J.M. Udias

FIRST: Fast Iterative Reconstruction Software for (PET) tomography

Small animal PET scanners require high spatial resolution and good sensitivity. To reconstruct high-resolution images in 3D-PET, iterative methods, such as OSEM, are superior to analytical reconstruction algorithms, although their high computational cost is still a serious drawback. The higher performance of modern computers could make iterative image reconstruction fast enough to be viable, provided we are able to deal with the large number of probability coefficients for the system response matrix in high-resolution PET scanners, which is a difficult task that prevents the algorithms from reaching peak computing performance. Considering all possible axial and in-plane symmetries, as well as certain quasi-symmetries, we have been able to reduce the memory requirements to store the system response matrix (SRM) well below 1 GB, which allows us to keep the whole response matrix of the system inside RAM of ordinary industry-standard computers, so that the reconstruction algorithm can achieve near peak performance. The elements of the SRM are stored as cubic spline profiles and matched to voxel size during reconstruction. In this way, the advantages of on-the-fly calculation and of fully stored SRM are combined. The on-the-fly part of the calculation (matching the profile functions to voxel size) of the SRM accounts for 10-30% of the reconstruction time, depending on the number of voxels chosen. We tested our approach with real data from a commercial small animal PET scanner. The results (image quality and reconstruction time) show that the proposed technique is a feasible solution.

PeneloPET, a Monte Carlo PET simulation tool based on PENELOPE: features and validation.

PENELOPE is a Monte Carlo code that simulates the transport in matter of electrons, positrons and photons with energies from a few hundred of eV to 1 GeV. It is robust, fast and very accurate, but it may be unfriendly for people not acquainted with the FORTRAN programming language. We have developed a tookit (`PeneloPET) to prepare simulations of PET and SPECT within PENELOPE. Sophisticated simulations can be setup by modifying just a few simple input files. The output data can be generated at different levels of detail and can be analyzed afterwards with the preferred programming language or tools. In this work, we present the features of PeneloPET as well as validations against other dedicated PET simulation programs and two real scanners.

Gyral and Sulcal Cortical Thinning in Adolescents with First Episode Early-Onset Psychosis

BACKGROUNDnPsychosis is associated with volumetric decreases of cortical structures. Whether these volumetric decreases imply abnormalities in cortical thickness, surface, or cortical folding is not clear. Due to differences in cytoarchitecture, cortical gyri and sulci might be differentially affected by psychosis. Therefore, we examined differences in gyral and sulcal cortical thickness, surface, folding, and volume between a minimally treated male adolescent population with early-onset first-episode psychosis (EOP) and a healthy control group, with surface-based morphometry.nnnMETHODSnMagnetic resonance imaging brain scans were obtained from 49 adolescent EOP patients and 34 healthy control subjects. Subjects were younger than 18 years (age range 12 years-18 years), and EOP patients had a duration of positive symptoms of <6 months.nnnRESULTSnEarly-onset first-episode psychosis was associated with local bilateral cortical thinning and volume deficits in both the gyri and sulci of the superior temporal cortex and the inferior, middle, medial, and superior prefrontal cortex. In the pars triangularis and opercularis cortex of patients, gyral cortical thickness was thinner, whereas sulcal thickness was not. Patients exhibited cortical thinning together with a decreased degree of cortical folding in the right superior frontal cortex.nnnCONCLUSIONSnCortical thinning of both gyri and sulci seem to underlie most cortical volume deficits in adolescent patients with EOP. Except for the right superior frontal region, the degree of cortical folding was normal in regions showing decreased cortical thickness, suggesting that the process of cortical thinning in adolescent patients with EOP primarily takes place after the formation of cortical folds.

Performance evaluation of SiPM detectors for PET imaging in the presence of magnetic fields

The multi-pixel photon counter (MPPC) or silicon photo-multiplier (SiPM), recently introduced as a solid-state photodetector, consists of an array of Geiger-mode photodiodes (microcells). It is a promising device for PET thanks to its potential for high photon detection efficiency (PDE) and immunity to high magnetic fields. It is also very easy to use, with simple electronic read-out, high gain and small size. In this work we evaluate the performance of three 1 × 1 mm2 and one 6 × 6 mm2 (2 × 2 array) SiPMs offered by Hamamatsu for their use in PET. We examine the dependence of the energy resolution and the gain of these devices on the thermal and reverse bias when coupled to LYSO scintillator crystals. We find that the 400 and 1600 microcells models and the 2 × 2 array are suitable for small size crystals, like those employed in high resolution small animal scanners. The good performance of these devices up to 7 Tesla has also been confirmed.

Positron range estimations with PeneloPET

Technical advances towards high resolution PET imaging try to overcome the inherent physical limitations to spatial resolution. Positrons travel in tissue until they annihilate into the two gamma photons detected. This range is the main detector-independent contribution to PET imaging blurring. To a large extent, it can be remedied during image reconstruction if accurate estimates of positron range are available. However, the existing estimates differ, and the comparison with the scarce experimental data available is not conclusive. In this work we present positron annihilation distributions obtained from Monte Carlo simulations with the PeneloPET simulation toolkit, for several common PET isotopes ((18)F, (11)C, (13)N, (15)O, (68)Ga and (82)Rb) in different biological media (cortical bone, soft bone, skin, muscle striated, brain, water, adipose tissue and lung). We compare PeneloPET simulations against experimental data and other simulation results available in the literature. To this end the different positron range representations employed in the literature are related to each other by means of a new parameterization for positron range profiles. Our results are generally consistent with experiments and with most simulations previously reported with differences of less than 20% in the mean and maximum range values. From these results, we conclude that better experimental measurements are needed, especially to disentangle the effect of positronium formation in positron range. Finally, with the aid of PeneloPET, we confirm that scaling approaches can be used to obtain universal, material and isotope independent, positron range profiles, which would considerably simplify range correction.

Positron range effects in high resolution 3D PET imaging

Positron range limits the spatial resolution of PET images. It has a different effect for different isotopes and propagation materials, therefore it is important to consider it during image reconstruction, in order to obtain the best image quality. Positron range distribution was computed using Monte Carlo simulations with PeneloPET. The simulation models positron trajectories and computes the spatial distribution of the annihilation coordinates for the most common isotopes used in PET: 18F, 11C, 13N, 15O, 68Ga and 82Rb. Range profiles are computed for different positron propagation materials, obtaining one kernel profile for each isotope-material combination. These range kernels were introduced in FIRST, a 3D-OSEM image reconstruction software, and employed to blur the object during forward projection. The blurring introduced takes into account the material in which the positron is annihilated, obtained for instance from a CT image. In this way, different positron range corrections for each material in the phantom are considered. We compare resolution and noise properties of the images reconstructed with and without positron range modelling. For this purpose, acquisitions of an Image Quality phantom filled with different isotopes have been simulated for the ARGUS small animal PET scanner.

GPU-Based Fast Iterative Reconstruction of Fully 3-D PET Sinograms

This work presents a graphics processing unit (GPU)-based implementation of a fully 3-D PET iterative reconstruction code, FIRST (Fast Iterative Reconstruction Software for [PET] Tomography), which was developed by our group. We describe the main steps followed to convert the FIRST code (which can run on several CPUs using the message passing interface [MPI] protocol) into a code where the main time-consuming parts of the reconstruction process (forward and backward projection) are massively parallelized on a GPU. Our objective was to obtain significant acceleration of the reconstruction without compromising the image quality or the flexibility of the CPU implementation. Therefore, we implemented a GPU version using an abstraction layer for the GPU, namely, CUDA C. The code reconstructs images from sinogram data, and with the same System Response Matrix obtained from Monte Carlo simulations than the CPU version. The use of memory was optimized to ensure good performance in the GPU. The code was adapted for the VrPET small-animal PET scanner. The CUDA version is more than 70 times faster than the original code running in a single core of a high-end CPU, with no loss of accuracy.

GPU acceleration of a fully 3D Iterative Reconstruction Software for PET using CUDA

A CUDA implementation of the existing software FIRST (Fast Iterative Reconstruction Software for (PET) Tomography) is presented. This implementation uses consumer graphics processing units (GPUs) to accelerate the compute-intensive parts of the reconstruction: forward and backward projection. FIRST was originally developed in FORTRAN, and it has been migrated to C language to be used with NVIDIA C for CUDA, as well as for a straightforward implementation and performance comparison between the C versions of the code running on the CPU and on the GPU. We measured the execution time of the CUDA version compared to the fastest available CPU. The CUDA implementation includes a loop re-ordering and an optimized memory allocation, which improves even more the performance of the reconstruction on the GPUs.

Statistical reconstruction methods in PET: resolution limit, noise, edge artifacts and considerations for the design of better scanners

Small animal positron emission tomography (PET) scanners are being increasingly used as a basic measurement tool in modern biomedical research. The new designs and technologies of these scanners and the modern reconstruction methods have allowed to reach high spatial resolution and sensitivity. Despite their successes, some important issues remain to be addressed in high resolution PET imaging. First, iterative reconstruction methods like maximum likelihood-expectation maximization (MLEM) are known to recover resolution, but also to create noisy images and edge artifacts if some kind of regularization is not imposed. Second, the limit of resolution achievable by iterative methods on high resolution scanners is not quantitatively understood. Third, the use of regularization methods like Sieves or maximum a posteriori (MAP) requires the determination of the optimal values of several adjustable parameter that may be object-dependent. In this work we review these problems in high resolution PET and establish that the origin of them is more related with the physical effects involved in the emission and detection of the radiation during the acquisition than with the kind of iterative reconstruction method chosen. These physical effects (positron range, non-collinearity, scatter inside the object and inside the detector materials) cause that the tube of response (TOR) that connects the voxels with a line of response (LOR) is rather thick. This implies that the higher frequencies of the patient organ structures are not recorded by the scanner and therefore cannot be recovered during the reconstruction. As iterations grow, ML-EM algorithms try to recover higher frequencies in the image. Once that a certain critic frequency is reached, this only maximizes high frequency noise. Using frequency response analyses techniques, we determine the maximum achievable resolution, before edge artifacts spoil the quality of the image, for a particular scanner as a function of the thickness of the TOR, and independently of the reconstruction method employed. With the same techniques, we can deduce well defined stopping criteria for reconstructions methods. Also, criteria for the highest number of subsets which should be used and how the design of the scanners can be optimized when statistical reconstruction methods are employed, is established.

Recovery and normalization of triple coincidences in PET.

PURPOSEnTriple coincidences in positron emission tomography (PET) are events in which three γ-rays are detected simultaneously. These events, though potentially useful for enhancing the sensitivity of PET scanners, are discarded or processed without special consideration in current systems, because there is not a clear criterion for assigning them to a unique line-of-response (LOR). Methods proposed for recovering such events usually rely on the use of highly specialized detection systems, hampering general adoption, and/or are based on Compton-scatter kinematics and, consequently, are limited in accuracy by the energy resolution of standard PET detectors. In this work, the authors propose a simple and general solution for recovering triple coincidences, which does not require specialized detectors or additional energy resolution requirements.nnnMETHODSnTo recover triple coincidences, the authors method distributes such events among their possible LORs using the relative proportions of double coincidences in these LORs. The authors show analytically that this assignment scheme represents the maximum-likelihood solution for the triple-coincidence distribution problem. The PET component of a preclinical PET/CT scanner was adapted to enable the acquisition and processing of triple coincidences. Since the efficiencies for detecting double and triple events were found to be different throughout the scanner field-of-view, a normalization procedure specific for triple coincidences was also developed. The effect of including triple coincidences using their method was compared against the cases of equally weighting the triples among their possible LORs and discarding all the triple events. The authors used as figures of merit for this comparison sensitivity, noise-equivalent count (NEC) rates and image quality calculated as described in the NEMA NU-4 protocol for the assessment of preclinical PET scanners.nnnRESULTSnThe addition of triple-coincidence events with the authors method increased peak NEC rates of the scanner by 26.6% and 32% for mouse- and rat-sized objects, respectively. This increase in NEC-rate performance was also reflected in the image-quality metrics. Images reconstructed using double and triple coincidences recovered using their method had better signal-to-noise ratio than those obtained using only double coincidences, while preserving spatial resolution and contrast. Distribution of triple coincidences using an equal-weighting scheme increased apparent system sensitivity but degraded image quality. The performance boost provided by the inclusion of triple coincidences using their method allowed to reduce the acquisition time of standard imaging procedures by up to ∼25%.nnnCONCLUSIONSnRecovering triple coincidences with the proposed method can effectively increase the sensitivity of current clinical and preclinical PET systems without compromising other parameters like spatial resolution or contrast.