Abstract
Objectives: Determination of kinetic tracer parameters using dynamic PET imaging requires knowledge of the arterial input function (IF), which is most easily acquired from time-activity curves of a blood pool. Due to a large amount of activity in the field of view (FOV) in the initial imaging phases post-injection, dynamic 3D imaging may suffer from significant dead-time (DT) effects. This work quantifies the effect of scanner DT on dynamic myocardial perfusion studies using the left ventricle (LV) as a blood pool.
Methods: A dynamic 3D PET scan was performed with 1150 MBq of 18F-FDG in a 15 ml cylindrical phantom to simulate an injection bolus. The phantom was placed entirely in the PET FOV, with 10-minute frames acquired continuously for 18 hours on a GE Discovery 710 scanner. DT corrected maximum activity concentrations were determined from image analysis (kBq/ml_max) and compared to true values (kBq/ml_true). A model of cardiac 13N-labeled ammonia kinetics was then applied to clinical data from 4 patients (mean activity = 541.7±21 MBq) to demonstrate the effect of DT on determination of global tracer flow. The percentage difference between kBq/ml_max and kBq/ml_true at an activity of 550 MBq was used to correct early phases of imaging assuming all activity in the FOV.
Results: Differences between kBq/ml_max and kBq/ml_true up to 33.2% were observed with 1150 MBq of activity, which reduced to 2% at 400 MBq. Error of 8.4% in activity concentration was observed at 550 MBq. Using the phantom results to correct the initial phases (30 seconds) of LV IF for 4 patients resulted in 8.9±0.6% higher global cardiac tracer flow compared to the uncorrected IF.
Conclusions: Our results indicate that scanner DT can contribute to error in the calculation of kinetic parameters from 3D cardiac studies. The error will be greater with higher activity studies such as 15O (1500 MBq), and 11C-acetate (750 MBq) and may lead to much larger error in derived kinetic variables.
Methods: A dynamic 3D PET scan was performed with 1150 MBq of 18F-FDG in a 15 ml cylindrical phantom to simulate an injection bolus. The phantom was placed entirely in the PET FOV, with 10-minute frames acquired continuously for 18 hours on a GE Discovery 710 scanner. DT corrected maximum activity concentrations were determined from image analysis (kBq/ml_max) and compared to true values (kBq/ml_true). A model of cardiac 13N-labeled ammonia kinetics was then applied to clinical data from 4 patients (mean activity = 541.7±21 MBq) to demonstrate the effect of DT on determination of global tracer flow. The percentage difference between kBq/ml_max and kBq/ml_true at an activity of 550 MBq was used to correct early phases of imaging assuming all activity in the FOV.
Results: Differences between kBq/ml_max and kBq/ml_true up to 33.2% were observed with 1150 MBq of activity, which reduced to 2% at 400 MBq. Error of 8.4% in activity concentration was observed at 550 MBq. Using the phantom results to correct the initial phases (30 seconds) of LV IF for 4 patients resulted in 8.9±0.6% higher global cardiac tracer flow compared to the uncorrected IF.
Conclusions: Our results indicate that scanner DT can contribute to error in the calculation of kinetic parameters from 3D cardiac studies. The error will be greater with higher activity studies such as 15O (1500 MBq), and 11C-acetate (750 MBq) and may lead to much larger error in derived kinetic variables.
| Original language | English |
|---|---|
| Title of host publication | The Journal of Nuclear Medicine |
| Pages | 605 |
| Volume | 55 |
| Publication status | Published - May 2014 |