HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Guo, Q. Articles by Gunn, R. PubMed Articles by Guo, Q. Articles by Gunn, R.

Predicting the in vivo performance of PET radioligands through in silico modelling

¹ Department of Engineering Science, University of Oxford, Oxford, United Kingdom; ² GlaxoSmithKline Clinical Imaging Center, London, United Kingdom

184

Objectives: The development of a successful radioligand requires a number of different criteria to be met. Here, a bio-mathematical model and a Monte-Carlo approach are used to derive a metric of radioligand performance. The long term goal is to predict the likely success of unlabelled radioligands from a library of candidates. Here, we evaluate the metric's utility and consider its performance for a range of established radioligands (N=22).

Methods: The model assumes that radioligand behaviour can be described by a single tissue compartment model in both reference/target regions and requires an estimate of influx rate (K₁), efflux rate (k₂) and binding potential (BP_ND) in order to determine the coefficient of variation of BP_ND (%COV) based on a Monte Carlo approach. A %COV volume was generated for the K₁, k₂ and BP_ND parameter space using a standard input function.

Results: The %COV decreased monotonically with an increase in K₁, and had a parabolic relationship with k₂ and BP_ND, consistent with a priori expectation. The model classified the established radioligands according to their %COV, as good (0-5%), intermediate (5-10%), and bad (>10%). Widely accepted radioligands such as ¹¹C-Flumazenil (2.46%), ¹¹C-Raclopride (3.85%), ¹¹C-WAY100365 (3.48%) and ¹¹C-NNC 112 (4.46%) were marked as good ligands by the model, whilst ¹⁸F-MPPF (6.23%) and ¹⁸F-CPFPX (6.95%) were marked as intermediate, and ¹¹C(S,S)-MRB (16.56%) and ¹¹C(R)-PK11195 (22.18%) were classified as bad ligands.

Conclusions: The model predictions are consistent with generally accepted ratings for these radioligands. In silico models to predict K₁, k₂, and BP_ND from in silico/in vitro data are now being developed so that the technique may be used to aid the discovery of novel molecular imaging probes.

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Guo, Q. Articles by Gunn, R. PubMed Articles by Guo, Q. Articles by Gunn, R.