Improvement of Accidental Coincidence Estimation for Scanditronix PC2048/4096 PET Scanners
Accidental coincidence (randoms) is one of the factors that degrade the accuracy of positron emission tomography (PET). In contrast to widely-held beliefs, we have previously shown that randoms can significantly affect quantitative PET even after current hardware/software randoms correction is applied. In order to have better estimation of randoms for our Scanditronix scanner, we investigated the limitation of the vendor-provided randoms estimation and proposed a simple modification. Methods: Multiple short-frame data of a homogeneous circular cylinder were collected over 12 hours. Summed data were generated by summing up selected frames. Randoms of the summed data estimated by the vendor-provided method and our modified method were compared. Results: The vendor-provided randoms estimation did not work as well as the modified method when frame duration is much larger than the half-life of the radioactive agent used. Conclusion: Our minimal modification of the vendor-provided method significantly improves randoms estimation.