Tracer kinetic model of regional pulmonary function using positron emission tomography

<jats:p> To determine the spatial distributions of pulmonary perfusion, shunt, and ventilation, we developed a compartmental model of regional <jats:sup>13</jats:sup>N-labeled molecular nitrogen (<jats:sup>13</jats:sup>NN) kinetics measured from positron emission tomography (PET) images. The model features a compartment for right heart and pulmonary vasculature and two compartments for each region of interest: 1) aerated alveolar units and 2) alveolar units with no gas content (shunting). The model was tested on PET data from normal animals (dogs and sheep) and from animals with experimentally injured lungs simulating acute respiratory distress syndrome. The analysis yielded estimates of regional perfusion, shunt fraction, and specific ventilation with excellent goodness-of-fit to the data ( R <jats:sup>2</jats:sup> > 0.99). Model parameters were estimated to within 10% accuracy in the presence of exaggerated levels of experimental noise by using a Monte Carlo sensitivity analysis. Main advantages of the present model are that 1) it separates intraregional blood flow to aerated alveolar units from that shunting across nonaerated units and 2) it accounts and corrects for intraregional tracer removal by shunting blood when estimating ventilation from subsequent washout of tracer. The model was thus found to provide estimates of regional parameters of pulmonary function in sizes of lung regions that could potentially approach the intrinsic resolution for PET images of <jats:sup>13</jats:sup>NN in lung (∼7.0 mm for a multiring PET camera). </jats:p>