Introduction
Although many models can be used to describe the unmetabolized tracer function (UF), usually, [carbonyl-11C]WAY-100635 UF is modeled in two different ways: with a linear interpolation or with an exponential model (one exponential plus a constant or a sum of two exponentials). Our goal was to understand if and how much a different UF modeling could affect the quantification of [carbonyl-11C]WAY-100635 images.
Materials & Methods
We considered [carbonyl-11C]WAY-100635 studies coming from 4 young healthy women and, first, modeled UF data with a linear interpolation (Lin) and calculated the relative metabolite-corrected plasma activity curve (Cp_Lin). Then, we modeled the same UF data with an exponential model (Exp), obtaining the relative metabolite-corrected plasma activity curve (Cp_Exp). Successively, we applied spectral analysis1, 2 to characterize the reversible and irreversible components of the system. SA was implemented by using the metabolite-corrected plasma activity curves as input function and considering the following ROIs: Anc (anterior cingulate cortex), Cer (cerebellum), Hip (hippocampus), Lof (lateral orbital frontal cortex), Mtc (mesial temporal cortex), and Occ (occipital cortex).
Results
The two UF models produce two different metabolite-corrected plasma activity curves. In particular, an important difference between Cp_Lin and Cp_Exp was point out in all the subjects during the first 5 minutes of the experiment while was undetectable afterwards. In spite of this, SA indicates that a difference (evaluated as 5 min AUC difference in percent) between Cp_Lin and Cp_Exp ≤ 30% does not have any relevant impact on the quantification of the ROIs. In particular, using Cp_lin or Cp_Exp, SA indicates the presence of the same three components (same amplitudes at the same frequencies) in the intermediate frequency range that corresponds to the presence of three reversible compartments. But, when the difference between Cp_Lin and Cp_Exp is >30%, SA returns two different spectra. In particular, when Cp_lin is used as input function, SA indicates the presence of three distinct components in the intermediate frequencies in all the ROIs. Differently, when Cp_Exp is employed and Anc, Cer and Occ are analyzed, SA till shows the presence of three distinct components in the intermediate frequencies but having amplitudes and frequency position different from whose obtained using Cp_Lin. When Cp_Exp is employed and Hip, Lof and Mtc are considered, SA detects only two distinct components in the intermediate frequencies.
Conclusion
The model chosen to obtain [carbonyl-11C]WAY-100635 metabolite-corrected plasma activity curves has an impact on the quantification of the spectrum components if the difference between Cp_lin and Cp_Exp is greater than 30%. In particular, the use of Cp_lin or Cp_Exp resolves in a different spectrum in terms of number of components and/or their amplitudes and position in the frequency range. When the difference between Cp_lin and Cp_Exp is ≤30%, SA gives the same results. We conclude that UF modeling is critical since it may affect the selection of the compartmental model to quantify [carbonyl-11C]WAY-100635 images.