Respective importance of scatter, attenuation, collimator response and partial volume effect corrections for accurate quantification in /sup 123/I dopamine receptor imaging

Abstract

The authors studied the relative importance of scatter correction (SC), attenuation correction (AC), depth-dependent collimator response (CC) and partial volume effect corrections (PVE) for accurate quantitative /sup 123/I brain SPECT. /sup 123/I activity was simulated in the putamen, caudate nuclei and background of the numerical Zubal brain phantom. Monte Carlo simulations including scatter, attenuation and collimator response effects were performed. The impact of all corrections was studied by reconstructing the projections using 29 schemes involving 2 reconstruction methods (FBP and OSEM), 3 SC (ideal, 20%, Jaszczak), 3 AC (ideal, iterative Chang, OSEM), 1 CC (frequency-distance principle) and 1 PVE correction (anatomically guided pixel-by-pixel algorithm). For each processing scheme, spatial resolution, binding potential (BP) and errors in striatal and background activity estimates were calculated. Results demonstrate: (1) the crucial role of CC for improved spatial resolution hence improved binding potential estimates; (2) the inaccurate striatal activity quantitation using SC+CC+AC only (errors /spl ges/50%); (3) the essential and effective role of PVE correction for accurate striatal activity quantitation (errors <13%). In conclusion, SC, AC and CC all help improve relative and absolute quantitation, but PVE correction is essential for absolute quantitation of striatal activity. PVE correction is effective provided high resolution anatomical information is available.