Modeling of carbon-11-acetate kinetics by simultaneously fitting data from multiple ROIs coupled by common parameters.

ABSTRACT

UNLABELLED One of the unique aspects of PET is its ability to noninvasively quantify metabolic processes. Metabolic rate parameters are estimated by fitting the time-activity curves from regions of interest (ROIs) placed on dynamic PET images with a kinetic model. In many cases it is possible to couple these datasets with common parameters, such as the time delay between arrival of tracer in the ROIs and the sampling site. METHODS: Data from eight ROIs placed about images of the myocardium were coupled by the parameters describing the metabolite concentration in the blood. The method was evaluated by comparing estimates of k2 made using the coupled region method and the standard process of fitting data from each region separately. In addition, comparisons were made between estimates of k2 and measured myocardial oxygen consumption. RESULTS: Very little change in mean values of k2 was obtained. The variances, however, were reduced by an average of 37%, compared to the standard method, when the common parameters were not constrained. When the values of the common metabolite parameters were constrained to values previously measured, the average variance in estimates of k2 was reduced by 30%. CONCLUSION: We have demonstrated that the use of this technique can significantly increase the precision of estimates of myocardial oxygen consumption utilizing 11C-acetate PET images. More precise estimates of such quantities can facilitate detection of small regional and/or temporal physiological changes measured with PET. Furthermore, this method can be utilized whenever it is known a priori that one or more kinetic model parameters has the same value for every set of ROI data.