Somehow my message was treated as spam and deleted!! So I repost it here. Dear Susana and GATE users, Thank you very much for your solutions. I have tested Susana's ideas by following 5 quick simulations: 1. air + back-to-back(object+source): 3409/2/20/20  (true/scatter/random/delay) 2. vacuum + back-to-back: 3483/0/20/23 3. water + back-to-back: 2208/383/17/18 4. no object + back-to-back: 29472/0/11/16 5. no object + F-18: 5818/0/19/28 All simulations are applied on ECAT 962, the object (if have one) is same shape to the source (cylinder with 4 cm radius and 3.88 cm long). Activity is 0.01 mCi and scan time is 1 second. (see the attached .mac) Obviously, simulation 4 is definitely wrong!! By using the absolute sensitivity of this system (about 9 cps/kBq according to Daube-Witherspoon, J Nucl Med 2002, 43: 1398-1409 ), the recorded true counts should be around 3300. Does this mean that we cannot use this combination for simulations? Simulation 5 is also questionable. When there is no medium around the source, how can Gate simulate annihilations without electrons? And the total counts is also too high. Simulation 1 & 2 are reasonable from my point of view. Did I do something wrong, or miss something here ? Any comments are very welcome. Thank you very much. Sincerely, Hongwei _____________________________________________________________________________ Scanned by IBM Email Security Management Services powered by MessageLabs. For more information please visit http://www.ers.ibm.com _____________________________________________________________________________