Simultaneous dual-isotope imaging based on an artificial neural network for evaluating myocardial perfusion and fatty acid metabolism

Abstract

BackgroundWe contrived a scatter correction method based on an artificial neural network (ANN) and applied it to the simultaneous evaluation of myocardial perfusion and fatty acid metabolism in single-photon emission computed tomography (SPECT). MethodsThe count data of three energy windows were used as inputs of the ANN. The count ratios of the estimated primary-to-total photons for 99mTc and 123I, which were used to reconstruct 99mTc and 123I images, were calculated using the ANN. In a phantom study, single- and dual-isotope imaging with 99mTc/123I and 201Tl/123I was performed by means of a cardiac phantom simulating patients with and without obesity. In a human study, five normal volunteers and ten patients with myocardial infarction underwent myocardial perfusion and fatty acid metabolism imaging with single and dual SPECT with combinations of 99mTc-methoxyisobutylisonitrile/123I-beta-methyl(p-iodophenyl)pentadecanoic acid (BMIPP) and 201Tl/123I-BMIPP as tracers. ResultsTechnetium-99m yielded more homogeneous images than 201Tl because of the lower degree of photon attenuation, especially in the condition of obese patients, resulting in clearer visualization of the perfusion-metabolism mismatch. Dual 99mTc/123I SPECT offered comparable images with single SPECT in assessing myocardial damage. ConclusionsThe method effectively separated 99mTc and 123I primary photons and proved applicable to 99mTc/123I dual-isotope myocardial SPECT.