Anthonin Reilhac

Characterisation of partial volume effect and region-based correction in small animal positron emission tomography (PET) of the rat brain

Accurate quantification of PET imaging data is required for a useful interpretation of the measured radioactive tracer concentrations. The partial volume effect (PVE) describes signal dilution and mixing due to spatial resolution and sampling limitations, which introduces bias in quantitative results. In the present study we investigated the magnitude of PVE for volumes of interest (VOIs) in the rat brain and the effect of positron range. In simulated (11)C-raclopride studies we examined the influence of PVE on time activity curves in striatal and cerebellar VOIs and binding potential estimation. The performance of partial volume correction (PVC) was studied using the region-based geometric transfer matrix (GTM) method including the question of whether a spatially variant point spread function (PSF) is necessary for PVC of a rat brain close to the centre of the field of view. Furthermore, we determined the effect of spillover from activity outside the brain. The results confirmed that PVE is significant in rat brain PET and showed that positron range is an important factor that needs to be included in the PSF. There was considerable bias in time activity curves for the simulated (11)C-raclopride studies and significant underestimation of binding potential even for very small centred VOIs. Good activity recovery was achieved with the GTM PVC using a spatially invariant simulated PSF when no activity was present outside the brain. PVC using a simple Gaussian fit point spread function was not sufficiently accurate. Spillover from regions outside the brain had a significant impact on measured activity concentrations and reduced the accuracy of PVC with the GTM method using rat brain regions alone, except for the smallest VOI size but at the cost of increased noise. Voxel-based partial volume correction methods which inherently compensate for spillover from outside the brain might be a more suitable choice.

Analytical positron range modelling in heterogeneous media for PET Monte Carlo simulation

Monte Carlo simulation codes that model positron interactions along their tortuous path are expected to be accurate but are usually slow. A simpler and potentially faster approach is to model positron range from analytical annihilation density distributions. The aims of this paper were to efficiently implement and validate such a method, with the addition of medium heterogeneity representing a further challenge. The analytical positron range model was evaluated by comparing annihilation density distributions with those produced by the Monte Carlo simulator GATE and by quantitatively analysing the final reconstructed images of Monte Carlo simulated data. In addition, the influence of positronium formation on positron range and hence on the performance of Monte Carlo simulation was investigated. The results demonstrate that 1D annihilation density distributions for different isotope-media combinations can be fitted with Gaussian functions and hence be described by simple look-up-tables of fitting coefficients. Together with the method developed for simulating positron range in heterogeneous media, this allows for efficient modelling of positron range in Monte Carlo simulation. The level of agreement of the analytical model with GATE depends somewhat on the simulated scanner and the particular research task, but appears to be suitable for lower energy positron emitters, such as (18)F or (11)C. No reliable conclusion about the influence of positronium formation on positron range and simulation accuracy could be drawn.

NEMA NU 4-2008 validation and applications of the PET-SORTEO Monte Carlo simulations platform for the geometry of the Inveon PET preclinical scanner

Monte Carlo-based simulation of positron emission tomography (PET) data plays a key role in the design and optimization of data correction and processing methods. Our first aim was to adapt and configure the PET-SORTEO Monte Carlo simulation program for the geometry of the widely distributed Inveon PET preclinical scanner manufactured by Siemens Preclinical Solutions. The validation was carried out against actual measurements performed on the Inveon PET scanner at the Australian Nuclear Science and Technology Organisation in Australia and at the Brain & Mind Research Institute and by strictly following the NEMA NU 4-2008 standard. The comparison of simulated and experimental performance measurements included spatial resolution, sensitivity, scatter fraction and count rates, image quality and Derenzo phantom studies. Results showed that PET-SORTEO reliably reproduces the performances of this Inveon preclinical system. In addition, imaging studies showed that the PET-SORTEO simulation program provides raw data for the Inveon scanner that can be fully corrected and reconstructed using the same programs as for the actual data. All correction techniques (attenuation, scatter, randoms, dead-time, and normalization) can be applied on the simulated data leading to fully quantitative reconstructed images. In the second part of the study, we demonstrated its ability to generate fast and realistic biological studies. PET-SORTEO is a workable and reliable tool that can be used, in a classical way, to validate and/or optimize a single PET data processing step such as a reconstruction method. However, we demonstrated that by combining a realistic simulated biological study ([(11)C]Raclopride here) involving different condition groups, simulation allows one also to assess and optimize the data correction, reconstruction and data processing line flow as a whole, specifically for each biological study, which is our ultimate intent.

In vivo PET imaging with [18F]FDG to explain improved glucose uptake in an apolipoprotein A-I treated mouse model of diabetes

Aims/hypothesisType 2 diabetes is characterised by decreased HDL levels, as well as the level of apolipoprotein A-I (apoA-I), the main apolipoprotein of HDLs. Pharmacological elevation of HDL and apoA-I levels is associated with improved glycaemic control in patients with type 2 diabetes. This is partly due to improved glucose uptake in skeletal muscle.MethodsThis study used kinetic modelling to investigate the impact of increasing plasma apoA-I levels on the metabolism of glucose in the db/db mouse model.ResultsTreatment of db/db mice with apoA-I for 2xa0h significantly improved both glucose tolerance (AUC 2574u2009±u200970xa0mmol/lu2009×u2009min vs 2927u2009±u2009137xa0mmol/lu2009×u2009min, for apoA-I and PBS, respectively; pu2009<u20090.05) and insulin sensitivity (AUC 388.8u2009±u200923.8xa0mmol/lu2009×u2009min vs 194.1u2009±u200919.6xa0mmol/lu2009×u2009min, for apoA-I and PBS, respectively; pu2009<u20090.001). ApoA-I treatment also increased glucose uptake by skeletal muscle in both an insulin-dependent and insulin-independent manner as evidenced by increased uptake of fludeoxyglucose ([18F]FDG) from plasma into gastrocnemius muscle in apoA-I treated mice, both in the absence and presence of insulin. Kinetic modelling revealed an enhanced rate of insulin-mediated glucose phosphorylation (k3) in apoA-I treated mice (3.5u2009±u20091.1u2009×u200910−2 min−1 vs 2.3u2009±u20090.7u2009×u200910−2 min−1, for apoA-I and PBS, respectively; pu2009<u20090.05) and an increased influx constant (3.7u2009±u20090.6u2009×u200910−3 ml min−1 g−1 vs 2.0u2009±u20090.3u2009×u200910−3 ml min−1 g−1, for apoA-I and PBS, respectively; pu2009<u20090.05). Treatment of L6 rat skeletal muscle cells with apoA-I for 2xa0h indicated that increased hexokinase activity mediated the increased rate of glucose phosphorylation.Conclusions/interpretationThese findings indicate that apoA-I improves glucose disposal in db/db mice by improving insulin sensitivity and enhancing glucose phosphorylation.

4D PET iterative deconvolution with spatiotemporal regularization for quantitative dynamic PET imaging

Quantitative measurements in dynamic PET imaging are usually limited by the poor counting statistics particularly in short dynamic frames and by the low spatial resolution of the detection system, resulting in partial volume effects (PVEs). In this work, we present a fast and easy to implement method for the restoration of dynamic PET images that have suffered from both PVE and noise degradation. It is based on a weighted least squares iterative deconvolution approach of the dynamic PET image with spatial and temporal regularization. Using simulated dynamic [(11)C] Raclopride PET data with controlled biological variations in the striata between scans, we showed that the restoration method provides images which exhibit less noise and better contrast between emitting structures than the original images. In addition, the method is able to recover the true time activity curve in the striata region with an error below 3% while it was underestimated by more than 20% without correction. As a result, the method improves the accuracy and reduces the variability of the kinetic parameter estimates calculated from the corrected images. More importantly it increases the accuracy (from less than 66% to more than 95%) of measured biological variations as well as their statistical detectivity.

Simulation-based optimisation of the PET data processing for Partial Saturation Approach protocols

Positron emission tomography (PET) with [(11)C]Raclopride is an important tool for studying dopamine D2 receptor expression in vivo. [(11)C]Raclopride PET binding experiments conducted using the Partial Saturation Approach (PSA) allow the estimation of receptor density (B(avail)) and the in vivo affinity appK(D). The PSA is a simple, single injection, single scan experimental protocol that does not require blood sampling, making it ideal for use in longitudinal studies. In this work, we generated a complete Monte Carlo simulated PET study involving two groups of scans, in between which a biological phenomenon was inferred (a 30% decrease of B(avail)), and used it in order to design an optimal data processing chain for the parameter estimation from PSA data. The impact of spatial smoothing, noise removal and image resolution recovery technique on the statistical detection was investigated in depth. We found that image resolution recovery using iterative deconvolution of the image with the system point spread function associated with temporal data denoising greatly improves the accuracy and the statistical reliability of detecting the imposed phenomenon. Before optimisation, the inferred B(avail) variation between the two groups was underestimated by 42% and detected in 66% of cases, while a false decrease of appK(D) by 13% was detected in more than 11% of cases. After optimisation, the calculated B(avail) variation was underestimated by only 3.7% and detected in 89% of cases, while a false slight increase of appK(D) by 3.7% was detected in only 2% of cases. We found during this investigation that it was essential to adjust a factor that accounts for difference in magnitude between the non-displaceable ligand concentrations measured in the target and in the reference regions, for different data processing pathways as this ratio was affected by different image resolutions.

Assessment of a fast generated analytical matrix for rotating slat collimation iterative reconstruction: a possible method to optimize the collimation profile

In SPECT imaging, improvement or deterioration of performance is mostly due to collimator design. Classical SPECT systems mainly use parallel hole or pinhole collimators. Rotating slat collimators (RSC) can be an interesting alternative to optimize the tradeoff between detection efficiency and spatial resolution. The present study was conducted using a RSC system for small animal imaging called CLiR. The CLiR system was used in planar mode only. In a previous study, planar 2D projections were reconstructed using the well-known filtered backprojection algorithm (FBP). In this paper, we investigated the use of the statistical reconstruction algorithm maximum likelihood expectation maximization (MLEM) to reconstruct 2D images with the CLiR system using a probability matrix calculated using an analytic approach. The primary objective was to propose a method to quickly generate a light system matrix, which facilitates its handling and storage, while providing accurate and reliable performance. Two other matrices were calculated using GATE Monte Carlo simulations to investigate the performance obtained using the matrix calculated analytically. The first matrix calculated using GATE took all the physics processes into account, where the second did not consider for the scattering, as the analytical matrix did not take this physics process into account either. 2D images were reconstructed using FBP and MLEM with the three different probability matrices. Both simulated and experimental data were used. A comparative study of these images was conducted using different metrics: the modulation transfert function, the signal-to-noise ratio and quantification measurement. All the results demonstrated the suitability of using a probability matrix calculated analytically. It provided similar results in terms of spatial resolution (about 0.6 mm with differences <5%), signal-to-noise ratio (differences <10%), or quality of image.

Iterative-based Partial Volume Effects correction with wavelet-based regularization for quantitative PET imaging

Two important technical limitations hampering quantitative measurements with dynamic PET imaging are 1) the limited spatial resolution of the detection system which induces quantitative biases due to Partial Volume Effects and 2), the poor counting statistics in the individual time frames which challenges further the subsequent kinetic modeling and analysis. In this work, we present an original method for the restoration of dynamic PET images that had suffered from both the consequences of the limited spatial resolution and the noise processes, which is based on the weighted least square iterative deconvolution in the wavelet space with temporal regularization of the wavelet coefficients. The solution image exhibits less noise, better contrasts between emitting structures and should therefore allow a better quantification compared with no correction‥

Comparative study of partial volume correction methods in small animal positron emission tomography (PET) of the rat brain

Quantitative accuracy in rat brain PET studies is reduced by partial volume effect. We investigated the performance of partial volume correction (PVC) in a realistic situation where activity is also taken up in the head and spills into the brain. The PVC approaches studied include the region-based geometric transfer matrix (GTM) method and voxel-based iterative deconvolution (reblurred Van Cittert and Richardson-Lucy). 8 realizations of dynamic rat brain PET studies of 11C-Raclopride with a binding potential BPND=3 in the striatum were simulated with the Monte Carlo simulator PET SORTEO. Synthetic time activity curves (TACs) were assigned to the striatum, cerebellum, remaining brain and head regions outside the brain of a rat head phantom. Different sized volumes of interest (VOIs) were sampled ranging from the full anatomical region to smaller VOIs containing only voxels with at least 50%, 70% or 90% of the maximum activity. BPND was calculated for the striatum using the simplified reference tissue model with the cerebellum as the reference tissue. Without PVC the accuracy of BPND was very low for all VOI sizes with biases between −44.7% and −20.9%. PVC using the GTM method was only accurate for the smallest 90% VOI with a bias of −7.7% but the standard deviation increased to 4.2% compared to less than 1% for the larger VOIs. Good accuracy was achieved for both iterative deconvolution methods using the 50% VOI (bias less than 8%) with standard deviations of less than 1.8%. Thus, in the presence of activity uptake outside the brain, iterative deconvolution methods outperform the GTM method. We are currently implementing PVC with a spatially variant PSF to better compensate for non-uniformities of spatial resolution away from the centre of the field of view.

Performance of an analytical positron range modelling approach in the context of whole body small animal and clinical PET

Monte Carlo simulation codes that model positron interactions along its tortuous path are expected to be accurate but slow. A simpler and potentially faster approach is to model positron range from analytical annihilation density distributions. We have implemented and validated such a method. 1D annihilation density distributions for different isotope-media combinations were fitted with Gaussian functions and described by simple look-up-tables of fitting coefficients. Together with a method developed for simulating positron range in heterogeneous media this allowed for efficient modeling of positron range. Previously, we focused on evaluating soft tissue-bone transitions and the performance of the modeling in small animal PET where positron range has a higher relative contribution to the spatial resolution than in clinical PET. Here, we extend the assessment to soft tissue-lung transitions in whole body small animal and clinical PET. The performance of the modeling was evaluated by comparing annihilation density distributions obtained in heterogeneous media with those produced by the Monte Carlo simulator GATE and by quantitatively analyzing the final reconstructed images of Monte Carlo simulated data for the small animal PET scanner microPET Focus 220 and the clinical PET scanner ECAT Exact HR+. Modelling heterogeneous media showed some limitations for soft tissue-lung transitions leading to systematic divergence from GATE at large positron range values depending on the isotope and the distance travelled before the transition occurs. The differences in the reconstructed point spread functions in water and lung showed that modeling medium heterogeneity is essential for small animal PET and would be beneficial for clinical PET. The level of agreement between the analytical model and GATE depends somewhat on the simulated scanner, but appears to be suitable for lower energy positron emitters, such as 18F or 11C. However, the method for heterogeneous media modelling could be used with any underlying positron range model.