Conventional cardiac PET modeling techniques for [13N]ammonia flow determination do not fully account for the effects of spillover of activity from the right ventricle (RV) onto the activity in the myocardial septum. The purpose of this study was to investigate and to quantitatively account and correct for this effect.
Methods: Simulations were performed to determine the error introduced by conventional quantitation using septal time-activity curves, which only account for left ventricle (LV) spillover. Furthermore, we explored two separate methods to account for the dual spillover problem: direct estimation of the RV and LV spillover fractions incorporated into the [13N]ammonia model by using the LV and RV input functions in the fit and estimation of the relative dispersion and time shift between the LV and RV input functions by fitting using only the LV input function. The simulated curves were fitted using a two-compartment [13N]ammonia model. Flow estimates from the conventional model and the models including either of the two correction procedures were compared with canine microsphere data.
Results: The influence of RV spillover on flow estimation in the septum is determined by several parameters (e.g., dispersion between the RV and LV input function). Depending on the value of these parameters, the septal flow may be underestimated by 0%-30%. The applied methods for correction of the dual spillover problem were comparable and allow for more accurate quantitation in the septum. The canine microsphere data revealed that flow underestimation in the septum is small but significant.
Conclusion: Dual spillover in the myocardial septum can introduce significant errors in the estimation of flow by the conventional [13N]ammonia model fitting method, which does not properly account for the RV spillover. Adjusting for the RV spillover in one of the two proposed methods allows for more accurate quantitation of myocardial septal flow with [13N]ammonia PET data.