Abstract
1292
Objectives This investigation was designed to explore the relationship between the amount of activity injected and the accuracy and precision of resulting kinetic modelling parameters for the D2/D3 PET radiotracer [11C]PHNO. The objective was to provide informed guidance for protocol design and execution.
Methods List mode data were acquired from a single [11C]PHNO dynamic human PET scan (acquired with a 211 MBq injection). The data set was subdivided by dealing out the recorded events into smaller data sets corresponding to simulated independent realisations of lower activity scans, and then reconstructed as though each were a separate acquisition. Dynamic PET images were produced corresponding to 106, 53, 26 and 13 MBq. All dynamic image sequences were analysed using an image analysis pipeline involving motion correction, registration to MRI and region-based kinetic analysis using the simplified reference tissue model to produce binding potential estimates in six brain regions: substantia nigra, globus pallidus, ventral striatum, thalamus, caudate and putamen.
Results Decreases in injected activity had a more significant effect on regions with low binding and / or low volumes. For example, for the substantia nigra the estimated coefficient of variation ranged from 13% for a 211 MBq scan to 48% for 13 MBq. If the desired COV were, say, 5% then injected activities greater than 211 MBq would be required for satisfactory quantification in this region. For thalamus, the 5% COV threshold was between 105 and 211 MBq, whereas for putamen the COV could be acceptable with activities as low as 13 MBq.
Conclusions A deeper understanding of the regional effect on outcome measures as a function of injected activities would enable more optimal design of PET studies. This could also provide valuable information that would allow a clinician to make more informed decisions on whether to administer radioactive doses and generalises to other tracers