PT - JOURNAL ARTICLE AU - Arman Rahmim AU - Katherine Dinelle AU - Sarah Lidstone AU - Stephan Blinder AU - Ju-Chieh (Kevin) Cheng AU - Geoffrey Topping AU - Olivier Rousset AU - Dean Wong AU - Vesna Sossi TI - Improved dynamic imaging using a novel event-driven motion correction method in high resolution human brain PET DP - 2007 May 01 TA - Journal of Nuclear Medicine PG - 416P--416P VI - 48 IP - supplement 2 4099 - http://jnm.snmjournals.org/content/48/supplement_2/416P.1.short 4100 - http://jnm.snmjournals.org/content/48/supplement_2/416P.1.full SO - J Nucl Med2007 May 01; 48 AB - 1738 Objectives: A new event-driven motion correction method for PET brain imaging that includes all detected lines of response (LOR) with appropriate detection sensitivity weighting (S-LOR) was validated with phantom studies [Rahmim, 2006]. Here we are testing the S-LOR method on human dynamic PET data and comparing the results against the cases of no motion correction (NoMC) and conventional event-driven motion correction (C-LOR) [Bloomfield, 2003], which neglects LORs that move in/out of the field of view due to motion. Methods: We examined 8 data sets acquired from 4 subjects with Parkinson's disease on a Siemens High Resolution Research Tomograph. Following injection of 11C-raclopride, subjects were scanned for 60 minutes and the data were reconstructed using NoMC, C-LOR, and S-LOR methods. Head motion was measured using a Polaris infrared tracking device. Motion correction methods were compared by evaluating time-activity curves (TAC) and binding potential (BP) values derived from regions of interest (ROI) in the striatum. Eight circular ROIs were placed on the striatum bilaterally (1 caudate, 3 putamen). A χ2 of the TAC points for each ROI was calculated with respect to a curve fit via a modelling approach to the measured data. BPs were calculated using a tissue input Logan analysis with the cerebellum as the reference region (BP = Bmax/Kd). Results: Data reconstructed using the C-LOR approach resulted in clearly visible image artifacts, and thus were not considered further. The χ2 obtained for the TACs derived from the S-LOR corrected images were significantly lower compared to the NoMC case (p ≤ 0.05 by Wilcoxon signed-rank test). A change between the BP values obtained with the NoMC and S-LOR reconstructed images was observed but did not reach statistical significance. Conclusions: The C-LOR method proved inadequate for the motion correction task. The decreased TAC χ2 observed using the S-LOR method indicates a positive effect of the motion correction; further studies are currently underway to ensure that the S-LOR derived BP values best represent the underlying biology. Research Support (if any): NSERC, MSFHR, CIHR