Validation of NEMA NU4-2008 Scatter Fraction estimation with 18F and 68Ga for the ARGUS small-animal PET scanner

The scatter fraction (SF) in PET data represents the fraction of coincidence events in which at least one of the two emitted photons have been scattered before being detected. It is usually estimated as the ratio of scattered events to total number of coincidences, when the number of random counts is negligible (less than 1% of true rates). SF provides a measurement of the relative sensitivity of the scanner to scattered radiation. It depends on object size, density and location inside the field of view, as well as on detector size, type of detector crystal and energy window. The performance evaluation guideline for small-animal PET NEMA (National Electrical Manufacturers Association) NU4-2008 proposes the estimation of the SF for three test phantoms made in proportion to the most widely used small animals in the laboratory: mouse, rat and monkey. The method estimates the different coincidence types in sinogram profiles from an off-centered line source inserted in these phantoms. We benchmark the procedure proposed by NEMA to estimate SF with F-18 and also with Ga-68, a radionuclide with lager positron range. Real data acquired with the ARGUS small-animal PET scanner (SEDECAL, Madrid, Spain) as well as simulations of the same scanner with peneloPET are used. The results show that, though SF should be practically the same with both F-18 and Ga-68 isotopes (and indeed our simulations indicate this) NEMA SF estimations with Ga-68 acquisitions are higher. This is due to the fact that Ga-68 positron range affects on the width of the line source profiles. Suggestions to modify the protocol to obtain similar SF estimations when using isotopes with larger positron range than F-18 are made.