TY - JOUR T1 - Role of partial volume correction in paired studies of C-11 PiB vs F-18 florbetapir: implications for centiloid? JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 486 LP - 486 VL - 59 IS - supplement 1 AU - Dean Wong AU - Olivier Rousset AU - Ayon Nandi AU - Michael Pontecorvo AU - Kelly Kitzmiller AU - Paul Rosenberg AU - Josh Roberts AU - Noble George AU - James Brasic AU - Wael Marashdeh AU - Lorena Gapasin AU - Mark Schmidt Y1 - 2018/05/01 UR - http://jnm.snmjournals.org/content/59/supplement_1/486.abstract N2 - 486Objectives: PET imaging is significantly compromised by a number of resolution degrading factors, resulting in undesirable cross-contamination between adjacent functional regions, i.e. the partial volume effect (PVE). Established methods require well-registered, well-segmented anatomic (e.g. MRI) images for partial volume correction (PVC), which in turn can help improve clinical outcome measurements. Methods: We applied the current Geometric Transfer Matrix (GTM)1 to a preliminary dataset of 5 people with Alzheimer’s disease (AD) (Age: M=63.0, SD=10.0, Range=51-74; MMSE: M=18.4, SD=8.2, Range=8-27) and 6 healthy controls (HC) (Age: M=62.8, SD=8.0, Range=51-75; MMSE: M=28.8, SD=1.0, Range=8-27) scanned with both AV45 and PiB (Days Between Scans: M=11.3, SD=7.4; excluding one outlier of 141 days) (Table 1). For both tracers, distribution volume ratio (DVR, reference region cerebellum) was calculated with and without PVC. Results: Our results show that PVC improves the agreement between the two amyloid tracers (AV45 and PiB) and increases the discrimination between AD and HC, in agreement with others2,3,4. Regression analysis (i.e., testing the linear model DVRAV45 = β[DVRPIB] + ϵ ) was used to test the relationship of DVR values across both tracers. The coefficient of determination R2 improved slightly from 0.7 (without PVC) to 0.8 (with PVC), while preserving ß around 0.6 (slope). This improvement in R2 was also seen when looking at AD subjects only (Figure 1). PVC also had a significant effect on the difference between AD and HC, most significantly in motor cortex (p=0.01), limbic region (p=0.04), temporal lobe (p=0.03), parietal lobe (p=0.04) and at a trend level in frontal cortex (p=0.06). Mean DVR values at all combinations of groups/tracers/PVC status is given in (Figure 2) L and R indicate left and right sides, respectively. Conclusions: This finding not only confirms the discriminatory value of dynamic PVC but is an important factor to consider in determining the local centiloid relationship of 18F amyloid tracer with [11C]PiB. References: 1. Rousset OG, et al. J Nucl Med. (1998). 39(5):904-911. 2. Rullman M, et al, J Ncul, Med. (2016) 57(2):198-203 3. Kanetaka H, et al. Eur J Nucl Med Mol Imaging (2004) 31(7):975-980 4. Meltzer CC, et al. J Nucl Med (2004) 41(11):1842-1848 View this table:Table 1: Subject Demographics: Mean +/- SD (Range) ER -