Abstract
581
Objectives To investigate and optimize acquisition time of dynamic 18F-FLT PET for accurate quantification of uptake/proliferation kinetics in order to reduce patient burden and PET system time. To establish an optimized dynamic protocol for FLT imaging to assess therapy response.
Methods 30min-dynamic FLT PET scans were acquired on a Philips Gemini TF 64 system in continuous list mode. PET data were reconstructed following a 26-frame protocol (8x15s, 6x30s, 5x1min, 5x2min, 2x5min). Maximum activity concentrations (Bq/mL) of both tumors and plasma in the descending aorta were obtained with 3D Volume of Interest placement along time frames. Ki values were calculated using Patlak Analysis by in-house developed software based on different data acquisition durations (Td) (10, 12, 14, 16, 20, 25 and 30min, separately) and different linear regression onset time (T0) points (0, 1, 2, 3, 4 and 5min). Ki of the 30min-data set was taken as the gold standard and compared with the rest shortened data series. Pearson product-moment correlation coefficient (R) of 0.9 was chosen as a limit for the correlation coefficient. A total of 26 data sets were evaluated.
Results Correlations of Ki between the 30min-dynamic PET and the shortened dynamic PET series increased as the increase of Td from 10 to 25min. It revealed poor correlations for PET with Td ≤14min (R=0.57±0.09, 0.73±0.04, 0.84±0.02 at Td=10, 12 and 14min, respectively), and excellent correlations when Td of PET ≥16min regardless of T0 value (R=0.90±0.01, 0.97±0.00, 1.00±0.00 under Td of 16, 20, 25min, respectively).
Conclusions A 16min-dynamic 18F-FLT PET acquisition is long enough to predict accurate Ki values of Patlak analysis. This study demonstrates the ability of shortened FLT dynamic PET imaging without influencing the kinetic quantification for assessment of changes for response assessment making the use of dynamic acquisitions more clinically feasible.
Research Support OIRAIN TECH 09-028