Adaptive Fourier rebinning for SPECT simultaneous transmission/emission protocol (STEP)

The adaptive Fourier rebinning algorithm (AFR) for SPECT STEP is a reconstruction technique that is similar to PET Fourier rebinning, in that a transfer function (TF) is constructed to connect the 3D Fourier coefficients of the projection sinograms with the 3D Fourier-Bessel coefficients of the object. This method may also be applied to the X-ray transform, and the authors have applied it to the reconstruction of transmission data in STEP as well. They demonstrate the use of this approach with emission and transmission studies from a Picker Prism-2000 STEP. The AFR method compared favorably with MLEM for both transmission and emission reconstruction For transmission reconstruction, AFR produced images that were quantitatively more accurate overall, with a lower noise level. The AFR provides objective, intelligent criteria to determine the amount of regularization to optimally balance noise and resolution in count-deficient studies. There criteria are based on total counts and attenuation. This is especially important to the scanning of large patients, in which attenuation reduces the number of photons available. The emission MLEM reconstructions were much noisier than AFR, and could not reliably detect a cold lesion in the phantom, whereas AFR was able to detect the lesion.