grant CA92781