Positron Emission Tomography Image Reconstruction Research Articles

Incomplete PET projection data reconstruction method and simulation experiment

In positron emission tomography (PET) imaging, the whole body reconstruction is not required if the region of interest (ROI) is only a part of the object. This paper researched a local image reconstruction algorithm based on the PI-lines, and proposed the derivative-backprojection-Hilbert (DBH) algorithm which can use the incomplete PET projection data to achieve accurate reconstruction of ROI. We designed simulation experiments to evaluate the performance of the DBH algorithm with thorax model as the object of PET image reconstruction, and compared root mean square error (RMSE) of DBH reconstruction and RMSE of filtered backprojection (FBP) reconstruction from simulations data sets with and without noise. The result shows that DBH algorithm has advantages over traditional FBP algorithm, the RMSE decreases by 49.07% and 51.19% with the coincidence events number of 20 million and 40 million, respectively. Compared with traditional maximum-likelihood expectation-maximization (ML-EM), the local iteration algorithm which uses DBH reconstruction as the initial image produces a better reconstructed image at fewer iterations with the coincidence events number of 60 million. DBH algorithm is expected to be applied to fast two-dimensional local image reconstruction and optimize the local iterative reconstruction algorithm in clinical diagnosis.

Super-Iterative Image Reconstruction in PET

Despite its success in many biomedical applications, Positron Emission Tomography (PET) has the drawback of typically having lower spatial resolution and higher noise respect to other medical imaging techniques. The best achievable spatial resolution in PET scanners is limited by factors such as the positron range, non-collinearity and the size of the detector crystals. In this work, we present a novel method that uses series of image reconstructions (super-iterations) to go beyond the expected resolution-noise limits for a given PET acquisition. The image quality improvement is achieved using the projections of the previous image reconstruction to redistribute the measured counts of each line-of-response (LOR) into several subLORs, from which a new activity distribution with better quality is reconstructed. The method was evaluated with data from the preclinical scanner 4R-SuperArgus PET/CT, using the NEMA NU4-2008 image quality phantom, a cold Derenzo phantom, and an in-vivo FDG cardiac study on a rat. Resolution and recovery coefficient (RC) improvement of ∼10% was achieved while keeping the same noise level. Qualitative results from the in-vivo study also confirm this improvement in image quality. The proposed method is able to achieve significantly better images at the expense of a modest increase of the computational time, and it could be also applied to other modalities, such as SPECT and Compton Cameras.

Deep-Learning Based Positron Range Correction of PET Images

Positron emission tomography (PET) is a molecular imaging technique that provides a 3D image of functional processes in the body in vivo. Some of the radionuclides proposed for PET imaging emit high-energy positrons, which travel some distance before they annihilate (positron range), creating significant blurring in the reconstructed images. Their large positron range compromises the achievable spatial resolution of the system, which is more significant when using high-resolution scanners designed for the imaging of small animals. In this work, we trained a deep neural network named Deep-PRC to correct PET images for positron range effects. Deep-PRC was trained with modeled cases using a realistic Monte Carlo simulation tool that considers the positron energy distribution and the materials and tissues it propagates into. Quantification of the reconstructed PET images corrected with Deep-PRC showed that it was able to restore the images by up to 95% without any significant noise increase. The proposed method, which is accessible via Github, can provide an accurate positron range correction in a few seconds for a typical PET acquisition.

A proof-of-concept study of an in-situ partial-ring time-of-flight PET scanner for proton beam verification.

Development of a PET system capable of in-situ imaging requires a design that can accommodate the proton treatment beam nozzle. Among the several PET instrumentation approaches developed thus far, the dual-panel PET scanner is often used as it is simpler to develop and integrate within the proton therapy gantry. Partial-angle coverage of these systems can however lead to limited-angle artefacts in the reconstructed PET image. We have previously demonstrated via simulations that time-of-flight (TOF) reconstruction reduces the artifacts accompanying limited-angle data, and permits proton range measurement with 1-2 mm accuracy and precision. In this work we show measured results from a small proof-of-concept dual-panel PET system that uses TOF information to reconstruct PET data acquired after proton irradiation. The PET scanner comprises of two detector modules, each comprised of an array of 4×4×30 mm3 lanthanum bromide scintillator. Measurements are performed with an oxygen-rich gel-water, an adipose tissue equivalent material, and in vitro tissue phantoms. For each phantom measurement, 2 Gy dose was deposited using 54 - 100 MeV proton beams. For each phantom, a Monte Carlo simulation generating the expected distribution of PET isotope from the corresponding proton irradiation was also performed. Proton range was calculated by drawing multiple depth-profiles over a central region encompassing the proton dose deposition. For each profile, proton range was calculated using two techniques (a) 50% pick-off from the distal edge of the profile, and (b) comparing the measured and Monte Carlo profile to minimize the absolute sum of differences over the entire profile. A 10 min PET acquisition acquired with minimal delay post proton-irradiation is compared with a 10 min PET scan acquired after a 20 min delay. Measurements show that PET acquisition with minimal delay is necessary to collect 15O signal, and maximize 11C signal collection with a short PET acquisition. In comparison with the 50% pick-off technique, the shift technique is more robust and offers better precision in measuring the proton range for the different phantoms. Range measurements from PET images acquired with minimal delay, and the shift technique demonstrate the ability to achieve <1.5 mm accuracy and precision in estimating proton range.

PET-enabled dual-energy CT: image reconstruction and a proof-of-concept computer simulation study

Standard dual-energy computed tomography (CT) uses two different x-ray energies to obtain energy-dependent tissue attenuation information to allow quantitative material decomposition. The combined use of dual-energy CT and positron emission tomography (PET) may provide a more comprehensive characterization of disease states in cancer and other diseases. However, the integration of dual-energy CT with PET is not trivial, either requiring costly hardware upgrades or increasing radiation exposure. This paper proposes a different dual-energy CT imaging method that is enabled by PET. Instead of using a second x-ray CT scan with a different energy, this method exploits time-of-flight PET image reconstruction via the maximum likelihood attenuation and activity (MLAA) algorithm to obtain a 511 keV gamma-ray attenuation image from PET emission data. The high-energy gamma-ray attenuation image is then combined with the low-energy x-ray CT of PET/CT to provide a pair of dual-energy CT images. A major challenge with the standard MLAA reconstruction is the high noise present in the reconstructed 511 keV attenuation map, which would not compromise the PET activity reconstruction too much but may significantly affect the performance of the gamma-ray attenuation image for material decomposition. To overcome the problem, we further propose a kernel MLAA algorithm to exploit the prior information from the available x-ray CT image. We conducted a computer simulation to test the concept and algorithm for the task of material decomposition. The simulation results demonstrate that this PET-enabled dual-energy CT method is promising for quantitative material decomposition. The proposed method can be readily implemented on time-of-flight PET/CT scanners to enable simultaneous PET and dual-energy CT imaging.

Penalized-Likelihood PET Image Reconstruction Using 3D Structural Convolutional Sparse Coding.

Positron emission tomography (PET) is widely used for clinical diagnosis. As PET suffers from low resolution and high noise, numerous efforts try to incorporate anatomical priors into PET image reconstruction, especially with the development of hybrid PET/CT and PET/MRI systems. In this work, we proposed a cube-based 3D structural convolutional sparse coding (CSC) concept for penalized-likelihood PET image reconstruction, named 3D PET-CSC. The proposed 3D PET-CSC takes advantage of the convolutional operation and manages to incorporate anatomical priors without the need of registration or supervised training. As 3D PET-CSC codes the whole 3D PET image, instead of patches, it alleviates the staircase artifacts commonly presented in traditional patch-based sparse coding methods. Compared with traditional coding methods in Fourier domain, the proposed method extends the 3D CSC to a straightforward approach based on the pursuit of localized cubes. Moreover, we developed the residual-image and order-subset mechanisms to further reduce the computational cost and accelerate the convergence for the proposed 3D PET-CSC method. Experiments based on computer simulations and clinical datasets demonstrate the superiority of 3D PET-CSC compared with other reference methods.

FBP-Net for direct reconstruction of dynamic PET images

Dynamic positron emission tomography (PET) imaging can provide information about metabolic changes over time, used for kinetic analysis and auxiliary diagnosis. Existing deep learning-based reconstruction methods have too many trainable parameters and poor generalization, and require mass data to train the neural network. However, obtaining large amounts of medical data is expensive and time-consuming. To reduce the need for data and improve the generalization of network, we combined the filtered back-projection (FBP) algorithm with neural network, and proposed FBP-Net which could directly reconstruct PET images from sinograms instead of post-processing the rough reconstruction images obtained by traditional methods. The FBP-Net contained two parts: the FBP part and the denoiser part. The FBP part adaptively learned the frequency filter to realize the transformation from the detector domain to the image domain, and normalized the coarse reconstruction images obtained. The denoiser part merged the information of all time frames to improve the quality of dynamic PET reconstruction images, especially the early time frames. The proposed FBP-Net was performed on simulation and real dataset, and the results were compared with the state-of-art U-net and DeepPET. The results showed that FBP-Net did not tend to overfit the training set and had a stronger generalization.

Crystal surface and reflector optimization for the SiPM-based dual-ended readout TOF-DOI PET detector

Silicon photomultipliers (SiPMs) are now widely used for positron emission tomography (PET) applications because of their high gain and low noise characteristics. The PET image quality has been improved with the advancement of time-of-flight (TOF) and depth-of-interaction (DOI) measurement techniques. For brain-dedicated PET systems, both TOF and DOI information are beneficial for enhancing the reconstructed PET image quality. In a previous study, we proposed SiPM-based dual-ended readout PET detectors that used a mean time method to achieve coincidence timing resolution (CTR) of 349 ps and DOI resolution of 2.9 mm. However, the coincidence timing resolution (CTR) was worse than 300 ps since the crystal surface and the reflector type were not optimized. This study aimed at investigating the optimal crystal surface treatment and the reflector material to achieve a sub-200 ps CTR and sub-3 mm DOI resolution with a dual-ended readout PET detector using an LYSO crystal (2.9 × 2.9 × 20 mm3). The scintillation light inside the LYSO crystal was read out by two SiPMs using the dual-ended readout method. The CTR and DOI resolution were measured with two different crystal surfaces (polished and saw-cut) and three different reflector material scenarios of ESR without grease (i.e., air coupling), ESR with optical grease and Teflon. We digitized the timing and energy signals by using a V775N TDC module (35 ps bit−1) and V965 QDC module, respectively. The combination of the saw-cut LYSO crystal and the ESR with air coupling resulted in the best CTR (188 ± 32 ps) and DOI resolution (2.9 ± 0.2 mm) with the dual-ended readout configuration. We concluded the dual-ended readout method in combination with the saw-cut crystal and the ESR reflector with air coupling can provide a sub-200 ps CTR and sub-3.0 mm DOI resolution simultaneously.

PET Image Reconstruction Using a Cascading Back-Projection Neural Network

Positron emission tomography (PET) imaging is a noninvasive technique that makes it possible to probe biological metabolic processes in vivo . However, PET image reconstruction is challenging due to the ill-posedness of the inverse problem. Many image reconstruction methods have been proposed over the past few years to improve diagnostic performance. However, most of these methods can compromise the reconstruction of important high-frequency structural details after aggressive denoising. To address this problem, in this work, we propose a novel deep learning method that reconstructs PET images using a cascading back-projection neural network (bpNet). This network consists of a domain translation operation, which acts as prior knowledge, cascaded with a modified encoder-decoder network. The image reconstruction pipeline ranges from the sinogram to the back-projection image and then to the PET image. Quantitative results from simulation data and Derenzo phantom experiments with the small animal PET prototype system developed in our laboratory clearly demonstrate that our proposed method provides favorable reconstructed image quality, especially for low-count PET image reconstruction.

Feasibility study of patient-specific dose verification in proton therapy utilizing positron emission tomography (PET) and generative adversarial network (GAN).

Online dose verification based on proton-induced positron emitters is a promising strategy for quality assurance in proton therapy. Because of the nonlinear correlation between dose and the activity distributions, a machine learning-based approach was developed to establish their relationship. Simulations were carried out using a pencil beam scanning system and a computed tomography (CT) image-based phantom. A DiscoGAN model was developed to perform dose verification for both central and off-center lines. Besides the activity as input, HU information from CT images and stopping power (SP) prior were incorporated as auxiliary features for the model. The performance was quantitatively studied in terms of mean absolute error (MAE) and mean relative error (MRE), under different signal-to-noise ratios (SNRs). In addition to a dataset comprising monoenergetic beams, two additional datasets were generated to evaluate the model's generalization capability: five reconstructed PET images based on an in-beam PET system and a dataset comprising spread-out Bragg peaks (SOBPs). The feasibility of dose verification was successfully demonstrated for all three datasets. For the monoenergetic case (i.e., raw activity of positron emitters), the MRE is found to be <1% for the central lines and 5% for the off-center lines, respectively. The range uncertainty is found to be less than 1mm. The prediction based on five PET images, which take into account the detection of 511-keV photons and image reconstruction, yields slightly inferior performance. For the SOBP case, the MRE of the center lines is found to be <3% and the range uncertainty is <1mm. The inclusion of anatomical information (HU and SP) improves both accuracy and generalization of the DiscoGAN model. The combination of proton-induced positron emitters, in-beam PET, and machine learning may become a useful tool allowing for patient-specific online dose verification in proton therapy.

APRICOT: Advanced Platform for Reproducible Infrastructures in the Cloud via Open Tools.

Background Scientific publications are meant to exchange knowledge among researchers but the inability to properly reproduce computational experiments limits the quality of scientific research. Furthermore, bibliography shows that irreproducible preclinical research exceeds 50%, which produces a huge waste of resources on nonprofitable research at Life Sciences field. As a consequence, scientific reproducibility is being fostered to promote Open Science through open databases and software tools that are typically deployed on existing computational resources. However, some computational experiments require complex virtual infrastructures, such as elastic clusters of PCs, that can be dynamically provided from multiple clouds. Obtaining these infrastructures requires not only an infrastructure provider, but also advanced knowledge in the cloud computing field. Objectives The main aim of this paper is to improve reproducibility in life sciences to produce better and more cost-effective research. For that purpose, our intention is to simplify the infrastructure usage and deployment for researchers. Methods This paper introduces Advanced Platform for Reproducible Infrastructures in the Cloud via Open Tools (APRICOT), an open source extension for Jupyter to deploy deterministic virtual infrastructures across multiclouds for reproducible scientific computational experiments. To exemplify its utilization and how APRICOT can improve the reproduction of experiments with complex computation requirements, two examples in the field of life sciences are provided. All requirements to reproduce both experiments are disclosed within APRICOT and, therefore, can be reproduced by the users. Results To show the capabilities of APRICOT, we have processed a real magnetic resonance image to accurately characterize a prostate cancer using a Message Passing Interface cluster deployed automatically with APRICOT. In addition, the second example shows how APRICOT scales the deployed infrastructure, according to the workload, using a batch cluster. This example consists of a multiparametric study of a positron emission tomography image reconstruction. Conclusion APRICOT's benefits are the integration of specific infrastructure deployment, the management and usage for Open Science, making experiments that involve specific computational infrastructures reproducible. All the experiment steps and details can be documented at the same Jupyter notebook which includes infrastructure specifications, data storage, experimentation execution, results gathering, and infrastructure termination. Thus, distributing the experimentation notebook and needed data should be enough to reproduce the experiment.

Deep Learning for PET Image Reconstruction

This article reviews the use of a subdiscipline of artificial intelligence (AI), deep learning, for the reconstruction of images in positron emission tomography (PET). Deep learning can be used either directly or as a component of conventional reconstruction, in order to reconstruct images from noisy PET data. The review starts with an overview of conventional PET image reconstruction and then covers the principles of general linear and convolution-based mappings from data to images, and proceeds to consider nonlinearities, as used in convolutional neural networks (CNNs). The direct deep-learning methodology is then reviewed in the context of PET reconstruction. Direct methods learn the imaging physics and statistics from scratch, not relying on a priori knowledge of these models of the data. In contrast, model-based or physics-informed deep-learning uses existing advances in PET image reconstruction, replacing conventional components with deep-learning data-driven alternatives, such as for the regularization. These methods use trusted models of the imaging physics and noise distribution, while relying on training data examples to learn deep mappings for regularization and resolution recovery. After reviewing the main examples of these approaches in the literature, the review finishes with a brief look ahead to future directions.

Management of Respiratory Motion Artefacts in 18F-fluorodeoxyglucose Positron Emission Tomography using an Amplitude-Based Optimal Respiratory Gating Algorithm.

Positron emission tomography (PET) combined with X-ray computed tomography (CT) is an important molecular imaging platform that is required for accurate diagnosis and clinical staging of a variety of diseases. The advantage of PET imaging is the ability to visualize and quantify a myriad of biological processes in vivo with high sensitivity and accuracy. However, there are multiple factors that determine image quality and quantitative accuracy of PET images. One of the foremost factors influencing image quality in PET imaging of the thorax and upper abdomen is respiratory motion, resulting in respiration-induced motion blurring of anatomical structures. Correction of these artefacts is required for providing optimal image quality and quantitative accuracy of PET images. Several respiratory gating techniques have been developed, typically relying on acquisition of a respiratory signal simultaneously with PET data. Based on the respiratory signal acquired, PET data is selected for reconstruction of a motion-free image. Although these methods have been shown to effectively remove respiratory motion artefacts from PET images, the performance is dependent on the quality of the respiratory signal being acquired. In this study, the use of an amplitude-based optimal respiratory gating (ORG) algorithm is discussed. In contrast to many other respiratory gating algorithms, ORG permits the user to have control over image quality versus the amount of rejected motion in the reconstructed PET images. This is achieved by calculating an optimal amplitude range based on the acquired surrogate signal and a user-specified duty cycle (the percentage of PET data used for image reconstruction). The optimal amplitude range is defined as the smallest amplitude range still containing the amount of PET data required for image reconstruction. It was shown that ORG results in effective removal of respiration-induced image blurring in PET imaging of the thorax and upper abdomen, resulting in improved image quality and quantitative accuracy.

Myocardial creep-induced misalignment artifacts in PET/MR myocardial perfusion imaging

PurposeMisalignment between positron emission tomography (PET) datasets and attenuation correction (AC) maps is a potential source of artifacts in myocardial perfusion imaging (MPI). We assessed the impact of adenosine on the alignment of AC maps derived from magnetic resonance (MR) and PET datasets during MPI on a hybrid PET/MR scanner.MethodsTwenty-eight volunteers underwent adenosine stress and rest 13N-ammonia MPI on a PET/MR. We acquired Dixon sequences for the creation of MRAC maps. After reconstruction of the original non-shifted PET images, we examined MRAC and PET datasets for cardiac spatial misalignment and, if necessary, reconstructed a second set of shifted PET images after manually adjusting co-registration. Summed rest, stress, and difference scores (SRS, SSS, and SDS) were compared between shifted and non-shifted PET images. Additionally, we measured the amount of cranial movement of the heart (i.e., myocardial creep) after termination of adenosine infusion.ResultsRealignment was necessary for 25 (89.3%) stress and 12 (42.9%) rest PET datasets. Median SRS, SSS, and SDS of the non-shifted images were 6 (IQR = 4–7), 12 (IQR = 7–18), and 8 (IQR = 2–11), respectively, and of the shifted images 2 (IQR = 1–6), 4 (IQR = 7–18), and 1 (IQR = 0–2), respectively. All three scores were significantly higher in non-shifted versus shifted images (all p < 0.05). The difference in SDS correlated moderately but significantly with the amount of myocardial creep (r = 0.541, p = 0.005).ConclusionMisalignment of MRAC and PET datasets commonly occurs during adenosine stress MPI on a hybrid PET/MR device, potentially leading to an increase in false-positive findings. Our results suggest that myocardial creep may substantially account for this and prompt for a careful review and correction of PET/MRAC data.

PET imaging of neurotransmission using direct parametric reconstruction

Receptor ligand-based dynamic Positron Emission Tomography (PET) permits the measurement of neurotransmitter release in the human brain. For single-scan paradigms, the conventional method of estimating changes in neurotransmitter levels relies on fitting a pharmacokinetic model to activity concentration histories extracted after PET image reconstruction. However, due to the statistical fluctuations of activity concentration data at the voxel scale, parametric images computed using this approach often exhibit low signal-to-noise ratio, impeding characterization of neurotransmitter release. Numerous studies have shown that direct parametric reconstruction (DPR) approaches, which combine image reconstruction and kinetic analysis in a unified framework, can improve the signal-to-noise ratio of parametric mapping. However, there is little experience with DPR in imaging of neurotransmission and the performance of the approach in this application has not been evaluated before in humans. In this report, we present and evaluate a DPR methodology that computes 3-D distributions of ligand transport, binding potential (BPND) and neurotransmitter release magnitude (γ) from a dynamic sequence of PET sinograms. The technique employs the linear simplified reference region model (LSRRM) of Alpert et al. (2003), which represents an extension of the simplified reference region model that incorporates time-varying binding parameters due to radioligand displacement by release of neurotransmitter. Estimation of parametric images is performed by gradient-based optimization of a Poisson log-likelihood function incorporating LSRRM kinetics and accounting for the effects of head movement, attenuation, detector sensitivity, random and scattered coincidences. A 11C-raclopride simulation study showed that the proposed approach substantially reduces the bias and variance of voxel-wise γ estimates as compared to standard methods. Moreover, simulations showed that detection of release could be made more reliable and/or conducted using a smaller sample size using the proposed DPR estimator. Likewise, images of BPND computed using DPR had substantially improved bias and variance properties. Application of the method in human subjects was demonstrated using 11C-raclopride dynamic scans and a reward task, confirming the improved quality of the estimated parametric images using the proposed approach.

MR-based Attenuation Correction for Brain PET Using 3D Cycle-Consistent Adversarial Network.

Attenuation correction (AC) is important for the quantitative merits of positron emission tomography (PET). However, attenuation coefficients cannot be derived from magnetic resonance (MR) images directly for PET/MR systems. In this work, we aimed to derive continuous AC maps from Dixon MR images without the requirement of MR and computed tomography (CT) image registration. To achieve this, a 3D generative adversarial network with both discriminative and cycle-consistency loss (Cycle-GAN) was developed. The modified 3D U-net was employed as the structure of the generative networks to generate the pseudo CT/MR images. The 3D patch-based discriminative networks were used to distinguish the generated pseudo CT/MR images from the true CT/MR images. To evaluate its performance, datasets from 32 patients were used in the experiment. The Dixon segmentation and atlas methods provided by the vendor and the convolutional neural network (CNN) method which utilized registered MR and CT images were employed as the reference methods. Dice coefficients of the pseudo-CT image and the regional quantification in the reconstructed PET images were compared. Results show that the Cycle-GAN framework can generate better AC compared to the Dixon segmentation and atlas methods, and shows comparable performance compared to the CNN method.

Model-Based Deep Learning PET Image Reconstruction Using Forward-Backward Splitting Expectation-Maximization.

We propose a forward-backward splitting algorithm to integrate deep learning into maximum-a-posteriori (MAP) positron emission tomography (PET) image reconstruction. The MAP reconstruction is split into regularization, expectation-maximization (EM), and a weighted fusion. For regularization, the use of either a Bowsher prior (using Markov-random fields) or a residual learning unit (using convolutional-neural networks) were considered. For the latter, our proposed forward-backward splitting EM (FBSEM), accelerated with ordered subsets (OS), was unrolled into a recurrent-neural network in which network parameters (including regularization strength) are shared across all states and learned during PET reconstruction. Our network was trained and evaluated using PET-only (FBSEM-p) and PET-MR (FBSEM-pm) datasets for low-dose simulations and short-duration in-vivo brain imaging. It was compared to OSEM, Bowsher MAPEM, and a post-reconstruction U-Net denoising trained on the same PET-only (Unet-p) or PET-MR (Unet-pm) datasets. For simulations, FBSEM-p(m) and Unet-p(m) nets achieved a comparable performance, on average, 14.4% and 13.4% normalized root-mean square error (NRMSE), respectively; and both outperformed OSEM and MAPEM methods (with 20.7% and 17.7% NRMSE, respectively). For in-vivo datasets, FBSEM-p(m), Unet-p(m), MAPEM, and OSEM methods achieved average root-sum-of-squared errors of 3.9%, 5.7%, 5.9%, and 7.8% in different brain regions, respectively. In conclusion, the studied U-Net denoising method achieved a comparable performance to a representative implementation of the FBSEM net.

Generative adversarial network based regularized image reconstruction for PET

Positron emission tomography (PET) is an ill-posed inverse problem and suffers high noise due to limited number of detected events. Prior information can be used to improve the quality of reconstructed PET images. Deep neural networks have also been applied to regularized image reconstruction. One method is to use a pretrained denoising neural network to represent the PET image and to perform a constrained maximum likelihood estimation. In this work, we propose to use a generative adversarial network (GAN) to further improve the network performance. We also modify the objective function to include a data-matching term on the network input. Experimental studies using computer-based Monte Carlo simulations and real patient datasets demonstrate that the proposed method leads to noticeable improvements over the kernel-based and U-net-based regularization methods in terms of lesion contrast recovery versus background noise trade-offs.

Acceleration of PET-based industrial internal defect image reconstruction using FPGA

Positron emission tomography (PET) can be used to measure the internal defects of industrial parts. However, PET requires a long execution time of image reconstruction, which hinders its practical usage in industrial measurements. A novel parallel scheme based on field-programmable gate arrays (FPGAs) is proposed in this study to accelerate PET image reconstruction. A fast maximum-likelihood expectation–maximization iteration reconstruction algorithm with prior estimation is implemented on the FPGA. This method can achieve satisfactory PET images with limited iteration times. The resources in the FPGA are divided into several groups, and each group supports the image reconstruction for a single sinogram. Thus, several sinograms can be processed in parallel. Two internal defect detection experiments are conducted to apply the proposed method to industrial measurements. Results show that the inner structure can be detected, whereas the inner defects can be visualized. A group of 104 slice images is reconstructed in parallel on FPGAs, and the final 3D PET image of the inner defects is acquired in 10 s.

DPIR-Net: Direct PET Image Reconstruction Based on the Wasserstein Generative Adversarial Network

Positron emission tomography (PET) is an advanced medical imaging technique widely used in various clinical applications, such as tumor detection and neurologic disorders. Reducing the radiotracer dose is desirable in PET imaging because it decreases the patient’s radiation exposure. However, reducing the dose can also increase noise, affecting the image quality. Therefore, an advanced image reconstruction algorithm based on low-dose PET data is needed to improve the quality of the reconstructed image. In this article, we propose the use of a direct PET image reconstruction network (DPIR-Net) using an improved Wasserstein generative adversarial network (WGAN) framework to enhance image quality. This article provides two main findings. First, our network uses sinogram data as input and outputs high-quality PET images direct, resulting in shorter reconstruction times compared with traditional model-based reconstruction networks. Second, we combine perceptual loss, mean square error, and the Wasserstein distance as the loss function, which effectively solves the problem of excessive smoothness and loss of detailed information in traditional network image reconstruction. We performed a comparative study using maximum-likelihood expectation maximization (MLEM) with a post-Gaussian filter, a total variation (TV)-norm regularization, a nonlocal means (NLMs) denoising method, a neural network denoising method, a traditional deep learning PET reconstruction network, and our proposed DPIR-Net method and evaluated the proposed method using both human and mouse data. The mouse data were obtained from a small animal PET prototype system developed by our laboratory. The quantitative and qualitative results show that our proposed method outperformed the conventional methods.