REAL-TIME RECONSTRUCTION AND NOISE PERFORMANCE FOR A TIME-OF-FLIGHT POSITRON EMISSION TOMOGRAPHY SYSTEM
PHILIPPE, EDOUARD ANTOINE
Doctor of Philosophy
Several researchers have proven the benefits of using time-of-flight data in Positron Emission Tomography (TOFPET). Related areas which are investigated here are: (1) The mathematical description of the measurement and image reconstruction processes. (2) The elucidation of the cause of noise performance improvement with utilization of TOF data. (3) The establishment of the optimal mode of utilization of TOF data for the best noise performance. (4) The establishment of the relationship between noise performance and TOF uncertainty. (5) The investigation of alternative modes of data acquisition and image reconstruction for high speed data handling. The above problems were investigated with the aid of a computer simulation, and the following conclusions were reached: (1) The utilization of TOF data in image reconstruction improves the noise performance because noise propagation is restricted to a smaller radius around activity sources than for conventional reconstruction. Thus, image degradation in adjacent structures from any active structure is reduced. (2) Confidence weighting with the probability distribution function of TOF error is the most optimal mode of improving image S/N ratio, within the class of weighting functions studied, and the phantoms utilized. (3) Confidence-weighting provides an improvement over conventional reconstruction that is object-size dependent. For a 40 cm. object a factor of two improvement in the S/N ratio is attained with currently available detectors. With a factor-of-two improvement in detector timing, a further enhancement of a factor of 1.4 can be expected. (4) Real-time reconstruction is possible, but is shown to be non-feasible with current technology, while a technique termed Real-Time Acquisition and Smoothing (RTAS) was shown to be feasible. The advantages of reconstruction using RTAS are three-fold: (1) Immediate visual feedback of image build-up. (2) Substantial savings of image memory. (3) Faster throughput of patients through the system. The implementation of the RTAS reconstruction scheme in the design of the TOFPET scanner currently under development at the University of Texas Medical School at Houston will yield a significantly more practical and economical system for clinical use.
Electronics; Electrical engineering