Evaluation of Optimal Scan Time by Bootstrap Approach for Quantitative Analysis in PET Receptor Study

Abstract

Quantification of the receptor binding potential (BP) in human brain has been performed with positron emission tomography. In this quantitative analysis, the uncertainty in estimated kinetic parameters depends on the SNR. Evaluation of the reliability of parameter estimates is important for the optimization of scan protocol and quantitative analysis methods. However, estimating the reliability is not easy for human data because the true noise level is not precisely known. In this study, we have evaluated a method for estimating the reliability of kinetic parameters with a bootstrap approach for both simulated and human data, and applied this method to evaluating the influence of scan time on the error in BP estimated from PET [11C] raclopride studies. As a result, it was possible to deduce the reliability of kinetic parameter estimates in region-of-interest analysis of human data using the replicated data generated by bootstrap method. In [11C] raclopride studies, the mean estimated BP value did not change when the scan time decreased from 90 to 15 min. However, the uncertainty in BP estimates became larger as the scan time became shorter. With the bootstrap method, it is possible to easily assess the reliability of parameter estimates using only the measured data, and this method can be applied to optimizing scan protocol.