Single-photon emission computed tomography (SPECT) is a well-established nuclear imaging modality for studying functional and pathological properties of the brain. Conventional general purpose SPECT systems typically offer a spatial resolution of about 10 mm with a sensitivity of 0.01-0.02%. A few dedicated brain SPECT scanners have been proposed, but resolutions and sensitivities are no better than 7 mm and 0.03% respectively, and some of these scanners are not manufactured anymore. This limited resolution hampers detection of localized brain abnormalities, while the low sensitivity requires a long scanning time that limits fast dynamic studies. Besides a compromised resolution and sensitivity, conventional SPECT systems require rotation of heavy detectors to obtain sufficient angular projections, which hamper fast dynamic imaging. @en

In clinical brain SPECT, correction for photon attenuation in the patient is essential to obtain images which provide quantitative information on the regional activity concentration per unit volume (kBq). This correction generally requires an attenuation map (map) denoting the attenuation coefficient at each voxel which is often derived from a CT or MRI scan. However, such an additional scan is not always available and the method may suffer from registration errors. Therefore, we propose a SPECT-only-based strategy for map estimation that we apply to a stationary multi-pinhole clinical SPECT system (G-SPECT-I) for 99mTc-HMPAO brain perfusion imaging. The method is based on the use of a convolutional neural network (CNN) and was validated with Monte Carlo simulated scans. Data acquired in list mode was used to employ the energy information of both primary and scattered photons to obtain information about the tissue attenuation as much as possible. Multiple SPECT reconstructions were performed from different energy windows over a large energy range. Locally extracted 4D SPECT patches (three spatial plus one energy dimension) were used as input for the CNN which was trained to predict the attenuation coefficient of the corresponding central voxel of the patch. Results show that Attenuation Correction using the Ground Truth maps (GT-AC) or using the CNN estimated maps (CNN-AC) achieve comparable accuracy. This was confirmed by a visual assessment as well as a quantitative comparison; the mean deviation from the GT-AC when using the CNN-AC is within 1.8% for the standardized uptake values in all brain regions. Therefore, our results indicate that a CNN-based method can be an automatic and accurate tool for SPECT attenuation correction that is independent of attenuation data from other imaging modalities or human interpretations about head contours.

@en

SPECT imaging with 123I-FP-CIT is used for diagnosis of neurodegenerative disorders like Parkinson's disease. Attenuation correction (AC) can be useful for quantitative analysis of 123I-FP-CIT SPECT. Ideally, AC would be performed based on attenuation maps (μ-maps) derived from perfectly registered CT scans. Such μ-maps, however, are most times not available and possible errors in image registration can induce quantitative inaccuracies in AC corrected SPECT images. Earlier, we showed that a convolutional neural network (CNN) based approach allows to estimate SPECT-aligned μ-maps for full brain perfusion imaging using only emission data. Here we investigate the feasibility of similar CNN methods for axially focused 123I-FP-CIT scans. We tested our approach on a high-resolution multi-pinhole prototype clinical SPECT system in a Monte Carlo simulation study. Three CNNs that estimate μ-maps in a voxel-wise, patch-wise and image-wise manner were investigated. As the added value of AC on clinical 123I-FP-CIT scans is still debatable, the impact of AC was also reported to check in which cases CNN based AC could be beneficial. AC using the ground truth μ-maps (GT-AC) and CNN estimated μ-maps (CNN-AC) were compared with the case when no AC was done (No-AC). Results show that the effect of using GT-AC versus CNN-AC or No-AC on striatal shape and symmetry is minimal. Specific binding ratios (SBRs) from localized regions show a deviation from GT-AC ≤ 2.5% for all three CNN-ACs while No-AC systematically underestimates SBRs by 13.1%. A strong correlation (r ≥ 0.99) was obtained between GT-AC based SBRs and SBRs from CNN-ACs and No-AC. Absolute quantification (in kBq ml-1) shows a deviation from GT-AC within 2.2% for all three CNN-ACs and of 71.7% for No-AC. To conclude, all three CNNs show comparable performance in accurate μ-map estimation and 123I-FP-CIT quantification. CNN-estimated μ-map can be a promising substitute for CT-based μ-map.

@en

Brain perfusion SPECT can be used in the diagnosis of various neurologic or psychiatric disorders, e.g. stroke, epilepsy, dementia and posttraumatic stress disorder. As traditional SPECT provides limited resolution and sensitivity, we recently proposed a high resolution focusing multi-pinhole clinical SPECT scanner dubbed G-SPECT-I (Beekman et al 2015, Eur. J. Nucl. Med. Mol. Imaging 42 S209). G-SPECT-I achieves data completeness in the scan region of interest (ROI) by making small translations of the patient bed while using projections from all bed positions together for image reconstruction. A strategy to restrict the number of bed translations is desired to minimize overhead time. Previously we presented optimized bed translation paths for focused partial brain imaging, while here we focus on whole brain imaging which is the common procedure in perfusion studies. Thus, a series of noise-free scans using a reduced number of bed positions were simulated and compared to an oversampled reference scan acquired with 128 bed positions. Noisy simulations were included to validate the utility of the optimized sequences in more realistic situations. Brain uptake ratios (BURs) and left-right Asymmetry Indices (AIs) in 51 selected regions of interest (ROIs) were calculated for assessment. Results show that images were barely affected by decreasing the number of bed positions from 128 down to 18 (mean deviation from the reference of only 2.2% and 1.5% for the BUR and AI, respectively) while slightly larger deviations (2.9% and 2.7%, respectively) were obtained when using 12 positions. For both 18- and 12-position sequences these deviations due to sampling were much smaller than those induced by noise (mean deviation of 6.5% and 8.6%, respectively). Given an associated total overhead for bed movement of half a minute (18 positions) or 20 s (12 positions), G-SPECT-I can be a clinical platform that brings new protocols for fast (dynamic) whole brain SPECT and motion correction into reach.

@en

The current situation with green gas emission requires the development of low-carbon energy solutions. However, a significant part of the modern energy industry still relies on fossil fuels. To combine these two contradictory targets, we investigate a strategy based on a combination of CO₂ sequestration with enhanced oil recovery (EOR) in the hydrocarbon reservoirs. In such technology, the development of miscibility is the most attractive strategy from both technological and economic aspects. Modeling of this process involves solving complex nonlinear problem describing compositional flow and transport in highly heterogeneous porous media. An accurate capture of the miscibility development usually requires an extensive number of components to be present in the compositional problem which makes simulation run-time prohibitive for optimization. Here, we apply a multi-scale reconstructing of compositional transport to the optimization of CO₂ injection. In this approach, a prolongation operator, based on the parametrization of injection and production tie-lines, is constructed following the fractional flow theory. This operator is tabulated as a function of pressure and pseudo-composition which then is used in the operator-based linearization (OBL) framework for simulation. As a result, a pseudo two-component solution of the multidimensional problem will match the position of trailing and leading shocks of the original problem which helps to accurately predict phase distribution. The reconstructed multicomponent solution can be used then as an effective proxy-model mimicking the behavior of the original multicomponent system. Next, we use this proxy-model in the optimization procedure which helps to improve the performance of the process several fold. An additional benefit of the proposed methodology is based on the fact that important technological features of CO₂ injection process can be captured with lower degrees of freedom which makes the optimization solution more feasible.

@en

The current situation with green gas emission requires the development of low carbon energy solutions. However, a significant part of the modern energy industry still relies on fossil fuels. To combine these two contradictory targets, we investigate a strategy based on a combination of CO2 sequestration with Enhanced Oil Recovery (EOR) in the hydrocarbon reservoirs. In such technology, the development of miscibility is the most attractive strategy from both technological and economic aspects. Modeling of this process involves solving complex nonlinear problem describing compositional flow and transport in highly heterogeneous porous media. An accurate capture of the miscibility development usually requires an extensive number of components to be present in the compositional problem which makes simulation run-time prohibitive for optimization. Here, we apply a multi-scale reconstructing of compositional transport to the optimization of CO2 injection. In this approach, a prolongation operator, based on the parametrization of injection and production tie-lines, is constructed following the fractional flow theory. This operator is tabulated as a function of pressure and pseudocomposition which then is used in the Operator-Based Linearization (OBL) framework for simulation. As a result, a pseudo two-component solution of the multidimensional problem will match the position of trailing and leading shocks of the original problem which helps to accurately predict phase distribution. The reconstructed multicomponent solution can be used then as an effective proxy-model mimicking the behavior of the original multicomponent system. Next, we use this proxy-model in the optimization procedure which helps to improve the performance of the process in several folds. An additional benefit of the proposed methodology is based on the fact that important technological features of CO2 injection process can be captured with lower degrees of freedom which makes the optimization solution more feasible.

@en

A compositional formulation is a reliable option for understanding the complex subsurface processes and the associated physical changes. However, this type of model has a great computational cost, since the number of equations that needs to be solved in each grid block increases proportionally with the number of components employed, thereby making them computationally demanding. In an effort to enhance the solution strategy of the hyperbolic problem, we herewith propose a multiscale reconstruction of compositional transport problem. Until recently, multiscale techniques have been seldom implemented on transport equations. Here, the ideology consists of two stages, wherein two different sets of restriction and prolongation operators are defined based on the dynamics of compositional transport. In the first stage, an operator restricting the arbitrary number of components to single transport equation is implemented with the objective of reconstructing the leading and trailing shock positions in space. The prediction of front propagation is the most critical aspect of the approach, as they involve a lot of uncertainty. Once their positions are identified, the full solution lying in the regions outside the shocks can be conservatively reconstructed based on the prolongation interpolation operator. Subsequently, the solution for the multicomponent problem (full system) in the two-phase region is reconstructed by solving just two transport equations with the aid of restriction operator defined based on an invariant thermodynamic path (based on Compositional Space Parameterization technique). We demonstrate applicability of the approach for the idealistic 1D test cases involving various gas drives with different number of components. Further, the first stage reconstruction was tested successfully on more realistic problems based on implementation in recently developed Operator-Based Linearization (OBL) platform.@en

SPECT can be used to image dopamine transporter (DaT) availability in the human striatum, e.g. for diagnosis of Parkinson's disease (PD). As traditional SPECT provides limited resolution and sensitivity, we proposed a full ring focusing multi-pinhole SPECT system (G-SPECT-I (Beekman 2015 Eur. J. Nucl. Med. Mol. Imaging 42 S209)) which demonstrated a 2.5 mm reconstructed resolution in phantom scans. G-SPECT-I achieves data completeness in the scan region of interest by translating the patient bed with an xyz-stage and combining projections from all bed positions into image reconstruction using a scanning focus method (SFM). This paper aims to develop dedicated SFM parameters for performing a DaTscan with high effective sensitivity and appropriate sampling. To this end, the axial scanning length was restricted and transaxial bed trajectories with a reduced number of positions based on a convex hull data-completeness model were tested. Quantitative accuracy was assessed using full G-SPECT-I simulations of an Alderson phantom based on measured system matrices. For each sampling strategy, the specific binding ratio (SBR) and asymmetry index (AI) in the left and right striatum, as well as the Localized SBR (L-SBR) and the Localized AI (L-AI) in eight striatal sub-regions were calculated and compared to those of the reference scan which performs full brain oversampling using 112 bed positions. Results show that structures essential for PD diagnosis were visually and quantitatively barely affected even when using the lowest number of bed translations (i.e. 4). The maximum deviation from the reference was only 1.5%, 1.5%, 5.5% and 7.0% for the SBR, AI, L-SBR and L-AI, respectively, when 4 positions were used. Thus, it is possible to perform an accurate DaTscan with a confined axial scan region and a limited number of focused bed positions. This enables protocols for extremely fast dynamic SPECT scans with less than half-minute time frames, which can be useful for motion correction.

@en

This paper presents a supervised retinal vessel segmentation by incorporating vessel filtering and wavelet transform features from orientation scores (OSs), and green intensity. Through an anisotropic wavelet-type transform, a 2D image is lifted to a 3D orientation score in the Lie-group domain of positions and orientations R²⋊S¹. Elongated structures are disentangled into their corresponding orientation planes and enhanced via multi-orientation vessel filtering. In addition, scale-selective OSs (in the domain of positions, orientations and scales R²⋊S¹×R⁺) are obtained by adding a scale adaptation to the wavelet transform. Features are optimally extracted by taking maximum orientation responses at multiple scales, to represent vessels of changing calibers. Finally, we train a Random Forest classifier for vessel segmentation. Extensive validations show that our method achieves a competitive segmentation, and better vessel preservation with less false detections compared with the state-of-the-art methods.

@en