Simple and Patlak models for myocardial blood flow measurements with nitrogen-13-ammonia and PET in humans

The Simple and Patlak models for estimating myocardial blood flow with N-13-ammonia have become attractive for clinical applications with PET because of their simplicity and ease of implementation. However, these models are sensitive to factors such as the data acquisition times and data integration times, which can cause errors in the estimation of myocardial blood flow, as demonstrated in this study. Limiting the application of these models to specific conditions can minimize the errors. Methods: Dynamic PET images of the uptake of N-13-ammonia in the heart were obtained in seven humans under rest and dipyridamole stress. Myocardial blood flow was estimated using the Simple and Patlak models for different data acquisition times and data integration times. Blood flow values were compared to flow values computed with the two-compartment model as a reference. Results: Blood flow values calculated with the Simple and Patlak models during the first 2 min of data acquisition were closely correlated to the two-compartment model values. Longer acquisition times resulted in significant underestimation of blood flow for the Simple model. Long integration times of greater than 60 sec also resulted in significant underestimation of blood flow for both models. Conclusion: The Simple and Patlak models produce estimates of myocardial blood flow that are well correlated with the two-compartment model estimated blood flows for the integration time of 60 sec from 60 to 120 sec postinjection, Because of the errors associated with longer data acquisition times and longer integration times, use of these models should be limited to a well-documented data acquisition paradigm.