Grootoonk S, Spinks T J, Sashin D, Spyrou N M, Jones T
Cyclotron Unit, MRC Clinical Services Centre, Hammersmith Hospital, London, UK.
Phys Med Biol. 1996 Dec;41(12):2757-74. doi: 10.1088/0031-9155/41/12/013.
A method for scatter correction using dual energy window acquisition has been developed and implemented on data collected with a brain-PET tomograph operated in the septa retracted, 3D mode. Coincidence events are assigned to (i) an upper energy window where both photons deposit energy between 380 keV and 850 keV or (ii) a lower energy window where one or both photons deposit within 200 keV and 380 keV. Scaling parameters are derived from measurements of the ratios of counts from line sources due to scattered and unscattered events in the two energy windows in head-sized phantoms. A scaled subtraction of the two energy windows produces a distribution of scatter which is smoothed prior to subtraction from the upper energy window. In phantoms, the correction was found to restore the uniformity, contrast and linearity of activity concentration. Relative activity concentrations were restored to within 7% of their true values in a multicompartment phantom. The method was found to provide accurate correction for scattered events arising from activity outside the direct detector field of view. In a three-compartment phantom containing water, 18F and 11C scanned in dynamic, multiframe mode, the half-lives of the two isotopes were restored to within 2% of their true value.
