Search

The aim of this study was to develop an accurate simulation and reconstruction model for the Philips Vereos digital whole-body PET/CT scanner. The model was developed using the SimSET Monte Carlo (MC) simulation code, storing all detected photons in SimSET list-mode (LM) format (history files). LM processing, binning and image reconstruction were performed using STIR recently added LM capabilities. In order to accurately process the listmode files generated by SimSET, a new interface between SimSET and STIR was developed. The performance of our MC model was evaluated using the NEMA NU2-2012 standard. Spatial resolution was in good agreement with published results(±12%). Count-rate performance was accurate overall, with NEC rates accurate for both low (32.5% vs. 31.7% at ≈ 5 KBq/mL), and high counts (146.2 kcps vs 153.4 at ≈ 55 KBq/mL), and consistent prompts for all the activity concentrations. LM-MLEM reconstruction including scatter and attenuation correction was performed to validate the SimSET-STIR developed interface.

A tool for the MC simulation of pinhole SPECT cameras that includes the main calibration parameters has been developed and validated.

The presence of the naturally occurring isotope 176Lu gives rise to background count rates in LSO/LYSO crystals to be used in PET and SPECT systems. The aim of this work is to measure 176Lu background count rates and its spatial distributions in a continuous LYSO block. Our results show that the low energy component of 176Lu background is related to partial energy deposition mainly produced near the crystal edge showing a hot-perimeter artifact. Furthermore, 176Lu background rate is observed to be x10 times lower for a 100-180 keV than for a 411-611 keV energy window. Our findings indicate that when using monolithic LYSO crystal blocks, 176Lu background is not expected to significantly deteriorate the performance of single photon detection systems.

Monte Carlo simulations were carried out in order to evaluate the performance of readout solutions for monolithic LYSO blocks, considering several 2D array configurations of discrete SiPMs. Results were compared to conventional multianode PMT readout. Our findings show that a detector module based on an array of 8×8 discrete SiPM may lead to better energy resolution and comparable spatial resolution. Our work shows the feasibility of using discrete SiPMs as readout for monolithic scintillator-based detector modules.

Multimedia applications like high definition digital television push the manufacturers to increase available bandwidth mainly in Powerline Communication (PLC) links. Because of this needs in this work a PLC link control is developed to maximize the information rate meanwhile the bit error rate (BER) is maintained under a desired reference limit. The developed control system has a unique variable called minimal total distance and the feedback loop signal is estimated from the low-density parity check (LDPC) decoder iterations. These mentioned characteristics enable a fast and simple control scheme. In order to evaluate the control system behavior a complete Powerline (PL) link was modeled in a matrix form and simulated with a measured PL channel model and with an additive white Gaussian noise (AWGN) channel. The implemented PL link reaches 570Mbps in AWGN. Finally the BER control shows a good response under a severe random walk SNR channel variation.

Positron Emission Tomography (PET) techniques have gained an important role in the diagnosis of human diseases as well as for pharmaceutical and biomedical research. In order to best optimize their capabilities, evaluation of their performance has to be achieved through standardized methodology from specific standards. The aim of this work was to assess the performance evaluation of the microPET R4 system for imaging rodents according to the new NEMA NU 4-2008 standards and compare it against its previous evaluation according the adapted clinical NEMA NU 2-2001. All measurements were made for an energy window of 350-650 keV. Spatial resolution of 2.12 mm FWHM in the radial direction, 2.66 mm FWHM in the tangential direction and 2.23 mm FWHM in the axial direction was measured in the centre of the axial FOV. Scatter fraction for mouse and rat phantoms were found to be 9.5% and 22% respectively. The maximum noise-equivalent count rate (NECR) was 110 kcps at 121.3 MBq for the rat phantom 287 kcp at 129.8 MBq for the mouse phantom. The absolute sensitivity of the scanner was found to be 1.66% at the centre of the field of view.

SiPM performance dynamic range limitation and possible effect on energy resolution of scintillation light readout has been addressed. Device response has been parametrized in terms of signal dispersion. An expected value of 7% dispersion in SiPM signal response obtained with our parametrization compares with the 8.4% energy resolution observed at a 22Na spectra acquired using a small dynamic range 400 pixel SiPM and a small LYSO crystal.

Monte Carlo simulations were carried out in order to guide the choice of crystal block material and thickness for a small FOV PET/SPECT camera. A continuous crystal block of 50mm x 50mm was simulated considering several thickness (8 mm, 16 mm and 32 mm) and two scintillation crystal types(LSO:Ce and NaI:Tl). Detection module parameters like spatial resolution at centre and border of FOV, efficiency and energy resolution were evaluated. Despite there is a common opinion that LSO is a less valuable material for SPECT detection, our findings demonstrate that LSO crystal blocks have interesting design features for the development of an hybrid SPECT and PET system. Our results show that by adequate corrections of non-linearity of a thick crystal block response, a high detection efficiency can be preserved (nearly 80% for PET and 99.9% for SPECT) without compromising high spatial resolution (about 1.9 mm for PET and 1.45 mm for SPECT).

We propose and evaluate a new scatter simulation algorithm to correct for scattered events in 3D PET. The algorithm simulates both single and double scatter. We show that single and double scatter distribution can have different shapes. This may cause difficulties if the total scatter is estimated by scaling the single scatter contribution. The whole implementation is based on the STIR library (software for tomographic image reconstruction)