Input recovery from noisy output measurements: a Monte Carlo method

摘要: Accurate determination of the input function is essential for absolute quantification physiological parameters in PET and SPECT imaging but it requires an invasive tedious procedure blood sampling that impractical clinical studies. We previously proposed a technique simultaneously estimates kinetic from tissue impulse response functions which only two samples. A nonlinear least squares method was used to estimate all fails occasionally due high noise levels data causing ill-conditioned cost function. This study investigates feasibility applying Monte Carlo called simulated annealing Time-activity curves teboroxime, very sensitive changes function, were based on published obtained canine model. The equations describing tracer kinetics different regions minimised by squares. found with are more accurate estimated closely resembled curve. conclude reduces bias estimation