Quantification of regional perfusion, shunt fraction and ventilation using positron emission tomography : a nonlinear tracer kinetics model

Abstract

Quantification of gas exchange in the lung using Positron Emission Tomography (PET) allows validation of mechanistic hypotheses regarding the local regulation of pulmonary function. We have developed a nonlinear lumped-parameter model to simulate the regional kinetics of the tracer 13N2 as measured by PET following a rapid intravenous injection of a bolus of 13N2 in saline solution. The model is made of a compartment representing the right heart and pulmonary vasculature through which the tracer is mixed and diluted in blood. From this compartment, blood is distributed to two compartments for each defined region of interest (ROI). One lumps all aerated alveolar units, and the other lumps all atelectatic, edematous or fluid filled alveolar units with no gas content. Differential equations of the model were solved numerically and the parameters estimated using gradient-descent search algorithms. The model was applied to PET data from supine sheep to study the effects of partial liquid ventilation (PLV) in bilaterally surfactant-depleted lungs. Images were divided in three ROI's along the vertical axis: nondependent, middle and dependent. It was concluded that PLV reduces shunt in the surfactant-depleted lung. In a second chapter, a new methodology to estimate the fractal dimension of two dimensional images is described. The method was applied to PET images of perfusion obtained from normal dogs. A relationship between the fractal dimension and Power Spectral Density was studied, leading to new insights for the physical interpretation of the fractal dimension.