Aim: Positron emission tomography (PET) imaging of non-standard positron emitters is influenced by gamma-coincidences, i.e. false coincidences produced by the coincident detection of an annihilation photon and a gamma-ray simultaneously emitted with the positron. The extent to which the PET study is disturbed by this effect is dependent on the kind of the positron emitter used, the kind and position of the object, the acquisition mode, i.e. the optional use of septa, and the reconstruction program. In order to demonstrate and study imaging problems with non-standard positron emitters, a phantom was scanned containing non-radioactive rods with different absorption materials and filled with either (124)I or (86)Y in the bidimensional (2D) as well as tridimensional (3D) acquisition mode.
Methods: For reconstruction, the PET manufacturer's standard software without any modification was used. To reduce errors caused by the gamma-coincidences, a simple linear background subtraction, estimated from the counts at the scanner's external radius, was applied.
Results: Without background subtraction, apparent positive and negative ''radioactivity concentrations'' were found in regions of interest positioned over the non-radioactive rods with values higher for (86)Y compared to (124)I and also higher for 3D compared to 2D. A complete subtraction of the background led to erroneous
Results: The errors in the phantom's non-radioactive rods and the difference between measured and true radioactivity became minimum, when about 75% of the background was subtracted. This refers to both the 2D and 3D mode.
Conclusion: Quantitation problems with the non-standard positron emitters (124)I and (86)Y could be minimized in the phantom study examined here by using a simple background subtraction together with the manufacturer's standard correction and reconstruction procedures.