Validation of MRI/SPECT similarity-based registration methods using realistic simulations of normal and pathological SPECT data

We propose a validation methodology to study the behaviour of SPECT/MRI similarity-based registration methods in normal and pathological conditions. Validation data sets were computed by simulating realistic SPECT data from 3D Tl-weighted MRI data, through Monte Carlo simulations, using a perfusion model built from measurements made on real SPECT data (normal or pathological). Validation of four similarity-based registration methods performed on these simulated data showed a registration mean accuracy significantly lower than simulated SPECT spatial resolution.