Deadtime and pile-up correction method based on the singles to coincidences ratio for PET

Abstract

The count rate in a PET scanner as a function of the activity in the Field of View (FOV) has a non-linear contribution coming from deadtime, pile-up and random coincidences. These effects must be estimated accurately and corrected in order to perform quantitative PET studies. For a given scanner and acquisition system, the relative importance of deadtime and pile-up effects still depends on the size and materials of the objects being imaged. These facts difficult to devise a universal correction method that yields accurate results for any kind of acquisition. In this work we show that, in a PET scanner, there is a linear relationship between the effective deadtime for coincidences, τ, (which takes into account deadtime and pile-up losses and gains within the energy window) and the Singles to Coincidences Ratio (SCRm) measured by the scanner. This relation has been recovered both in simulations and real data. This allows us to devise a simple method which, requiring only two calibration acquisitions for each energy window, one with high SCRm and one with low SCRm, is able to estimate accurately deadtime and pile up corrections for any other acquisition performed in the same scanner. Simulations show that corrected count rates are accurate within 5%, even when high activities are present in the FOV.