Extraction of a Plasma Time-Activity Curve From Dynamic Brain PET Images Based on Independent Component Analysis

Abstract

A compartment model has been used for kinetic analysis of dynamic positron emission tomography (PET) data [e.g., 2-deoxy-2-18F-fluoro-D-glucose (FDG)]. The input function of the model [the plasma time-activity curve (pTAC)] was obtained by serial arterial blood sampling. It is of clinical interest to develop a method for PET studies that estimates the pTAC without needing serial arterial blood sampling. For this purpose, we propose a new method to extract the pTAC from the dynamic brain PET images using a modified independent component analysis [extraction of the pTAC using independent component analysis (EPICA). Source codes of EPICA are freely available at http://www5f.biglobe.ne.jp/ũkimura/Software/top.html]. EPICA performs the appropriate preprocessing and independent component analysis (ICA) using an objective function that takes the various properties of the pTAC into account. After validation of EPICA by computer simulation, EPICA was applied to human brain FDG-PET studies. The results imply that the EPICA-estimated pTAC was similar to the actual measured pTAC, and that the estimated blood volume image was highly correlated with the blood volume image measured using 15O-CO inhalation. These results demonstrated that EPICA is useful for extracting the pTAC from dynamic PET images without the necessity of serial arterial blood sampling.

Comparison of Eight Methods for the Estimation of the Image-Derived Input Function in Dynamic [18F]-FDG PET Human Brain Studies

“…In particular, we observed underestimated peaks and slower descending slopes. Therefore, the normalization of the estimated curves using the arterial peak, as proposed by Naganawa et al (2005), would have given largely overestimated IFs. Thus, we performed the normalizations by using a blood sample at 30 mins from the injection.…”

“…The dynamic PET time sequence comprised 12 frames of 10 secs each, 2 Â 20 secs, 2 Â 150 secs, 5 Â 5 mins, 1 Â 7 mins, 1 Â 10 mins, and 1 Â 20 mins. Of note, most of the methods tested in the present study use a minimum frame duration of 10 secs or more in the original papers (Litton, 1997;Naganawa et al, 2005;Su et al, 2005;Parker and Feng, 2005;Mourik et al, 2008a;Zanotti-Fregonara et al, 2007). Image reconstructions were the same as for the phantom studies.…”

“…Naganawa et al: Naganawa et al (2005) used an approach based on the ICA. They used the principal component analysis as a first step to reduce the dimensions of the dynamic PET image sequence to two and apply fixed-point algorithm for ICA to extract kinetics.…”

Image-Derived Input Function for Brain PET Studies: Many Challenges and Few Opportunities

“…With the exception of algorithms that work without any prior anatomical assumption, such as blind source separation algorithms when they are used to derive directly the input function through the source signal mixing process (Bodvarsson et al, 2006;Naganawa et al, 2005a), the initial step in calculating IDIF is carotid segmentation, which is required for obtaining raw blood time-activity curves. Carotid segmentation has been performed by both taking advantage of coregistered magnetic resonance images (MRI) (Fung et al, 2009;Litton, 1997;Trebossen et al, 1999), and by placing carotid ROIs directly on PET images (Chen et al, 1998;Liptrot et al, 2004;Mourik et al, 2008a;Su et al, 2005).…”

Arterial Input Function Derived from Pairwise Correlations Between PET-image Voxels

“…The performance of the method was compared with that of three other methods, previously described by Chen et al 20 Mourik et al, 21 and Naganawa et al 22 These methods were selected from the literature since they, among others, have shown the best performance in the estimates of cerebral metabolic rate of glucose. 15 …”