Abstract
With the increase in 3D acquisition of positron emission tomography (PET) data, triple event is becoming a more and more relevant issue. Triple event may occur when three photons are detected within a coincidence time window. For traditional PET system, the triple coincidence is discarded. However, the triple coincidence usually contains a valid true coincidence and an unrelated gamma. Thus, the purpose of the study is to propose a novel method to enhance the sensitivity of PET imaging via recycling the triple coincidence. A combined likelihoods model involving time, geometry, energy, and activity information is proposed to recover the true coincidence form the triple coincidence. Monte Carlo simulations of 3D PET on an NEMA phantom and a uniform phantom were conducted to validate the proposed approach. Results showed that the amounts of triple event increase with the increase of the activities. The ratios between the true triple and true double coincidence can achieved about 3% to 48% at the activity levels from 1 mCi to 20 mCi. By using the proposed method, more than 92% true coincidence can be recovered from triple coincidence on both phantoms. Furthermore, the NECR gains can be achieved by 3% ~ 57% with the activity levels from 1 to 15mCi. We conclude that ours proposed methods can improve the counting statistics for positron emitters and are readily applicable to clinical studies as no hardware modification is needed.