Presentation Biomathematical modeling approach to predict clinical SUVR in amyloid PET imaging towards efficient radioligand discovery and development

荒川, 悠馬  ,  志田原, 美保  ,  Hwey Nai, Ying  ,  古本, 祥三  ,  関, 千江  ,  岡村, 信行  ,  田代, 学

2015-10-11
Description
Biomathematical modeling approach to predict clinical SUVR in amyloid PET imaging towards efficient radioligand discovery and developmentYuma Arakawa1, Miho Shidahara1, 2, YingHwey Nai3, Shozo Furumoto4, Chie Seki5, Nobuyuki Okamura6, Manabu Tashiro2, Yukitsuka Kudo7, Kazuhiko Yanai6 Kohsuke Gonda1 and Hiroshi Watabe31.Division of Medical Physics, Tohoku University School of Medicine, Sendai, JAPAN.2.Division of Cyclotron Nuclear Medicine, Cyclotron and Radioisotope Center, Tohoku University, Sendai, JAPAN.3.Division of Radiation Protection and Safety control, Cyclotron and Radioisotope Center, Tohoku University, Sendai, JAPAN.4.Division of Radiopharmaceutical Chemistry, Cyclotron and Radioisotope Center, Tohoku University, Sendai, JAPAN.5.Biophysics Program, Molecular Imaging Center, National Institute of Radiological Sciences, Chiba, JAPAN6.Department of Pharmacology, Tohoku University School of Medicine, Sendai, JAPAN.7.Innovation of New Biomedical Engineering Center, Tohoku University, Sendai, JAPAN.Abstract (<400words)Aim: Purpose of the study is to develop a new methodology to predict clinical SUVR of amyloid PET probes by extending biomathematical modeling, which was previously proposed by Guo et al. [1]. Methods: 6 amyloid imaging agents, [11C]PIB, [11C]BF-227, [11C]AZD2184, [18F]FACT, [18F]flobetapir and [18F]AZD4694 were conducted in this study.In the methodology, for each tracers, time-activity curves (TACs) with one-tissue compartment model were generated using arterial plasma input function and kinetic parameters (K1, k2 and BPND). In detail, K1, k2 and BPND were predicted by biomathematical modeling approach using lipophilicity (logP), apparent volume (Vx), free fraction in plasma (fP), free fraction in tissue (fND), dissociation constant (KD) and the density of Amyloid  (Bavail) [1]. LogP was computed in two ways, ClogP by chemoffice ver. 2012 (HUlinks Inc.) and moriguchi method logP (MlogP) by dproperties (Affinity science corp.). Vx was also computed by dproperties. Both fP and fND is calculated by relational expressions among logP, fND and fP. Regression lines of logP vs. fND and fND vs. fP. were derived from logP, fND and fP. in three publications, Guo et al. [1], Summerfield et al. [2] and Wan et al.[3], respectively. KD was refered from publications for each tracers. Bavail was fixed 3nM for healthy control, HC and 50nM for severe Alzheimar Disease (AD) patient. Predicted SUVRs of HC and AD were regarded as dividing summed TACs in the target over that in the reference region.Predicted SUVRs of HC and AD were also compared each tracer with in vivo SUVRs of HC and AD groups which were previously reported. These correlations were compared in 6 combinations of logP and regression line (ClogP-Guo’s, ClogP-Summerfield’s, ClogP-Wan’s, MlogP-Guo’s, MlogP-summerfield’s and MlogP-Wan’s).Results: Good correlations between predicted SUVR(y) and in vivo SUVR(x) were observed in case of both MlogP-Summerfield’s (y = 1.17 x -0.16, r2 = 0.71) and MlogP-Wan’s (y = 2.96 x – 1.95, r2 = 0.70). On the other hands, poor correlation was observed in case of ClogP-Guo’s (y=0.59x+0.36, r2=0.38).Conclusion: Proposed methodology (MlogP-Summerfield’s and MlogP-Wan’s) predicts SUVR with good correlation against in vivo SUVR for 6 tracers and may have the potential to apply to other amyloid radioligands as well. Reference:[1] J Nucl Med, 2009, 50(10):1715-23.[2] J Pharmacol Exp Ther, 2006; 316:1282–90[3] J Med Chem, 2007, 50:4606-15
Annual Congress of the Europran Association of Nuclear Medicine

Number of accesses :  

Other information