EXPECTATION MAXIMIZATION RECONSTRUCTION OF POSITRON EMISSION TOMOGRAPHY IMAGES USING ANATOMICAL MAGNETIC-RESONANCE INFORMATION

Using statistical methods the reconstruction of positron emission tomo graphy (PET) images can be improved by high-resolution anatomical info rmation obtained from magnetic resonance (MR) images, We implemented t wo approaches that utilize MR data for PET reconstruction, The anatomi cal MR information is modeled as a priori distribution of the PET imag e and combined with the distribution of the measured PET data to gener ate the a posteriori function from which the expectation maximization (EM)-type algorithm with a maximum a posteriori (MAP) estimator is der ived, One algorithm (Markov-GEM) uses a Gibbs function to model intera ctions between neighboring pixels within the anatomical regions, The o ther (Gauss-EM) applies a Gauss function with the same mean for all pi xels in a given anatomical region, A basic assumption of these methods is that the radioactivity is homogeneously distributed inside anatomi cal regions, Simulated and phantom data are investigated under the fol lowing aspects: count density, object size, missing anatomical informa tion, and misregistration of the anatomical information, Compared with the maximum likelihood-expectation maximization (ML-EM) algorithm the results of both algorithms show a large reduction of noise with a bet ter delineation of borders, Of the two algorithms tested, the Gauss-EM method is superior in noise reduction (up to 50%). Regarding incorrec t a priori information the Gauss-EM algorithm is very sensitive, where as the Markov-GEM algorithm proved to be stable with a small change of recovery coefficients between 0.5 and 3%.