The Patlak graphical analysis technique is a popular tool for estimating blood-to-tissue transfer constants from multiple-time uptake data. Our objective was to extend this technique to tracers with labeled metabolites, the presence of which can cause errors in the standard Patlak analysis.
Methods: Based on previously described formulations, we generalized the graphical technique for use under specific conditions. To test the extended graphical approach, we applied the method to both simulated and patient data using a preliminary compartmental model for the PET tumor proliferation marker, 2-[11C]-thymidine.
Results: When given conditions are met, a linear relationship exists between the normalized tissue activity (tissue activity/blood activity) and a new set of graphical analysis basis functions, including a new definition of normalized time, which takes the presence of labeled metabolites into account. Graphical estimations of the tumor thymidine incorporation rate for simulated data were accurate and showed close agreement to the results of detailed compartmental analysis. In patient studies, the graphical and compartmental estimates showed good agreement but a somewhat poorer correlation than in the simulations.
Conclusion: The extended graphical analysis approach provides an efficient method for estimating blood-tissue transfer constants for tracers with labeled metabolites.