Program Information
Correction of the Intra-Prostatic Spill-in Effect in FCh PET Using Monte Carlo Simulations
S Laberge1*, L Archambault2 , (1) ,,,(2) CHUQ Pavillon Hotel-Dieu de Quebec, Quebec, QC
Presentations
TU-C3-GePD-IT-5 (Tuesday, August 1, 2017) 10:30 AM - 11:00 AM Room: Imaging ePoster Theater
Purpose: In FCh PET imaging of prostate cancer, the strong accumulation of radiotracer within the bladder is suspected to provoke spill-in counts in the neighbouring prostate area. The goal of this study is to image a rudimentary phantom of the lower abdomen using the GATE simulator and to quantify the variation of the prostate signal after cancelling the bladder signal during iterative reconstruction using STIR.
Methods: The model used is made of three embedded water cylinders. A larger cylinder (diameter 250 mm) stands for the patient’s body and provides background activity (5e3 Bq/mL). A second cylinder centered on the first (60 mm) contains varying amounts of bladder activity (1e4,3e4,5e4,7e4 Bq/mL). A third cylinder representing the prostate (40 mm, 2 Bq/mL) is offset vertically by 50 mm and overlaps the bladder axially by 17 mm. Acquisitions of 120 seconds were done for each of the four increasing bladder activities using a validated model of the GE LS scanner, thus imitating a dynamic acquisition. The correction method consists of reconstructing the image a first time, segmenting the bladder signal, projecting it into sinogram space then using it as an additive element in the projection model of the OSEM iterative algorithm (including attenuation correction and scatters/randoms correction).
Results: The mean prostate signal was overestimated by up to 2.76% for the highest bladder activity and the maximum prostate activity was overestimated by up to 46%.
Conclusion: It was demonstrated that the GATE simulator can be used conjointly with the STIR reconstruction software to measure intra-prostatic spill-in counts from the bladder in prostate cancer dynamic acquisitions. A next step would be to replicate this study using realistic phantoms derived from clinical data.
Contact Email: