Simultaneous activity and crystal efficiencies reconstruction: TOF patient-based detector Quality Control

PET scanner calibration is a routine procedure that is performed on a daily basis. In Siemens scanners, for example, data are acquired for about 20-30 minutes using a 20 cm diameter uniform cylinder. Assuming a known object, the estimation of a crystal efficiencies normalization component is carried out, since the rest of the normalization components are fixed for a given scanner type. Recently it was demonstrated that each line of response (LOR), attenuation can be estimated from time of flight (TOF) data of an unknown object. The formalism can be applied to a normalization array estimation instead of to attenuation. The normalization array can be further constrained by crystal efficiencies as unknowns. Therefore, scanner performance can be monitored as patient scans occur, eliminating the need for the frequent calibration scan. We propose a monotonic iterative algorithm for the estimation of crystal efficiencies simultaneous with patient activity, based on the TOF Poisson Maximum Likelihood Model. The method produces a non-negative efficiencies solution and is applicable to compressed data, such as azimuthally (possibly TOF) mashed and axially (possibly TOF) rebinned data. Patient data from a Siemens mCT scanner were used to validate the algorithm's performance. The proposed method produced crystal efficiencies maps similar to those of the daily scans.