Positional dependence of activity determination in single photon emission computed tomography.

Emlyn Price,Andrew P Robinson,Jill Tipping,Nick Calvert,Emma Page,David M Cullen,David Hamilton,Ben Pietras,Sophia Pells

doi:10.1097/mnm.0000000000001034

Abstract

Accurate image quantification requires accurate calibration of the detector and is vital if dosimetry is to be performed in molecular radiotherapy. A dependence on the position of calibration has been observed in single photon emission computed tomography images when attenuation correction (AC) and scatter correction are applied. This work investigates the origin of this dependence in single photon emission computed tomography scans of phantom inserts filled with 177Lu solution. A 113 ml sphere and inserts representing a mathematical model of a spleen and an anatomical model of a patient spleen were imaged at the centre and edge of elliptical phantoms. For these inserts, the difference in calibration factor between the positions was around 10% for images reconstructed with AC and triple energy window scatter correction. A combination of experimental imaging and Monte Carlo simulation was used to isolate possible causes due to imaging or reconstruction in turn. Inconsistent application of AC between different reconstruction systems was identified as the origin of the positional dependence.

Highlights

The 3D quantification of the activity distribution in a patient is essential if meaningful dosimetry is to be carried out in nuclear medicine therapies
The counts extracted from reconstructed single photon emission computed tomography (SPECT) images must be related to this activity distribution
It can be seen that the calibration factors for each insert are consistent when the insert is radially displaced in vacuum for both the EM1 and EM2 windows

Summary

Introduction

The 3D quantification of the activity distribution in a patient is essential if meaningful dosimetry is to be carried out in nuclear medicine therapies. The counts extracted from reconstructed single photon emission computed tomography (SPECT) images must be related to this activity distribution. The relationship between counts and activity depends on the scanner used, the corrections applied during image reconstruction and the volume of interest (VOI) [1]. The spheres were imaged on GE Infinia Hawkeye, GE Discovery 670 and Siemens Symbia T cameras. The images were reconstructed with native and vendor-neutral software. The cause of the positional dependence of the calibration factor was suggested to be the depth-dependent spatial resolution (DDSR) of SPECT scanners [2].

Methods

Results

Conclusion