Enhancement of SPECT reconstructions by means of coregistered MR data

The diagnostic potential of brain SPECT imaging is limited by its low spatial resolution. Here an iterative reconstruction technique is proposed which is based on the Conjugate Gradient (CG) algorithm and offers the possibility of improving the spatial resolution of the SPECT activity map of a patient, provided that an MR scan of the same patient is available. The SPECT data are reconstructed twice. The first reconstruction is to permit an accurate coregistration of the MR image to the SPECT image. Then, the coregistered MR image is used as the starting point for the CG algorithm, which solves a "shifted problem". As the iterations proceed, the MR signal of low spatial frequency is replaced by the SPECT signal. However, the high spatial frequency anatomical details contained in the MR image are slowly destroyed by the SPECT projector-backprojector operator since they are nearly outside its range. They remain in the iterates and determine their spatial accuracy. The number of performed iterations and another parameter tune the MR bias on the SPECT reconstruction.