Application of correlated component analysis to dynamic PET time-activity curves denoising.

Positron emission tomography (PET) is a physiological, non-invasive imaging technique, which forms an essential part of nuclear medicine. The data obtained in a PET scan represent the concentration of an administered radiotracer in tissues over time. Quantitative analysis of PET data makes possible the assessments of in-vivo physiological processes. The Logan graphical analysis (LGA) is one of the methods that are used for quantitative analysis of PET data. LGA transforms PET data into a simple linear relationship. The slope of the LGA linear relationship is a physiological quantity denoting receptor availability. This quantity is termed distribution volume ratio (DVR). LGA-based estimates of the DVR are negatively affected by the noise in PET data -leading to the DVR being underestimated. A number of approaches proposed to address this issue have been observed to reduce the bias at the cost precision. An alternative regression method, least-squares cubic (LSC), was recently applied to estimate the DVR in order to reduce the bias. LSC was observed to reduce the bias in the LGA-based estimates. However, slight increases were also observed in the variance of the LSC-based estimates. This calls for methods to act against the variance in the LSC-based estimates. In this study, an alternative method is applied for tTAC denoising. This method is referred to as correlated component analysis (CorrCA). CorrCA transform the data by searching for dimensions of maximum correlation. This technique is closely related to other well-known methods such as principal component analysis and independent component analysis. In this study, the data were denoised by CorrCA (to act against the variance in the estimate) and the DVR was estimated by LSC, which provides for minimal bias. The resulting method LSC-CorrCA, gave less-biased estimated with increased precision. This was observed for both simulation results as well as for clinical data, both for 11C Pittsburgh compound B. Simulation data revealed reduced variances in LCS-CorrCA-based estimates, and the clinical data showed improved contrast between gray and white matter regions.Clinical Relevance-Improved DVR estimates would ease the interpretation of medical images, which will in turn positively influence the clinical processes, from diagnosis to treatment and follow-ups.

Improving contrast between gray and white matter of Logan graphical analysis' parametric images in positron emission tomography through least-squares cubic regression and principal component analysis.

Logan graphical analysis (LGA) is a method forin vivoquantification of tracer kinetics in positron emission tomography (PET). The shortcoming of LGA is the presence of a negative bias in the estimated parameters for noisy data. Various approaches have been proposed to address this issue. We recently applied an alternative regression method called least-squares cubic (LSC), which considers the errors in both the predictor and response variables to estimate the LGA slope. LSC reduced the bias in non-displaceable binding potential estimates while causing slight increases in the variance. In this study, we combined LSC with a principal component analysis (PCA) denoising technique to counteract the effects of variance on parametric image quality, which was assessed in terms of the contrast between gray and white matter. Tissue time-activity curves were denoised through PCA, prior to estimating the regression parameters using LSC. We refer to this approach as LSC-PCA. LSC-PCA was assessed against OLS-PCA (PCA with ordinary least-squares (OLS)), LSC, and conventional OLS-based LGA. Comparisons were made for simulated11C-carfentanil and11C Pittsburgh compound B (11C-PiB) data, and clinical11C-PiB PET images. PCA-based methods were compared over a range of principal components, varied by the percentage variance they account for in the data. The results showed reduced variances in distribution volume ratio estimates in the simulations for LSC-PCA compared to LSC, and lower bias compared to OLS-PCA and OLS. Contrasts were not significantly improved in clinical data, but they showed a significant improvement in simulation data -indicating a potential advantage of LSC-PCA over OLS-PCA. The effects of bias reintroduction when many principal components are used were also observed in OLS-PCA clinical images. We therefore encourage the use of LSC-PCA. LSC-PCA can allow the use of many principal components with minimal risk of bias, thereby strengthening the interpretation of PET parametric images.