Piero Calvini, Daniele Grosso, Maria Antonietta Penco, Sandro Squarcia

Laboratorio di Fisica e Statistica Medica
Dipartimento di Fisica – Università degli Studi di Genova e Sezione INFN di Genova

The spatial resolution displayed by brain SPECT images is rather limited, if compared with the resolution of the corresponding anatomical images (MR or CT images). To some extent this fact limits the diagnostic potential of the functional SPECT scans of the brain. For instance, it may be difficult to distinguish between low tracer uptake due to a functional deficit, where brain tissue still exists, from low uptake generated by focal atrophy, where tissue is lost and replaced by CerebroSpinal Fluid (CSF). In the pathologies where both phenomena occur the interpretation of SPECT images may be problematic. In the last few years the so called iterative reconstruction methods have been found to represent a valuable tool for a thorough exploitation of the informational content of the acquired functional data (the SPECT projections). However, the computational burden associated with these techniques is rather high, since at each iteration step a simulation of the data collection process is performed by applying the so called projector, denoted by A, to the current iterate.