Initial Tests for a Wrist-Based Quantitative PET Scanner: Flow Rates, Activity, Geometry, and Scatter
Abstract
Purpose
To identify and characterize how blood flow, activity concentration, geometry, and scatter affect the counting abilities of a prototype device to quantitatively measure PET radiotracer blood input function at the wrist.
Methods
A model system for measuring FDG in the wrist was constructed using 0.7 mm tubing and two sodium iodide detectors. A syringe autoinjector controlled the flow of FDG through the tubing at various rates (400 – 800 mL/min for a 1 mm syringe) to model blood flow. Spectra were collected in 10 second intervals over three minutes. Performance at different activity levels was also assessed. Scatter was modeled by passing the tubing through an acrylic cylinder; spectra with and without the acrylic scatterer were compared. Statistical analysis was performed in MATLAB.
Results
Performance was assessed in each trial by summing counts in each spectrum and comparing measured count rates to expected count rates. Regression analysis found no significant correlation between peak counts and blood flow (R2 = 0.169), which was confirmed with ANOVA (p-value = 0.492), but did find that the count rate as a percentage of expected rate was correlated with the blood flow (R2 = 0.828, ANOVA p-value = 0.032). Scatter increased counts by more than 15% compared to the absence of scattering material. Counts in spectra scaled as expected with activity.
Conclusion
This work demonstrated the factors that must be considered to dynamically measure PET tracer in the wrist for quantitative modeling of tracer uptake. Blood flow rate, scatter, and activity concentration all influence the relative measured count rate as compared to true count rate, indicating that the proper quantitation from this system must account for these factors.