Determination of tracer arrival delay with spectral analysis

This work presents a general and fast method for the detection of the tracer arrival delay in brain PET studies. The technique adopts the nonnegative least squares (NNLS) algorithm which is commonly used in the spectral analysis of positron emission tomography (PET) data. The method is able to account for a vascular contribution to the tissue response function and does not involve multiple iterations. No assumptions on the model order of the tissue response function are required. It is based on the whole scanner count rate of true coincidences every second in the field-of-view and the whole blood activity concentration from the on-line blood detector system. The presented method has been used for the analysis of neuroreceptor ligand studies. An illustrative example with the serotonin 5-HT_2A receptor ligand [¹¹C]MDL 100 907, a simulation study and a comparison with the iterative nonlinear least squares procedure are included.