Axial resolution of helical-orbit pinhole SPECT with synchronized and unsynchronized motion

Abstract

1520

Objectives: Pinhole SPECT imaging is widely used in small animal imaging. Helical orbits provide better sampling and axial resolution compared to circular ones; however the unsynchronized axial translation of the object and the camera rotation may cause blurring in the image resulting in poorer resolution. In an earlier study, the axial resolution for synchronized and unsynchronized motions was investigated with a MicroDefrise phantom. That study's results for synchronized motion disagreed with Monte Carlo expectations. Since then we have implemented angular-dependent axial mechanical shifts corrections that substantially improve axial resolution in this regime. In addition, a new phantom (UltraMicroDefrise), more representative of a mouse, has become avaliable. Thus there is motivation to repeat the previous set of experiments. We show that synchronized motion provides better results.

Methods: A small phantom, consisting of 9 hot disks with 1.6 mm thickness and 2.7 cm diameter, filled with ~20 mCi of ^99m-Tc was scanned with synchronized and unsynchronized helical orbits. The phantom was mounted on a computer controlled linear stage. Synchronization was achieved by translating the phantom while the camera was moving, i.e. while not acquiring data. In the unsynchronized method, the linear stage was moving continuously with constant speed during the entire data acquisition. Eight different image sets (20,30,40,50,60,90,120, and 180 views) were acquired with a triple-head Prism 3000XP SPECT scanner in step-and-shoot mode. The data acquisition time was such that each scan had the same number of expected emissions. An iterative OSEM reconstruction algorithm was used for image reconstruction. Resolutions from synchronized and unsynchronized data were calculated by fitting nine square pulses convolved with a Gaussian and compared.

Results: Synchronized motion yields better resolution (~0.65 mm) for any number of views. However, as the number of views decreases, unsynchronized motion resolution degrades considerably; for 20 views the resolution is 1.25 mm, and the image is affected by severe artifacts.

Conclusions: For synchronized motion axial resolution does not degrade as the number of views decreases abd is better than unsynchronized, especially for small number of views.

Research Support (if any): This work was supported by the National Institute for Biomedical Imaging and Bioengineering under Grants R01-EB-001910 and R33-EB-001543.