Study of small animal PET imaging optimization

• Κύριος συγγραφέας: Φώτος, Νικόλαος
• Άλλοι συγγραφείς: Κωσταρίδου, Ελένη
• Μορφή: Thesis
• Γλώσσα: English
• Έκδοση: 2016
• Θέματα: Positron emission tomography; NEMA protocol; Τομογραφία εκπομπής ποζιτρονίων; Πρωτόκολλο ΝΕΜΑ; 616.075 75

The hybrid imaging combined by the positron emission (PET: Positron emission Tomography) and computed tomography (CT), PET/CT it is consider to be a very promising technology in medicine. The total time required for a PET imaging, is the sum of the acquisition time in PET procedure, and the scan time in CT procedure. In clinical task, the positron emission tomography follows a variety of acquisition time protocols for the PET procedure, depending on the pathological issue, the organ, and the size of the patient. According to our literature research, it is introduced that there is not an optimum acquisition time for PET. Most of the protocols in preclinical task, introduce methodologies referring to the procedure of an experiment applied by a micro-PET/CT, and the studies are focused in evaluating the performance between different PET/CT systems. The Biomedical Research Foundation of Academy of Athens, in June 2014, added to its facilities, a brand new small animal PET/CT, the nanoScan of Mediso S.A. The aim of this thesis, is to evaluate the performance and stability of the new micro-PET/CT system, for various acquisition times, and the evaluation of an optimum acquisition time in preclinical task. This effort is focused on the investigation of the effect of acquisition time in image quality. Two experiments were performed, the one with phantom, and a second with mice. In addition, a Matlab code, was developed, for all the experiments, for recording and processing measurements of image quality characteristics. The Matlab results are in agreement with the System’s software. The phantom study, focused on the evaluation of performance of the systems from the scope of image quality, hence to ensure the stability of the new system over the time. Three scannings were performed with the same phantom, without changing the acquisition and reconstruction parameters. This was an effort for evaluating the precision of our measurements. For the image reconstruction in phantom studies so as in animal studies, was used the reconstruction algorithm OSEM 3D, with 4 iteration – 4 Subjects, in low Regularization level. The guide to our experiment was the NEMA-NU4 2008, using the IQ-Phantom as an object of study, and a radionuclide 18F, with a 3.7 MBq initial activity. It was followed an acquisition time method of 100 min List-Mode, divided in five time frames of 10,15,20,25, and 30 min, for which, five reconstruction was applied, by alternating the order of the frames, in a duration of 100 min List –Mode. In the phantom experiment, the stability of the system was evaluated by the reaction of the average activity concentration over the time, also the noise of the reconstructed image over the time was calculated, in the uniform region of the IQ-Phantom. Particularly, the average activity concentration found approximately equal to 191 KBq/ml, with a percentage standard deviation %STD ~4%. The results are in agreement with the NEMA protocol and the literature [42]. The dependence of average activity concentration from the total time acquisition, was displayed in diagrams (Average Activity Concentration vs Total Time Acquisition) for all of the reconstruction sets. The average activity concentration was not affected dramatically over the time. The %STD, it is an important characteristic of uniformity quality of the reconstructed image, and is indicative of the noise. Furthermore, the graphs %STD vs Total Time acquisition, presented that the noise is decreasing while the acquisition time increasing, a condition that it is expected. In addition the recovery coefficients were defined by the ratio of the measured activity concentration in the Rods’ region divided by the average activity concentration in the uniform region, and it ranges 0<RC≤1, with an ideal value equal to one RC=1. Indicatively, an average of the values are 0.04, 0.3, 0.7, 0.85, 0.9 for the rods with diameter of 1mm, 2mm, 3mm, 4mm, 5mm representatively and 10 mm long (z-axis) each. According to the characteristic curves of RC vs Rod diameter, for the five time frames, it is obvious that each curve overlap the other, and this encourage us to say that the RC is not differentiated significantly from 10 min to 30 min acquisition frame. Moreover, the Spill-Over ratio was defined by the ratio of the average measured activity of the cold region divided by the average activity concentration in the uniform region, and the values are reported as 0.11 for the water region and 0.12 for the air region. SOR ranges 0<SOR<1, with and ideal value SOR=0 The SOR values did not show significant differentiation over the time. The scope of the animal studies, is to evaluate the effect of acquisition time on image quality, by calculating the Standardized Uptake Value (SUV), for three different acquisition frames. Six experiments were performed, working on six mice with melanoma, which were injected with an 18F-FDG solution. The acquisition method was a 3D dynamic scan in a 30 min List-Mode. Consequently, three different reconstructions were performed, one for the first 10 min, a second for the first 20 min and a third for the 30 min, in the List-mode of 30 min. For the three reconstructions there was a semi quantitative analysis, by calculating the following SUVs of the reconstructed image: the max value SUVmax of the VOI, and four SUVsmean that were calculated by applying a threshold of 10%, 40%, 50% and 60% to the SUVmax. Finally, three population were created by the SUVs of the six mice, for the three reconstructions (10 min, 20 min, 30 min), and a Wilcoxon-Test was performed. The Wilcoxon-Test concluded that the three population are significant equal in the 5% significance level. Hence, the different acquisition time of 10, 20, and 30 min do not occur differentiation on the SUV.