A new scatter compensation method for Ga-67 imaging using artificial neural networks

A new scatter correction method for Ga-67 based on artificial neural networ ks (ANN) with error back-propagation was designed and evaluated. The ANN co nsisted of a 37-node input layer (37 energy channels in the range 60-370 ke V), an 18-node hidden layer, and a 3-node output layer to estimate the scat ter-free distribution in the 93-, 185-, and 300-keV photopeaks. Two separat e activity and attenuation distribution sets, based on a segmented realisti c anthropomorphic torso phantom, were simulated. The first set was used for ANN learning and the second to evaluate the scatter correction. Our Monte Carlo simulation modeled all photon interactions in the patient, collimator , and detector. Interactions simulated in the collimator included Compton a nd coherent scatter and photoelectric absorption with forced production of lead K-shell X rays. Ninety very high count projections were simulated and used as a basis for generating 15 Poisson noise realizations for each angle ; noise levels were characteristic of 72-h post-injection Ga-67 studies. Th e energy window images (WIN) used clinically were also generated for compar ison. Bias and variance were computed with respect to the primary distribut ions over reconstructed volumes of interest in the lungs, abdomen, liver, a nd tumors. ANN overall bias in all structures was less than 16% (8% in the abdomen) as compared to 85% with WIN. The variance of the activity estimate s was systematically greater with WIN than ANN. ANN is a promising approach to scatter correction in Ga-67 studies.