Correction methods for random coincidences in 3D wholebody PET imaging

Abstract

With the advantages of the increased sensitivity of 3D PET imaging for wholebody imaging come the challenges of more complicated quantitative corrections, and in particular an increase in the random coincidence field of view (FOV) relative to the true coincidence FOV. The most common method of correcting for random coincidences is the on-line subtraction of a delayed coincidence channel, which does not add bias but increases noise. An alternative approach is the post-acquisition subtraction of a low noise random coincidence estimate, which can be from a smoothed delayed coincidence channel, a calibration scan, or directly estimated. Each method makes different tradeoffs between noise amplification, bias, and data processing requirements. These tradeoffs are dependent on activity injected, the local imaging environment (e.g. near the bladder), and the reconstruction algorithm. Using 3D wholebody simulations and phantom studies, we show that the gains in sinogram NEC by using a noiseless random coincidence estimation method are translated to improvements in image SNR. The image SNR, however, depends on the image reconstruction method and the local imaging environment. For 3D wholebody imaging, a low noise estimate of random coincidences based on the single photon rates improves sinogram and image SNRs by approximately 15% compared to on-line subtraction of delayed coincidences, and performs only slightly worse than using a 3D extension of the Casey-Hoffman smoothing of a separately acquired delayed coincidence sinogram.