Practical considerations for image-based PSF and blobs reconstruction in PET

Resolution recovery on list mode MLEM reconstruction for dynamic cardiac SPECT system

The Dynamic Cardiac SPECT (DC-SPECT) system is being developed at the Massachusetts General Hospital, featuring a static cardio focus asymmetrical geometry enabling simultaneous high-resolution and high-sensitivity imaging. Among 14 design iterations of the DC-SPECT with varying number of detector heads, system sensitivity and resolution, the current version under development features 10 mm FWHM geometrical resolution (without resolution recovery) and 0.07% sensitivity at the center of the FOV, this is 1.5× resolution gain and 7× sensitivity gain compared to a conventional dual head gamma camera (0.01% sensitivity and 15-mm resolution). This work presents improvement in imaging resolution by implementing a spatially variant point spread function (SV-PSF) with list mode MLEM reconstruction. A resolution recovery method by PSF deconvolution is validated on list mode MLEM reconstruction for the DC-SPECT. A spatial invariant PSF is included as an additional test to show the influence of the PSF modelling accuracy on reconstructed image quality. We compare the MLEM reconstruction with and without PSF deconvolution; an analytic model is used for the calculation of system response, and the results are compared to the reconstruction with system modelling using Monte Carlo (MC) based methods. Results show that with PSF modelling applied, the quality of the reconstructed image is improved, and the DC-SPECT system can achieve a 4.5 mm central spatial resolution with average 795 counts/Mbq. Both the SV-PSF and the spatial-invariant PSF improve the image quality, and the reconstruction with SV-PSF generates line profiles closer to the ground truth. The results show substantial improvement over the GE Discovery 570c performance (7 mm spatial resolution with an average 460 counts/MBq, 5.8 mm resolution at the FOV center). The impact of PSF deconvolution is significant, improvement of the reconstructed image quality is evident in comparison to MC simulated system matrix with the same sampling size in the simulation.

Micro-Networks for Robust MR-Guided Low Count PET Imaging

Noise suppression is particularly important in low count positron emission tomography (PET) imaging. Post-smoothing (PS) and regularization methods which aim to reduce noise also tend to reduce resolution and introduce bias. Alternatively, anatomical information from another modality such as magnetic resonance (MR) imaging can be used to improve image quality. Convolutional neural networks (CNNs) are particularly well suited to such joint image processing, but usually require large amounts of training data and have mostly been applied outside the field of medical imaging or focus on classification and segmentation, leaving PET image quality improvement relatively understudied. This article proposes the use of a relatively low-complexity CNN (micro-net) as a post-reconstruction MR-guided image processing step to reduce noise and reconstruction artefacts while also improving resolution in low count PET scans. The CNN is designed to be fully 3-D, robust to very limited amounts of training data, and to accept multiple inputs (including competitive denoising methods). Application of the proposed CNN on simulated low (30 M) count data (trained to produce standard (300 M) count reconstructions) results in a 36% lower normalized root mean squared error (NRMSE, calculated over ten realizations against the ground truth) compared to maximum-likelihood expectation maximization (MLEM) used in clinical practice. In contrast, a decrease of only 25% in NRMSE is obtained when an optimized (using knowledge of the ground truth) PS is performed. A 26% NRMSE decrease is obtained with both RM and optimized PS. Similar improvement is also observed for low count real patient datasets. Overfitting to training data is demonstrated to occur as the network size is increased. In an extreme case, a U-net (which produces better predictions for training data) is shown to completely fail on test data due to overfitting to this case of very limited training data. Meanwhile, the resultant images from the proposed CNN (which has low training data requirements) have lower noise, reduced ringing, and partial volume effects, as well as sharper edges and improved resolution compared to conventional MLEM.

Time-of-flight (TOF) implementation for PET reconstruction in practice

The time-of-flight (TOF) feature of PET scanners has been used for a long time in PET reconstruction, but many implementational aspects are still incomplete or ambiguous in the literature. Here we formalize and present theoretical and practical implementation details for the reconstruction of clinical TOF histogram and list-mode data using ML-EM. Relevant aspects include the computation of the TOF component of the system matrix, the processing of TOF bins, the use of estimations of random and scattered coincidences, and differences between histogram and list-mode ML-EM TOF reconstruction. Several approaches and approximations have been implemented in the CASToR platform and compared for OSEM reconstructions of patient data from the GE Signa PET/MR scanner. Differences between implementations are not larger than the typical bias in clinical data reconstruction. The largest difference and contrast loss occur when the processing of histogram TOF bins is simplified, and list-mode reconstruction is most sensitive to the truncation of the Gaussian TOF probability distribution.

Intercomparison of MR-informed PET image reconstruction methods

PURPOSE

Numerous image reconstruction methodologies for positron emission tomography (PET) have been developed that incorporate magnetic resonance (MR) imaging structural information, producing reconstructed images with improved suppression of noise and reduced partial volume effects. However, the influence of MR structural information also increases the possibility of suppression or bias of structures present only in the PET data (PET-unique regions). To address this, further developments for MR-informed methods have been proposed, for example, through inclusion of the current reconstructed PET image, alongside the MR image, in the iterative reconstruction process. In this present work, a number of kernel and maximum a posteriori (MAP) methodologies are compared, with the aim of identifying methods that enable a favorable trade-off between the suppression of noise and the retention of unique features present in the PET data.

METHODS

The reconstruction methods investigated were: the MR-informed conventional and spatially compact kernel methods, referred to as KEM and KEM largest value sparsification (LVS) respectively; the MR-informed Bowsher and Gaussian MR-guided MAP methods; and the PET-MR-informed hybrid kernel and anato-functional MAP methods. The trade-off between improving the reconstruction of the whole brain region and the PET-unique regions was investigated for all methods in comparison with postsmoothed maximum likelihood expectation maximization (MLEM), evaluated in terms of structural similarity index (SSIM), normalized root mean square error (NRMSE), bias, and standard deviation. Both simulated BrainWeb (10 noise realizations) and real [18 F] fluorodeoxyglucose (FDG) three-dimensional datasets were used. The real [18 F]FDG dataset was augmented with simulated tumors to allow comparison of the reconstruction methodologies for the case of known regions of PET-MR discrepancy and evaluated at full counts (100%) and at a reduced (10%) count level.

RESULTS

For the high-count simulated and real data studies, the anato-functional MAP method performed better than the other methods under investigation (MR-informed, PET-MR-informed and postsmoothed MLEM), in terms of achieving the best trade-off for the reconstruction of the whole brain and PET-unique regions, assessed in terms of the SSIM, NRMSE, and bias vs standard deviation. The inclusion of PET information in the anato-functional MAP method enables the reconstruction of PET-unique regions to attain similarly low levels of bias as unsmoothed MLEM, while moderately improving the whole brain image quality for low levels of regularization. However, for low count simulated datasets the anato-functional MAP method performs poorly, due to the inclusion of noisy PET information in the regularization term. For the low counts simulated dataset, KEM LVS and to a lesser extent, HKEM performed better than the other methods under investigation in terms of achieving the best trade-off for the reconstruction of the whole brain and PET-unique regions, assessed in terms of the SSIM, NRMSE, and bias vs standard deviation.

CONCLUSION

For the reconstruction of noisy data, multiple MR-informed methods produce favorable whole brain vs PET-unique region trade-off in terms of the image quality metrics of SSIM and NRMSE, comfortably outperforming the whole image denoising of postsmoothed MLEM.

Factor Analysis of Dynamic PET Images: Beyond Gaussian Noise

Factor analysis has proven to be a relevant tool for extracting tissue time-activity curves (TACs) in dynamic PET images, since it allows for an unsupervised analysis of the data. Reliable and interpretable results are possible only if it is considered with respect to suitable noise statistics. However, the noise in reconstructed dynamic PET images is very difficult to characterize, despite the Poissonian nature of the count rates. Rather than explicitly modeling the noise distribution, this paper proposes to study the relevance of several divergence measures to be used within a factor analysis framework. To this end, the β -divergence, widely used in other applicative domains, is considered to design the data-fitting term involved in three different factor models. The performances of the resulting algorithms are evaluated for different values of β , in a range covering Gaussian, Poissonian, and Gamma-distributed noises. The results obtained on two different types of synthetic images and one real image show the interest of applying non-standard values of β to improve the factor analysis.

Spatially-Compact MR-Guided Kernel EM for PET Image Reconstruction

Positron emission tomography (PET) is a highly sensitive functional and molecular imaging modality which can measure picomolar concentrations of an injected radionuclide. However, the physical sensitivity of PET is limited, and reducing the injected dose leads to low count data and noisy reconstructed images. A highly effective way of reducing noise is to reparameterise the reconstruction in terms of MR-derived spatial basis functions. Spatial basis functions derived using the kernel method have demonstrated excellent noise reduction properties and maintain shared PET-MR detailed structures. However, as previously shown in the literature, the MR-guided kernel method may lead to excessive smoothing of structures that are only present in the PET data. This work makes two main contributions in order to address this problem: first, we exploit the potential of the MR-guided kernel method to form more spatially-compact basis functions which are able to preserve PET-unique structures, and secondly, we consider reconstruction at the native MR resolution. The former contribution notably improves the recovery of structures which are unique to the PET data. These adaptations of the kernel method were compared to the conventional implementation of the MR-guided kernel method and also to MLEM, in terms of ability to recover PET unique structures for both simulated and real data. The spatially-compact kernel method showed clear visual and quantitative improvements in the reconstruction of the PET unique structures, relative to the conventional kernel method for all sizes of PET unique structures investigated, whilst maintaining to a large extent the impressive noise mitigating and detail preserving properties of the conventional MR-guided kernel method. We therefore conclude that a spatially-compact parameterisation of the MR-guided kernel method, should be the preferred implementation strategy in order to obviate unnecessary losses in PET-unique details.

Molecular Imaging-Guided Radiotherapy for the Treatment of Head-and-Neck Squamous Cell Carcinoma: Does it Fulfill the Promises?

With the routine use of intensity modulated radiation therapy for the treatment of head-and-neck squamous cell carcinoma allowing highly conformed dose distribution, there is an increasing need for refining both the selection and the delineation of gross tumor volumes (GTV). In this framework, molecular imaging with positron emission tomography and magnetic resonance imaging offers the opportunity to improve diagnostic accuracy and to integrate tumor biology mainly related to the assessment of tumor cell density, tumor hypoxia, and tumor proliferation into the treatment planning equation. Such integration, however, requires a deep comprehension of the technical and methodological issues related to image acquisition, reconstruction, and segmentation. Until now, molecular imaging has had a limited value for the selection of nodal GTV, but there are increasing evidences that both FDG positron emission tomography and diffusion-weighted magnetic resonance imaging has a potential value for the delineation of the primary tumor GTV, effecting on dose distribution. With the apprehension of the heterogeneity in tumor biology through molecular imaging, growing evidences have been collected over the years to support the concept of dose escalation/dose redistribution using a planned heterogeneous dose prescription, the so-called "dose painting" approach. Validation trials are ongoing, and in the coming years, one may expect to position the dose painting approach in the armamentarium for the treatment of patients with head-and-neck squamous cell carcinoma.

Multimodal partial volume correction: Application to [11C]PIB PET/MRI myelin imaging in multiple sclerosis

The [11C]PIB PET tracer, originally developed for amyloid imaging, has been recently repurposed to quantify demyelination and remyelination in multiple sclerosis (MS). Myelin PET imaging, however, is limited by its low resolution that deteriorates the quantification accuracy of white matter (WM) lesions. Here, we introduce a novel partial volume correction (PVC) method called Multiresolution-Multimodal Resolution-Recovery (MM-RR), which uses the wavelet transform and a synergistic statistical model to exploit MRI structural images to improve the resolution of [11C]PIB PET myelin imaging. MM-RR performance was tested on a phantom acquisition and in a dataset comprising [11C]PIB PET and MR T1- and T2-weighted images of 8 healthy controls and 20 MS patients. For the control group, the MM-RR PET images showed an average increase of 5.7% in WM uptake while the grey-matter (GM) uptake remained constant, resulting in +31% WM/GM contrast. Furthermore, MM-RR PET binding maps correlated significantly with the mRNA expressions of the most represented proteins in the myelin sheath (R2 = 0.57 ± 0.09). In the patient group, MM-RR PET images showed sharper lesion contours and significant improvement in normal-appearing tissue/WM-lesion contrast compared to standard PET (contrast improvement > +40%). These results were consistent with MM-RR performances in phantom experiments.

Image-based Modeling of PSF Deformation with Application to Limited Angle PET Data

The point-spread-functions (PSFs) of reconstructed images can be deformed due to detector effects such as resolution blurring and parallax error, data acquisition geometry such as insufficient sampling or limited angular coverage in dual-panel PET systems, or reconstruction imperfections/simplifications. PSF deformation decreases quantitative accuracy and its spatial variation lowers consistency of lesion uptake measurement across the imaging field-of-view (FOV). This can be a significant problem with dual panel PET systems even when using TOF data and image reconstruction models of the detector and data acquisition process. To correct for the spatially variant reconstructed PSF distortions we propose to use an image-based resolution model (IRM) that includes such image PSF deformation effects. Originally the IRM was mostly used for approximating data resolution effects of standard PET systems with full angular coverage in a computationally efficient way, but recently it was also used to mitigate effects of simplified geometric projectors. Our work goes beyond this by including into the IRM reconstruction imperfections caused by combination of the limited angle, parallax errors, and any other (residual) deformation effects and testing it for challenging dual panel data with strongly asymmetric and variable PSF deformations. We applied and tested these concepts using simulated data based on our design for a dedicated breast imaging geometry (B-PET) consisting of dual-panel, time-of-flight (TOF) detectors. We compared two image-based resolution models; i) a simple spatially invariant approximation to PSF deformation, which captures only the general PSF shape through an elongated 3D Gaussian function, and ii) a spatially variant model using a Gaussian mixture model (GMM) to more accurately capture the asymmetric PSF shape in images reconstructed from data acquired with the B-PET scanner geometry. Results demonstrate that while both IRMs decrease the overall uptake bias in the reconstructed image, the second one with the spatially variant and accurate PSF shape model is also able to ameliorate the spatially variant deformation effects to provide consistent uptake results independent of the lesion location within the FOV.

Optimization of brain PET imaging for a multicentre trial: the French CATI experience

Background

CATI is a French initiative launched in 2010 to handle the neuroimaging of a large cohort of subjects recruited for an Alzheimer’s research program called MEMENTO. This paper presents our test protocol and results obtained for the 22 PET centres (overall 13 different scanners) involved in the MEMENTO cohort. We determined acquisition parameters using phantom experiments prior to patient studies, with the aim of optimizing PET quantitative values to the highest possible per site, while reducing, if possible, variability across centres.

Methods

Jaszczak’s and 3D-Hoffman’s phantom measurements were used to assess image spatial resolution (ISR), recovery coefficients (RC) in hot and cold spheres, and signal-to-noise ratio (SNR). For each centre, the optimal reconstruction parameters were chosen as those maximizing ISR and RC without a noticeable decrease in SNR. Point-spread-function (PSF) modelling reconstructions were discarded. The three figures of merit extracted from the images reconstructed with optimized parameters and routine schemes were compared, as were volumes of interest ratios extracted from Hoffman acquisitions. The net effect of the 3D-OSEM reconstruction parameter optimization was investigated on a subset of 18 scanners without PSF modelling reconstruction.

Results

Compared to the routine parameters of the 22 PET centres, average RC in the two smallest hot and cold spheres and average ISR remained stable or were improved with the optimized reconstruction, at the expense of slight SNR degradation, while the dispersion of values was reduced.

For the subset of scanners without PSF modelling, the mean RC of the smallest hot sphere obtained with the optimized reconstruction was significantly higher than with routine reconstruction. The putamen and caudate-to-white matter ratios measured on 3D-Hoffman acquisitions of all centres were also significantly improved by the optimization, while the variance was reduced.

Conclusions

This study provides guidelines for optimizing quantitative results for multicentric PET neuroimaging trials.

Improved image quality in pinhole SPECT by accurate modeling of the point spread function in low magnification systems

PURPOSE

Single photon emission computed tomography (SPECT) has become an important noninvasive imaging technique in small-animal research. Due to the high resolution required in small-animal SPECT systems, the spatially variant system response needs to be included in the reconstruction algorithm. Accurate modeling of the system response should result in a major improvement in the quality of reconstructed images. The aim of this study was to quantitatively assess the impact that an accurate modeling of spatially variant collimator/detector response has on image-quality parameters, using a low magnification SPECT system equipped with a pinhole collimator and a small gamma camera.

METHODS

Three methods were used to model the point spread function (PSF). For the first, only the geometrical pinhole aperture was included in the PSF. For the second, the septal penetration through the pinhole collimator was added. In the third method, the measured intrinsic detector response was incorporated. Tomographic spatial resolution was evaluated and contrast, recovery coefficients, contrast-to-noise ratio, and noise were quantified using a custom-built NEMA NU 4-2008 image-quality phantom.

RESULTS

A high correlation was found between the experimental data corresponding to intrinsic detector response and the fitted values obtained by means of an asymmetric Gaussian distribution. For all PSF models, resolution improved as the distance from the point source to the center of the field of view increased and when the acquisition radius diminished. An improvement of resolution was observed after a minimum of five iterations when the PSF modeling included more corrections. Contrast, recovery coefficients, and contrast-to-noise ratio were better for the same level of noise in the image when more accurate models were included. Ring-type artifacts were observed when the number of iterations exceeded 12.

CONCLUSIONS

Accurate modeling of the PSF improves resolution, contrast, and recovery coefficients in the reconstructed images. To avoid the appearance of ring-type artifacts, the number of iterations should be limited. In low magnification systems, the intrinsic detector PSF plays a major role in improvement of the image-quality parameters.

Experimental evaluation and basis function optimization of the spatially variant image-space PSF on the Ingenuity PET/MR scanner

PURPOSE

The Ingenuity time-of-flight (TF) PET/MR is a recently developed hybrid scanner combining the molecular imaging capabilities of PET with the excellent soft tissue contrast of MRI. It is becoming common practice to characterize the system's point spread function (PSF) and understand its variation under spatial transformations to guide clinical studies and potentially use it within resolution recovery image reconstruction algorithms. Furthermore, due to the system's utilization of overlapping and spherical symmetric Kaiser-Bessel basis functions during image reconstruction, its image space PSF and reconstructed spatial resolution could be affected by the selection of the basis function parameters. Hence, a detailed investigation into the multidimensional basis function parameter space is needed to evaluate the impact of these parameters on spatial resolution.

METHODS

Using an array of 12 × 7 printed point sources, along with a custom made phantom, and with the MR magnet on, the system's spatially variant image-based PSF was characterized in detail. Moreover, basis function parameters were systematically varied during reconstruction (list-mode TF OSEM) to evaluate their impact on the reconstructed resolution and the image space PSF. Following the spatial resolution optimization, phantom, and clinical studies were subsequently reconstructed using representative basis function parameters.

RESULTS

Based on the analysis and under standard basis function parameters, the axial and tangential components of the PSF were found to be almost invariant under spatial transformations (~4 mm) while the radial component varied modestly from 4 to 6.7 mm. Using a systematic investigation into the basis function parameter space, the spatial resolution was found to degrade for basis functions with a large radius and small shape parameter. However, it was found that optimizing the spatial resolution in the reconstructed PET images, while having a good basis function superposition and keeping the image representation error to a minimum, is feasible, with the parameter combination range depending upon the scanner's intrinsic resolution characteristics.

CONCLUSIONS

Using the printed point source array as a MR compatible methodology for experimentally measuring the scanner's PSF, the system's spatially variant resolution properties were successfully evaluated in image space. Overall the PET subsystem exhibits excellent resolution characteristics mainly due to the fact that the raw data are not under-sampled/rebinned, enabling the spatial resolution to be dictated by the scanner's intrinsic resolution and the image reconstruction parameters. Due to the impact of these parameters on the resolution properties of the reconstructed images, the image space PSF varies both under spatial transformations and due to basis function parameter selection. Nonetheless, for a range of basis function parameters, the image space PSF remains unaffected, with the range depending on the scanner's intrinsic resolution properties.

Efficient fully 3D list-mode TOF PET image reconstruction using a factorized system matrix with an image domain resolution model

A factorized system matrix utilizing an image domain resolution model is attractive in fully 3D time-of-flight PET image reconstruction using list-mode data. In this paper, we study a factored model based on sparse matrix factorization that is comprised primarily of a simplified geometrical projection matrix and an image blurring matrix. Beside the commonly-used Siddon's ray-tracer, we propose another more simplified geometrical projector based on the Bresenham's ray-tracer which further reduces the computational cost. We discuss in general how to obtain an image blurring matrix associated with a geometrical projector, and provide theoretical analysis that can be used to inspect the efficiency in model factorization. In simulation studies, we investigate the performance of the proposed sparse factorization model in terms of spatial resolution, noise properties and computational cost. The quantitative results reveal that the factorization model can be as efficient as a non-factored model, while its computational cost can be much lower. In addition we conduct Monte Carlo simulations to identify the conditions under which the image resolution model can become more efficient in terms of image contrast recovery. We verify our observations using the provided theoretical analysis. The result offers a general guide to achieve the optimal reconstruction performance based on a sparse factorization model with an image domain resolution model.