A short-scan method for k[sub]3 estimation with moderately reversible PET ligands: Application of irreversible model to early-phase PET data

• Published in: NeuroImage (Orlando, Fla.) Vol. 59; no. 4; pp. 3149 - 3158
• Main Authors: Sato, Koichi, Fukushi, Kiyoshi, Shinotoh, Hitoshi, Shimada, Hitoshi, Tanaka, Noriko, Hirano, Shigeki, Irie, Toshiaki
• Format: Journal Article
• Language: English
• Published: 15.02.2012
• Subjects: Alzheimer's disease
• ISSN: 1053-8119
• DOI: 10.1016/j.neuroimage.2011.10.087

Long dynamic scans (60-120 min) are often required for estimating the k[sub]3 value, an index of receptor density, by positron emission tomography (PET). However, the precision of k[sub]3 is usually low in kinetic analyses for reversible PET ligands compared with irreversible ligands. That is largely due to unstable estimation of the dissociation rate constant, k[sub]4. We propose a novel '3P+' method for estimating k[sub]3 of moderately reversible ligands, where a 3-parameter model without k[sub]4 is applied to early-phase PET data to obtain a good model-fit of k[sub]3 estimation. By using [[super]11C] Pittsburgh compound B (PIB) (k[sub]4 = 0.018/min) as an example of a moderately reversible ligand, the 3P+ method simulation with a 28 min PET scan yielded less than 3% k[sub]3 relative bias with a +100% k[sub]3 change. In [[super]11C]PIB PET scans of 15 normal controls (NC) and nine patients with Alzheimer's disease (AD), the 3P+ method provided a precise k[sub]3 estimate (mean SE of 13.6% in parietal cortex; covariance matrix method). The results revealed linear correlations (r = 0.964) of parietal k[sub]3 values in 24 subjects between 28 minute 3P+ method and conventional 90 minute 4-parameter method. A good separation of k[sub]3 between NC and AD groups (P < 0.001; t-test) was replicated in 28 minute 3P+ method. The short-scan 3P+ method may be a practical alternative method for analyzing reversible ligands.