Positron Emission Tomography Reconstruction Research Articles

Effect of attenuation mismatches in time of flight PET reconstruction

While the pursuit of better time resolution in positron emission tomography (PET) is rapidly evolving, little work has been performed on time of flight (TOF) image quality at high time resolution in the presence of modelling inconsistencies. This works focuses on the effect of using the wrong attenuation map in the system model, causing perturbations in the reconstructed radioactivity image. Previous work has usually considered the effects to be local to the area where there is attenuation mismatch, and has shown that the quantification errors in this area tend to reduce with improved time resolution. This publication shows however that errors in the PET image at a distance from the mismatch increase with time resolution. The errors depend on the reconstruction algorithm used. We quantify the errors in the hypothetical case of perfect time resolution for maximum likelihood reconstructions. In addition, we perform reconstructions on simulated and patient data. In particular, for respiratory-gated reconstructions from a wrong attenuation map, increased errors are observed with improved time resolutions in areas close to the lungs (e.g. from 13.3% in non-TOF to up to 20.9% at 200 ps in the left ventricle).

Analytic time-of-flight positron emission tomography reconstruction: three-dimensional case

In a positron emission tomography (PET) scanner, the time-of-flight (TOF) information gives us rough event position along the line-of-response (LOR). Using the TOF information for PET image reconstruction is able to reduce image noise. The state-of-the-art TOF PET image reconstruction uses iterative algorithms. This study introduces an analytic TOF PET algorithm that focuses on three-dimensional (3D) reconstruction. The proposed algorithm is in the form of backprojection filtering, in which the backprojection is performed first by using a time-resolution profile function, and then a 3D filter is applied to the backprojected image. For the list-mode data, the backprojection is carried out in the event-by-event fashion, and the timing resolution determined weighting function is used along the projection LOR. Computer simulations are carried out to verify the feasibility of the proposed algorithm.

PET image reconstruction using physical and mathematical modelling for time of flight PET-MR scanners in the STIR library.

This work demonstrates how computational and physical modelling of the positron emission tomography (PET) image acquisition process for a state-of-the-art integrated PET and magnetic resonance imaging (PET-MR) system can produce images comparable to the manufacturer. The GE SIGNA PET/MR scanner is manufactured by General Electric and has time-of-flight (TOF) capabilities of about 390ps. All software development took place in the Software for Tomographic Image Reconstruction (STIR: http://stir.sf.net) library, which is a widely used open source software to reconstruct data as exported from emission tomography scanners. The new software developments will be integrated into STIR, providing the opportunity for researchers worldwide to establish and expand their image reconstruction methods. Furthermore, this work is of particular significance as it provides the first validation of TOF PET image reconstruction for real scanner datasets using the STIR library. This paper presents the methodology, analysis, and critical issues encountered in implementing an independent reconstruction software package. Acquired PET data were processed via several appropriate algorithms which are necessary to produce an accurate and precise quantitative image. This included mathematical, physical and anatomical modelling of the patient and simulation of various aspects of the acquisition. These included modelling of random coincidences using 'singles' rates per crystals, detector efficiencies and geometric effects. Attenuation effects were calculated by using the STIR's attenuation correction model. Modelling all these effects within the system matrix allowed the reconstruction of PET images which demonstrates the metabolic uptake of the administered radiopharmaceutical. These implementations were validated using measured phantom and clinical datasets. The developments are tested using the ordered subset expectation maximisation (OSEM) and the more recently proposed kernelised expectation maximisation (KEM) algorithm which incorporates anatomical information from MR images into PET reconstruction.

Embedded time of arrival estimation for digital silicon photomultipliers with in-pixel TDCs

Including precise time-of-flight measurements in the reconstruction of positron emission tomography (PET) images reduces the required radiotracer dose, scan time and improves the contrast-to-noise ratio. However, major improvements of photodetectors and scintillators are still required to fully exploit the projected 10 ps FWHM coincidence timing resolution (CTR). On the photodetector side, single photon avalanche diode (SPAD) detectors based on digital architectures promise better CTR by timestamping multiple scintillation photons from a single PET event with an array of in-pixel time-to-digital converters (TDCs). The digital nature of these architectures allow the introduction of data processing schemes to optimally combine the produced timestamps for a scintillation event with a time of arrival estimator. This processing is typically performed on an external computing platform, which requires the transfer of large quantities of raw data and leads to a higher dead time owing to the large number of TDCs. To circumvent this issue, we propose a photodetector architecture with in-pixel TDCs that embeds the data processing scheme, reducing the dataset produced by the TDC array to a single timestamp per scintillation event. Hardware-in-the-loop simulations of the fabricated CMOS 65 nm post-processing circuit coupled with simulated SPADs and 1 × 1 × 10 mm3 LYSO scintillators demonstrate a CTR improvement from 160 ps to 126 ps, a bandwidth reduction from 34.7 Mbit/s to 0.5 Mbit/s and a serial transmission dead time reduction from 69 μs to 1μs.

Improved PET quantification and harmonization by adaptive denoising

Purpose: Quantification in positron emission tomography (PET) is subject to bias due to physical and technical limitations. The goal of quantitative harmonization is to achieve comparable measurements between different scanners, thus enabling multicenter clinical trials. Clinical guidelines, such as those from the European Association of Nuclear Medicine (EANM), recommend harmonizing PET reconstructions to bring contrast recovery coefficients (CRCs) within specifications. However, these harmonized reconstructions can show quantitative biases. In this work we improve harmonization by using a novel adaptive filtering scheme. Our goal was to obtain low quantification bias and high peak signal to noise ratio (PSNR) values at the same time. Methods: a novel three-stage adaptive denoising filter was implemented. Filter parameters were optimized to achieve both high PSNR in a digital brain phantom and low quantitative bias of maximum CRC values (CRCmax) obtained from a National Electrical Manufacturers Association (NEMA) PET image quality phantom. The NEMA phantom was scanned on several PET/CT scanners and reconstructed without postfilters. The optimal filter settings found for a training dataset were then applied to testing reconstructions from other scanners. Harmonization limits were defined using the 95% confidence intervals across reconstructions. Results: Average CRCmax values close to unity (± 5%) were achieved for spheres with diameter equal or greater than 13 mm for the training dataset. PSNR values were comparable to other state-of-the-art filter results. Using the same optimal filter settings for the testing datasets, similar quantitative results were found. Lesion conspicuity was improved on clinical scans when compared with EANM reconstructions, with no visible artifacts. Conclusions: Our three-stage adaptive filter achieved state-of-the-art quantitative performance for PET imaging. Harmonization tolerances with lower bias and variance than EANM guidelines were achieved for a variety of scanner models. CRCmax values were close to unity and the quantification variability was reduced when compared with standard reconstructions.

SIRF: Synergistic Image Reconstruction Framework

The combination of positron emission tomography (PET) with magnetic resonance (MR) imaging opens the way to more accurate diagnosis and improved patient management. At present, the data acquired by PET-MR scanners are essentially processed separately, but the opportunity to improve accuracy of the tomographic reconstruction via synergy of the two imaging techniques is an active area of research.In this paper, we present Release 2.1.0 of the CCP-PETMR Synergistic Image Reconstruction Framework (SIRF) software suite, providing an open-source software platform for efficient implementation and validation of novel reconstruction algorithms. SIRF provides user-friendly Python and MATLAB interfaces built on top of C++ libraries. SIRF uses advanced PET and MR reconstruction software packages and tools. Currently, for PET this is Software for Tomographic Image Reconstruction (STIR); for MR, Gadgetron and ISMRMRD; and for image registration tools, NiftyReg. The software aims to be capable of reconstructing images from acquired scanner data, whilst being simple enough to be used for educational purposes.The most recent version of the software can be downloaded from http://www.ccppetmr.ac.uk/downloads and https://github.com/CCPPETMR/. Program summaryProgram Title: Synergistic Image Reconstruction Framework (SIRF)Program Files DOI:http://dx.doi.org/10.17632/s45f5jh55j.1Licensing provisions: GPLv3 and Apache-2.0Programming languages: C++, C, Python, MATLABNature of problem: In current practice, data acquired by PET-MR scanners are processed separately. Methods for improving the accuracy of the tomographic reconstruction using the synergy of the two imaging techniques are actively being investigated by the PET-MR research and development community, however, practical application is heavily reliant on software. Open-source software available to the PET-MR community – such as the PET package (STIR) (Thielemans et al., 2012) and the MR package Gadgetron (Hansen and Sørensen, 2013) – provide a basis for new synergistic PET-MR software. However, these two software packages are independent and have very different software architectures. They are mostly written in C++ but many researchers in the PET-MR community are more familiar with script-style languages, such as Python and MATLAB, which enable rapid prototyping of novel reconstruction algorithms. In the current situation it is difficult for researchers to exploit any synergy between PET and MR data. Furthermore, techniques from one field cannot easily be applied in the other.Solution method: In SIRF, the bulk of computation is performed by available advanced open-source reconstruction and registration software (currently STIR, Gadgetron and NiftyReg) that can use multithreading and GPUs. The SIRF C++ code provides a thin layer on top of these existing libraries. The SIRF layer has unified data-containers and access mechanisms. This C++ layer provides the basis for a simple and intuitive Python and MATLAB interface, enabling users to quickly develop and test their reconstruction algorithms using these scripting languages only. At the same time, advanced users proficient in C++ can directly utilise wider SIRF functionality via the SIRF C++ libraries that we provide.

Analytic time-of-flight positron emission tomography reconstruction: two-dimensional case

In a positron emission tomography (PET) scanner, the time-of-flight (TOF) information gives us rough event position along the line-of-response (LOR). Using the TOF information for PET image reconstruction is able to reduce image noise. The state-of-the-art TOF PET image reconstruction uses iterative algorithms. Analytical image reconstruction algorithm exits for TOF PET which emulates the iterative Landweber algorithm. This paper introduces such an algorithm, focusing on two-dimensional (2D) reconstruction. The proposed algorithm is in the form of backprojection filtering, in which the backprojection is performed first, and then a 2D filter is applied to the backprojected image. For the list-mode data, the backprojection is carried out in the event-by-event fashion, and a profile function may be used along the projection LOR. The 2D filter depends on the TOF timing resolution as well as the backprojection profile function. In order to emulate the iterative algorithm effects, a Fourier-domain window function is suggested. This window function has a parameter, k, which corresponds to the iteration number in an iterative algorithm.

3D Tensor Based Nonlocal Low Rank Approximation in Dynamic PET Reconstruction.

Reconstructing images from multi-view projections is a crucial task both in the computer vision community and in the medical imaging community, and dynamic positron emission tomography (PET) is no exception. Unfortunately, image quality is inevitably degraded by the limitations of photon emissions and the trade-off between temporal and spatial resolution. In this paper, we develop a novel tensor based nonlocal low-rank framework for dynamic PET reconstruction. Spatial structures are effectively enhanced not only by nonlocal and sparse features, but momentarily by tensor-formed low-rank approximations in the temporal realm. Moreover, the total variation is well regularized as a complementation for denoising. These regularizations are efficiently combined into a Poisson PET model and jointly solved by distributed optimization. The experiments demonstrated in this paper validate the excellent performance of the proposed method in dynamic PET.

Faster PET reconstruction with non-smooth priors by randomization and preconditioning

Uncompressed clinical data from modern positron emission tomography (PET) scanners are very large, exceeding 350 million data points (projection bins). The last decades have seen tremendous advancements in mathematical imaging tools many of which lead to non-smooth (i.e. non-differentiable) optimization problems which are much harder to solve than smooth optimization problems. Most of these tools have not been translated to clinical PET data, as the state-of-the-art algorithms for non-smooth problems do not scale well to large data. In this work, inspired by big data machine learning applications, we use advanced randomized optimization algorithms to solve the PET reconstruction problem for a very large class of non-smooth priors which includes for example total variation, total generalized variation, directional total variation and various different physical constraints. The proposed algorithm randomly uses subsets of the data and only updates the variables associated with these. While this idea often leads to divergent algorithms, we show that the proposed algorithm does indeed converge for any proper subset selection. Numerically, we show on real PET data (FDG and florbetapir) from a Siemens Biograph mMR that about ten projections and backprojections are sufficient to solve the MAP optimisation problem related to many popular non-smooth priors; thus showing that the proposed algorithm is fast enough to bring these models into routine clinical practice.

CT radiomics and PET radiomics: ready for clinical implementation?

Today, rapid technical and clinical developments result in an increasing number of treatment options for oncological diseases. Thus, decision support systems are needed to offer the right treatment to the right patient. Imaging biomarkers hold great promise in patient-individual treatment guidance. Routinely performed for diagnosis and staging, imaging datasets are expected to hold more information than used in the clinical practice. Radiomics describes the extraction of a large number of meaningful quantitative features from medical images, such as computed tomography (CT) and positron emission tomography (PET). Due to the non-invasive nature and ability to capture 3D image-based heterogeneity, radiomic features are potential surrogate markers of the cancer phenotype. Several radiomic studies are published per day, owing to encouraging results of many radiomics-based patient outcome models. Despite this comparably large number of studies, radiomics is mainly studied in proof of principle concept. Hence, a translation of radiomics from a hot topic research field into an essential clinical decision-making tool is lacking, but of high clinical interest. Herein, we present a literature review addressing the clinical evidence of CT and PET radiomics. An extensive literature review was conducted in PubMed, including papers on robustness and clinical applications. We summarize image-modality related influences on the robustness of radiomic features and provide an overview of clinical evidence reported in the literature. Today, more evidence has been provided for CT imaging, however, PET imaging offers the promise of direct imaging of biological processes and functions. We provide a summary of future research directions, which needs to be addressed in order to successfully introduce radiomics into clinical medicine. In comparison to CT, more focus should be directed towards harmonization of PET acquisition and reconstruction protocols, which is important for transferable modelling. Both CT and PET radiomics are promising pre-treatment and intra-treatment biomarkers for outcome prediction. Most studies are performed in retrospective setting, however their validation in prospective data collections is ongoing.

Three-dimensional convolutional neural networks for simultaneous dual-tracer PET imaging

Dual-tracer positron emission tomography (PET) is a promising technique to measure the distribution of two tracers in the body by a single scan, which can improve the clinical accuracy of disease diagnosis and can also serve as a research tool for scientists. Most current research on dual-tracer PET reconstruction is based on mixed images pre-reconstructed by algorithms, which restricts the further improvement of the precision of reconstruction. In this study, we present a hybrid loss-guided deep learning based framework for dual-tracer PET imaging using sinogram data, which can achieve reconstruction by naturally unifying two functions: the reconstruction of the mixed images and the separation for individual tracers. Combined with volumetric dual-tracer images, we adopted a three-dimensional (3D) convolutional neural network (CNN) to learn full features, including spatial information and temporal information simultaneously. In addition, an auxiliary loss layer was introduced to guide the reconstruction of the dual tracers. We used Monte Carlo simulations with data augmentation to generate sufficient datasets for training and testing. The results were analyzed by the bias and variance both spatially (different regions of interest) and temporally (different frames). The analysis verified the feasibility of the 3D CNN framework for dual-tracer reconstruction. Furthermore, we compared the reconstruction results with a deep belief network (DBN), which is another deep learning based technique for the separation of dual-tracer images based on time–activity curves (TACs). The comparison results provide insights into the superior features and performance of the 3D CNN. Furthermore, we tested the [11C]FMZ–[11C]DTBZ images with three total-counts levels (, , ), which indicate different noise ratios. The analysis results demonstrate that our method can successfully recover the respective distribution of lower total counts with nearly the same accuracy as that of the higher total counts in the total counts range we applied, which also also indicates the proposed 3D CNN framework is more robust to noise compared with DBN.

Ultra-low-dose PET reconstruction using generative adversarial network with feature matching and task-specific perceptual loss.

Our goal was to use a generative adversarial network (GAN) with feature matching and task-specific perceptual loss to synthesize standard-dose amyloid Positron emission tomography (PET) images of high quality and including accurate pathological features from ultra-low-dose PET images only. Forty PET datasets from 39 participants were acquired with a simultaneous PET/MRI scanner following injection of 330±30MBq of the amyloid radiotracer 18F-florbetaben. The raw list-mode PET data were reconstructed as the standard-dose ground truth and were randomly undersampled by a factor of 100 to reconstruct 1% low-dose PET scans. A 2D encoder-decoder network was implemented as the generator to synthesize a standard-dose image and a discriminator was used to evaluate them. The two networks contested with each other to achieve high-visual quality PET from the ultra-low-dose PET. Multi-slice inputs were used to reduce noise by providing the network with 2.5D information. Feature matching was applied to reduce hallucinated structures. Task-specific perceptual loss was designed to maintain the correct pathological features. The image quality was evaluated by peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and root mean square error (RMSE) metrics with and without each of these modules. Two expert radiologists were asked to score image quality on a 5-point scale and identified the amyloid status (positive or negative). With only low-dose PET as input, the proposed method significantly outperformed Chen et al.'s method (Chen et al. Radiology. 2018;290:649-656) (which shows the best performance in this task) with the same input (PET-only model) by 1.87dB in PSNR, 2.04% in SSIM, and 24.75% in RMSE. It also achieved comparable results to Chen et al.'s method which used additional magnetic resonance imaging (MRI) inputs (PET-MR model). Experts' reading results showed that the proposed method could achieve better overall image quality and maintain better pathological features indicating amyloid status than both PET-only and PET-MR models proposed by Chen et al. CONCLUSION: Standard-dose amyloid PET images can be synthesized from ultra-low-dose images using GAN. Applying adversarial learning, feature matching, and task-specific perceptual loss are essential to ensure image quality and the preservation of pathological features.

PET Reconstruction With Non-Negativity Constraint in Projection Space: Optimization Through Hypo-Convergence.

Standard positron emission tomography (PET) reconstruction techniques are based on maximum-likelihood (ML) optimization methods, such as the maximum-likelihood expectation-maximization (MLEM) algorithm and its variations. Most methodologies rely on a positivity constraint on the activity distribution image. Although this constraint is meaningful from a physical point of view, it can be a source of bias for low-count/high-background PET, which can compromise accurate quantification. Existing methods that allow for negative values in the estimated image usually utilize a modified log-likelihood, and therefore break the data statistics. In this paper, we propose to incorporate the positivity constraint on the projections only, by approximating the (penalized) log-likelihood function by an adequate sequence of objective functions that are easily maximized without constraint. This sequence is constructed such that there is hypo-convergence (a type of convergence that allows the convergence of the maximizers under some conditions) to the original log-likelihood, hence allowing us to achieve maximization with positivity constraint on the projections using simple settings. A complete proof of convergence under weak assumptions is given. We provide results of experiments on simulated data where we compare our methodology with the alternative direction method of multipliers (ADMM) method, showing that our algorithm converges to a maximizer, which stays in the desired feasibility set, with faster convergence than ADMM. We also show that this approach reduces the bias, as compared with MLEM images, in necrotic tumors-which are characterized by cold regions surrounded by hot structures-while reconstructing similar activity values in hot regions.

Probabilistic Graphical Models for Dynamic PET: A Novel Approach to Direct Parametric Map Estimation and Image Reconstruction.

In the context of dynamic emission tomography, the conventional processing pipeline consists of independent image reconstruction of single-time frames, followed by the application of a suitable kinetic model to time-activity curves (TACs) at the voxel or region-of-interest level. Direct 4D positron emission tomography (PET) reconstruction, by contrast, seeks to move beyond this scheme and incorporate information from multiple time frames within the reconstruction task. Established direct methods are based on a deterministic description of voxelwise TACs, captured by the chosen kinetic model, considering the photon counting process the only source of uncertainty. In this paper, we introduce a new probabilistic modeling strategy based on the key assumption that activity time course would be subject to uncertainty even if the parameters of the underlying dynamic process are known. This leads to a hierarchical model that we formulate using the formalism of probabilistic graphical modeling. The inference is addressed using a new iterative algorithm, in which kinetic modeling results are treated as prior expectation of activity time course, rather than as a deterministic match, making it possible to control the trade-off between a data-driven and a model-driven reconstruction. The proposed method is flexible to an arbitrary choice of (linear and nonlinear) kinetic models, it enables the inclusion of arbitrary (sub)differentiable priors for parametric maps, and it is simple to implement. Computer simulations and an application to a real-patient scan show how the proposed method is able to generalize over conventional indirect and direct approaches, providing a bridge between them by properly tuning the impact of the kinetic modeling step on image reconstruction.

Brain-Dedicated Emission Tomography Systems: A Perspective on Requirements for Clinical Research and Clinical Needs in Brain Imaging

Research applications of human brain positron emission tomography (PET) imaging have been in place for over 40 years. The unique combination of quantitative PET instrumentation and reconstruction, novel radiotracers, and kinetic modeling has led to the development and application of a huge number of imaging paradigms. These have been applied to the understanding of normal function and pathophysiology, the development of drugs for neuropsychiatric disorders, and the study of drug interactions with the brain. I have had been fortunate to be part of this great endeavor throughout my research career, using brain-dedicated PET scanners, such as the HRRT [item [1] in the Appendix]. Over time, these systems have offered important advantages over whole-body systems. However, due to the limited clinical market for brain PET imaging, scanner manufacturers have focused exclusively on large field-of-view systems for the past two decades [item [2] in the Appendix]. The goal of this perspective is to share a viewpoint on the utility of brain-dedicated PET systems for human research, particularly in the context of novel radiotracers with dynamic studies using kinetic modeling.

Left Ventricular Remodeling Results in Homogenization of Myocardial Work Distribution.

The interaction between regional left ventricular (LV) myocardial work and metabolism in remodeled hearts has not yet been well established. Our aim was to investigate the effect of inhomogeneous LV work distribution on regional metabolism and remodeling in our animal model with reversible dyssynchrony due to pacing. In 12 sheep, 8 weeks of right atrial and right ventricular free wall (DDD) pacing lead to LV dilatation, a thinned septum, and thickened lateral wall. Left bundle branch block-like dyssynchrony caused by DDD pacing could be acutely reverted by right atrial pacing (AAI) only. Invasive hemodynamics and echocardiography were used to assess regional work by stress-strain loop area and compared with regional glucose metabolism measured by 18F-fluorodeoxyglucose positron emission tomography with and without improved spatial resolution by motion and anatomy correction on gated reconstructions. Glucose metabolism by positron emission tomography with anatomic correction on gated positron emission tomography reconstruction showed a different regional distribution than with clinical reconstructions and correlated best and significantly with regional myocardial work. At baseline, work was homogeneously distributed with normal conduction (AAI pacing), whereas during dyssynchrony (DDD pacing), the lateral wall was more loaded, and the septum was unloaded. After 8 weeks of remodeling under DDD pacing, however, an almost homogeneous work distribution was found with DDD pacing, whereas with AAI pacing, the thin septum showed exaggerated loading and the lateral walls a low load. Our experimental observations were confirmed in 5 patient responders to cardiac resynchronization therapy. Regional LV glucose metabolism closely correlates with regional work. Our data indicate that regionally different LV remodeling after exposure to inhomogeneous loading conditions, such as during LV dyssynchrony, is an adaptive process that helps to equilibrate work distribution. Correction of the inhomogeneous loading conditions, such as during cardiac resynchronization therapy, then triggers a reverse LV remodeling through the same mechanism.

Bias in PET Images of Solid Phantoms Due to CT-Based Attenuation Correction

The use of computed tomography (CT) images to correct for photon attenuation in positron emission tomography (PET) produces unbiased patient images, but it is not optimal for synthetic materials. For test objects made from epoxy, image bias and artifacts have been observed in well-calibrated PET/CT scanners. An epoxy used in commercially available sources was infused with long-lived 68Ge/68Ga nuclide and measured on several PET/CT scanners as well as on older PET scanners that measured attenuation with 511-keV photons. Bias in attenuation maps and PET images of phantoms was measured as imaging parameters and methods varied. Changes were made to the PET reconstruction to show the influence of CT-based attenuation correction. Additional attenuation measurements were made with a new epoxy intended for use in radiology and radiation treatment whose photonic properties mimic water. PET images of solid phantoms were biased by between 3% and 24% across variations in CT X-ray energy and scanner manufacturer. Modification of the reconstruction software reduced bias, but object-dependent changes were required to generate accurate attenuation maps. The water-mimicking epoxy formulation showed behavior similar to water in limited testing. For some solid phantoms, transformation of CT data to attenuation maps is a major source of PET image bias. The transformation can be modified to accommodate synthetic materials, but our data suggest that the problem may also be addressed by using epoxy formulations that are more compatible with PET/CT imaging.

A Krasnoselskii-Mann Algorithm With an Improved EM Preconditioner for PET Image Reconstruction.

This paper presents a preconditioned Krasnoselskii-Mann (KM) algorithm with an improved EM preconditioner (IEM-PKMA) for higher-order total variation (HOTV) regularized positron emission tomography (PET) image reconstruction. The PET reconstruction problem can be formulated as a three-term convex optimization model consisting of the Kullback-Leibler (KL) fidelity term, a nonsmooth penalty term, and a nonnegative constraint term which is also nonsmooth. We develop an efficient KM algorithm for solving this optimization problem based on a fixed-point characterization of its solution, with a preconditioner and a momentum technique for accelerating convergence. By combining the EM precondtioner, a thresholding, and a good inexpensive estimate of the solution, we propose an improved EM preconditioner that can not only accelerate convergence but also avoid the reconstructed image being "stuck at zero." Numerical results in this paper show that the proposed IEM-PKMA outperforms existing state-of-the-art algorithms including, the optimization transfer descent algorithm and the preconditioned L-BFGS-B algorithm for the differentiable smoothed anisotropic total variation regularized model, the preconditioned alternating projection algorithm, and the alternating direction method of multipliers for the nondifferentiable HOTV regularized model. Encouraging initial experiments using clinical data are presented.

Improving PET Imaging Acquisition and Analysis With Machine Learning: A Narrative Review With Focus on Alzheimer's Disease and Oncology.

Machine learning (ML) algorithms have found increasing utility in the medical imaging field and numerous applications in the analysis of digital biomarkers within positron emission tomography (PET) imaging have emerged. Interest in the use of artificial intelligence in PET imaging for the study of neurodegenerative diseases and oncology stems from the potential for such techniques to streamline decision support for physicians providing early and accurate diagnosis and allowing personalized treatment regimens. In this review, the use of ML to improve PET image acquisition and reconstruction is presented, along with an overview of its applications in the analysis of PET images for the study of Alzheimer's disease and oncology.

PET Image Reconstruction Using Deep Image Prior.

Recently, deep neural networks have been widely and successfully applied in computer vision tasks and have attracted growing interest in medical imaging. One barrier for the application of deep neural networks to medical imaging is the need for large amounts of prior training pairs, which is not always feasible in clinical practice. This is especially true for medical image reconstruction problems, where raw data are needed. Inspired by the deep image prior framework, in this paper, we proposed a personalized network training method where no prior training pairs are needed, but only the patient's own prior information. The network is updated during the iterative reconstruction process using the patient-specific prior information and measured data. We formulated the maximum-likelihood estimation as a constrained optimization problem and solved it using the alternating direction method of multipliers algorithm. Magnetic resonance imaging guided positron emission tomography reconstruction was employed as an example to demonstrate the effectiveness of the proposed framework. Quantification results based on simulation and real data show that the proposed reconstruction framework can outperform Gaussian post-smoothing and anatomically guided reconstructions using the kernel method or the neural-network penalty.