Sequential deep learning image enhancement models improve diagnostic confidence, lesion detectability, and image reconstruction time in PET.

BACKGROUND: Investigate the potential benefits of sequential deployment of two deep learning (DL) algorithms namely DL-Enhancement (DLE) and DL-based time-of-flight (ToF) (DLT). DLE aims to enhance the rapidly reconstructed ordered-subset-expectation-maximisation algorithm (OSEM) images towards block-sequential-regularised-expectation-maximisation (BSREM) images, whereas DLT aims to improve the quality of BSREM images reconstructed without ToF. As the algorithms differ in their purpose, sequential application may allow benefits from each to be combined. 20 FDG PET-CT scans were performed on a Discovery 710 (D710) and 20 on Discovery MI (DMI; both GE HealthCare). PET data was reconstructed using five combinations of algorithms:1. ToF-BSREM, 2. ToF-OSEM + DLE, 3. OSEM + DLE + DLT, 4. ToF-OSEM + DLE + DLT, 5. ToF-BSREM + DLT. To assess image noise, 30 mm-diameter spherical VOIs were drawn in both lung and liver to measure standard deviation of voxels within the volume. In a blind clinical reading, two experienced readers rated the images on a five-point Likert scale based on lesion detectability, diagnostic confidence, and image quality. RESULTS: Applying DLE + DLT reduced noise whilst improving lesion detectability, diagnostic confidence, and image reconstruction time. ToF-OSEM + DLE + DLT reconstructions demonstrated an increase in lesion SUVmax of 28 ± 14% (average ± standard deviation) and 11 ± 5% for data acquired on the D710 and DMI, respectively. The same reconstruction scored highest in clinical readings for both lesion detectability and diagnostic confidence for D710. CONCLUSIONS: The combination of DLE and DLT increased diagnostic confidence and lesion detectability compared to ToF-BSREM images. As DLE + DLT used input OSEM images, and because DL inferencing was fast, there was a significant decrease in overall reconstruction time. This could have applications to total body PET.

Deep learning-based time-of-flight (ToF) image enhancement of non-ToF PET scans.

PURPOSE: To improve the quantitative accuracy and diagnostic confidence of PET images reconstructed without time-of-flight (ToF) using deep learning models trained for ToF image enhancement (DL-ToF). METHODS: A total of 273 [18F]-FDG PET scans were used, including data from 6 centres equipped with GE Discovery MI ToF scanners. PET data were reconstructed using the block-sequential-regularised-expectation-maximisation (BSREM) algorithm with and without ToF. The images were then split into training (n = 208), validation (n = 15), and testing (n = 50) sets. Three DL-ToF models were trained to transform non-ToF BSREM images to their target ToF images with different levels of DL-ToF strength (low, medium, high). The models were objectively evaluated using the testing set based on standardised uptake value (SUV) in 139 identified lesions, and in normal regions of liver and lungs. Three radiologists subjectively rated the models using testing sets based on lesion detectability, diagnostic confidence, and image noise/quality. RESULTS: The non-ToF, DL-ToF low, medium, and high methods resulted in - 28 ± 18, - 28 ± 19, - 8 ± 22, and 1.7 ± 24% differences (mean; SD) in the SUVmax for the lesions in testing set, compared to ToF-BSREM image. In background lung VOIs, the SUVmean differences were 7 ± 15, 0.6 ± 12, 1 ± 13, and 1 ± 11% respectively. In normal liver, SUVmean differences were 4 ± 5, 0.7 ± 4, 0.8 ± 4, and 0.1 ± 4%. Visual inspection showed that our DL-ToF improved feature sharpness and convergence towards ToF reconstruction. Blinded clinical readings of testing sets for diagnostic confidence (scale 0-5) showed that non-ToF, DL-ToF low, medium, and high, and ToF images scored 3.0, 3.0, 4.1, 3.8, and 3.5 respectively. For this set of images, DL-ToF medium therefore scored highest for diagnostic confidence. CONCLUSION: Deep learning-based image enhancement models may provide converged ToF-equivalent image quality without ToF reconstruction. In clinical scoring DL-ToF-enhanced non-ToF images (medium and high) on average scored as high as, or higher than, ToF images. The model is generalisable and hence, could be applied to non-ToF images from BGO-based PET/CT scanners.

Image enhancement of whole-body oncology [18F]-FDG PET scans using deep neural networks to reduce noise.

PURPOSE: To enhance the image quality of oncology [18F]-FDG PET scans acquired in shorter times and reconstructed by faster algorithms using deep neural networks. METHODS: List-mode data from 277 [18F]-FDG PET/CT scans, from six centres using GE Discovery PET/CT scanners, were split into ¾-, ½- and »-duration scans. Full-duration datasets were reconstructed using the convergent block sequential regularised expectation maximisation (BSREM) algorithm. Short-duration datasets were reconstructed with the faster OSEM algorithm. The 277 examinations were divided into training (n = 237), validation (n = 15) and testing (n = 25) sets. Three deep learning enhancement (DLE) models were trained to map full and partial-duration OSEM images into their target full-duration BSREM images. In addition to standardised uptake value (SUV) evaluations in lesions, liver and lungs, two experienced radiologists scored the quality of testing set images and BSREM in a blinded clinical reading (175 series). RESULTS: OSEM reconstructions demonstrated up to 22% difference in lesion SUVmax, for different scan durations, compared to full-duration BSREM. Application of the DLE models reduced this difference significantly for full-, ¾- and ½-duration scans, while simultaneously reducing the noise in the liver. The clinical reading showed that the standard DLE model with full- or ¾-duration scans provided an image quality substantially comparable to full-duration scans with BSREM reconstruction, yet in a shorter reconstruction time. CONCLUSION: Deep learning-based image enhancement models may allow a reduction in scan time (or injected activity) by up to 50%, and can decrease reconstruction time to a third, while maintaining image quality.

Motion-corrected and high-resolution anatomically assisted (MOCHA) reconstruction of arterial spin labeling MRI.

PURPOSE: A model-based reconstruction framework is proposed for motion-corrected and high-resolution anatomically assisted (MOCHA) reconstruction of arterial spin labeling (ASL) data. In this framework, all low-resolution ASL control-label pairs are used to reconstruct a single high-resolution cerebral blood flow (CBF) map, corrected for rigid-motion, point-spread-function blurring and partial volume effect. METHODS: Six volunteers were recruited for CBF imaging using pseudo-continuous ASL labeling, two-shot 3D gradient and spin-echo sequences and high-resolution T1 -weighted MRI. For 2 volunteers, high-resolution scans with double and triple resolution in the partition direction were additionally collected. Simulations were designed for evaluations against a high-resolution ground-truth CBF map, including a simulated hyperperfused lesion and hyperperfusion/hypoperfusion abnormalities. The MOCHA technique was compared with standard reconstruction and a 3D linear regression partial-volume effect correction method and was further evaluated for acquisitions with reduced control-label pairs and k-space undersampling. RESULTS: The MOCHA reconstructions of low-resolution ASL data showed enhanced image quality, particularly in the partition direction. In simulations, both MOCHA and 3D linear regression provided more accurate CBF maps than the standard reconstruction; however, MOCHA resulted in the lowest errors and well delineated the abnormalities. The MOCHA reconstruction of standard-resolution in vivo data showed good agreement with higher-resolution scans requiring 4-times and 9-times longer acquisitions. The MOCHA reconstruction was found to be robust for 4-times-accelerated ASL acquisitions, achieved by reduced control-label pairs or k-space undersampling. CONCLUSION: The MOCHA reconstruction reduces partial-volume effect by direct reconstruction of CBF maps in the high-resolution space of the corresponding anatomical image, incorporating motion correction and point spread function modeling. Following further evaluation, MOCHA should promote the clinical application of ASL.

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.

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.

Enhancement of Partial Volume Correction in MR-Guided PET Image Reconstruction by Using MRI Voxel Sizes.

Positron emission tomography (PET) suffers from poor spatial resolution which results in quantitative bias when evaluating the radiotracer uptake in small anatomical regions, such as the striatum in the brain which is of importance in this paper of neurodegenerative diseases. These partial volume effects need to be compensated for by employing partial volume correction (PVC) methods in order to achieve quantitatively accurate images. Two important PVC methods applied during the reconstruction are resolution modeling, which suffers from Gibbs artifacts, and penalized likelihood using anatomical priors. The introduction of clinical simultaneous PET-MR scanners has attracted new attention for the latter methods and brought new opportunities to use MRI information to assist PET image reconstruction in order to improve image quality. In this context, MR images are usually down-sampled to the PET resolution before being used in MR-guided PET reconstruction. However, the reconstruction of PET images using the MRI voxel size could achieve a better utilization of the high resolution anatomical information and improve the PVC obtained with these methods. In this paper, we evaluate the importance of the use of MRI voxel sizes when reconstructing PET images with MR-guided maximum a posteriori (MAP) methods, specifically the modified Bowsher method. We also propose a method to avoid the artifacts that arise when PET reconstructions are performed in a higher resolution matrix than the standard for a given scanner. The MR-guided MAP reconstructions were implemented with a modified Lange prior that included anatomical information with the Bowsher method. The methods were evaluated with and without resolution modeling for simulated and real brain data. We show that the use of the MRI voxel sizes when reconstructing PET images with MR-guided MAP enhances PVC by improving the contrast and reducing the bias in six different regions of the brain using regional metrics for a single simulated data set and ensemble metrics for ten noise realizations. Similar results were obtained for real data, where a good enhancement of the contrast was achieved. The combination of MR-guided MAP reconstruction with point-spread function modeling and MRI voxel sizes proved to be an attractive method to achieve considerable enhancement of PVC, while reducing and controlling the noise level and Gibbs artifacts.

Multi-modal synergistic PET and MR reconstruction using mutually weighted quadratic priors.

PURPOSE: To propose a framework for synergistic reconstruction of PET-MR and multi-contrast MR data to improve the image quality obtained from noisy PET data and from undersampled MR data. THEORY AND METHODS: Weighted quadratic priors were devised to preserve common boundaries between PET-MR images while reducing noise, PET Gibbs ringing, and MR undersampling artifacts. These priors are iteratively reweighted using normalized multi-modal Gaussian similarity kernels. Synergistic PET-MR reconstructions were built on the PET maximum a posteriori expectation maximization algorithm and the MR regularized sensitivity encoding method. The proposed approach was compared to conventional methods, total variation, and prior-image weighted quadratic regularization methods. Comparisons were performed on a simulated [18 F]fluorodeoxyglucose-PET and T1 /T2 -weighted MR brain phantom, 2 in vivo T1 /T2 -weighted MR brain datasets, and an in vivo [18 F]fluorodeoxyglucose-PET and fluid-attenuated inversion recovery/T1 -weighted MR brain dataset. RESULTS: Simulations showed that synergistic reconstructions achieve the lowest quantification errors for all image modalities compared to conventional, total variation, and weighted quadratic methods. Whereas total variation regularization preserved modality-unique features, this method failed to recover PET details and was not able to reduce MR artifacts compared to our proposed method. For in vivo MR data, our method maintained similar image quality for 3× and 14× accelerated data. Reconstruction of the PET-MR dataset also demonstrated improved performance of our method compared to the conventional independent methods in terms of reduced Gibbs and undersampling artifacts. CONCLUSION: The proposed methodology offers a robust multi-modal synergistic image reconstruction framework that can be readily built on existing established algorithms.

MR-Guided Kernel EM Reconstruction for Reduced Dose PET Imaging.

PET image reconstruction is highly susceptible to the impact of Poisson noise, and if shorter acquisition times or reduced injected doses are used, the noisy PET data become even more limiting. The recent development of kernel expectation maximisation (KEM) is a simple way to reduce noise in PET images, and we show in this work that impressive dose reduction can be achieved when the kernel method is used with MR-derived kernels. The kernel method is shown to surpass maximum likelihood expectation maximisation (MLEM) for the reconstruction of low-count datasets (corresponding to those obtained at reduced injected doses) producing visibly clearer reconstructions for unsmoothed and smoothed images, at all count levels. The kernel EM reconstruction of 10% of the data had comparable whole brain voxel-level error measures to the MLEM reconstruction of 100% of the data (for simulated data, at 100 iterations). For regional metrics, the kernel method at reduced dose levels attained a reduced coefficient of variation and more accurate mean values compared to MLEM. However, the advances provided by the kernel method are at the expense of possible over-smoothing of features unique to the PET data. Further assessment on clinical data is required to determine the level of dose reduction that can be routinely achieved using the kernel method, whilst maintaining the diagnostic utility of the scan.

Synergistic PET and SENSE MR Image Reconstruction Using Joint Sparsity Regularization.

In this paper, we propose a generalized joint sparsity regularization prior and reconstruction framework for the synergistic reconstruction of positron emission tomography (PET) and under sampled sensitivity encoded magnetic resonance imaging data with the aim of improving image quality beyond that obtained through conventional independent reconstructions. The proposed prior improves upon the joint total variation (TV) using a non-convex potential function that assigns a relatively lower penalty for the PET and MR gradients, whose magnitudes are jointly large, thus permitting the preservation and formation of common boundaries irrespective of their relative orientation. The alternating direction method of multipliers (ADMM) optimization framework was exploited for the joint PET-MR image reconstruction. In this framework, the joint maximum a posteriori objective function was effectively optimized by alternating between well-established regularized PET and MR image reconstructions. Moreover, the dependency of the joint prior on the PET and MR signal intensities was addressed by a novel alternating scaling of the distribution of the gradient vectors. The proposed prior was compared with the separate TV and joint TV regularization methods using extensive simulation and real clinical data. In addition, the proposed joint prior was compared with the recently proposed linear parallel level sets (PLSs) method using a benchmark simulation data set. Our simulation and clinical data results demonstrated the improved quality of the synergistically reconstructed PET-MR images compared with the unregularized and conventional separately regularized methods. It was also found that the proposed prior can outperform both the joint TV and linear PLS regularization methods in assisting edge preservation and recovery of details, which are otherwise impaired by noise and aliasing artifacts. In conclusion, the proposed joint sparsity regularization within the presented a ADMM reconstruction framework is a promising technique, nonetheless our clinical results showed that the clinical applicability of joint reconstruction might be limited in current PET-MR scanners, mainly due to the lower resolution of PET images.

MR-guided joint reconstruction of activity and attenuation in brain PET-MR.

With the advent of time-of-flight (TOF) PET scanners, joint maximum-likelihood reconstruction of activity and attenuation (MLAA) maps has recently regained attention for the estimation of PET attenuation maps from emission data. However, the estimated attenuation and activity maps are scaled by unknown scaling factors. We recently demonstrated that in hybrid PET-MR, the scaling issue of this algorithm can be effectively addressed by imposing MR spatial constraints on the estimation of attenuation maps using a penalized MLAA (P-MLAA+) algorithm. With the advent of simultaneous PET-MR systems, MRI-guided PET image reconstruction has also gained attention for improving the quantitative accuracy of PET images, usually degraded by noise and partial volume effects. The aim of this study is therefore to increase the benefits of MRI information for improving the quantitative accuracy of PET images by exploiting MRI-based anatomical penalty functions to guide the reconstruction of both activity and attenuation maps during their joint estimation. We employed an anato-functional joint entropy penalty function for the reconstruction of activity and an anatomical quadratic penalty function for the reconstruction of attenuation. The resulting algorithm was referred to as P-MLAA++ since it exploits both activity and attenuation penalty functions. The performance of the P-MLAA algorithms were compared with MLAA and the widely used activity reconstruction algorithms such as maximum likelihood expectation maximization (MLEM) and penalized MLEM (P-MLEM) both corrected for attenuation using a conventional MRI segmentation-based attenuation correction (MRAC) method. The studied methods were evaluated using simulations and clinical studies taking the PET image reconstructed using reference CT-based attenuation maps as a reference. The simulation results showed that the proposed method can notably improve the visual quality of the PET images by reducing noise while preserving structural boundaries and at the same time improving the quantitative accuracy of the PET images. Our clinical reconstruction results showed that the MLEM-MRAC, P-MLEM-MRAC, MLAA, P-MLAA+ and P-MLAA++ algorithms result in, on average, quantification errors of -13.5 ± 3.1%, -13.4 ± 3.1%, -2.0 ± 6.5%, -3.0 ± 3.5% and -4.2 ± 3.6%, respectively, in different regions of the brain. In conclusion, whilst the P-MLAA+ algorithm showed the best overall quantification performance, the proposed P-MLAA++ algorithm provided simultaneous partial volume and attenuation corrections with only a minor compromise of PET quantification.

PET image reconstruction using multi-parametric anato-functional priors.

In this study, we investigate the application of multi-parametric anato-functional (MR-PET) priors for the maximum a posteriori (MAP) reconstruction of brain PET data in order to address the limitations of the conventional anatomical priors in the presence of PET-MR mismatches. In addition to partial volume correction benefits, the suitability of these priors for reconstruction of low-count PET data is also introduced and demonstrated, comparing to standard maximum-likelihood (ML) reconstruction of high-count data. The conventional local Tikhonov and total variation (TV) priors and current state-of-the-art anatomical priors including the Kaipio, non-local Tikhonov prior with Bowsher and Gaussian similarity kernels are investigated and presented in a unified framework. The Gaussian kernels are calculated using both voxel- and patch-based feature vectors. To cope with PET and MR mismatches, the Bowsher and Gaussian priors are extended to multi-parametric priors. In addition, we propose a modified joint Burg entropy prior that by definition exploits all parametric information in the MAP reconstruction of PET data. The performance of the priors was extensively evaluated using 3D simulations and two clinical brain datasets of [18F]florbetaben and [18F]FDG radiotracers. For simulations, several anato-functional mismatches were intentionally introduced between the PET and MR images, and furthermore, for the FDG clinical dataset, two PET-unique active tumours were embedded in the PET data. Our simulation results showed that the joint Burg entropy prior far outperformed the conventional anatomical priors in terms of preserving PET unique lesions, while still reconstructing functional boundaries with corresponding MR boundaries. In addition, the multi-parametric extension of the Gaussian and Bowsher priors led to enhanced preservation of edge and PET unique features and also an improved bias-variance performance. In agreement with the simulation results, the clinical results also showed that the Gaussian prior with voxel-based feature vectors, the Bowsher and the joint Burg entropy priors were the best performing priors. However, for the FDG dataset with simulated tumours, the TV and proposed priors were capable of preserving the PET-unique tumours. Finally, an important outcome was the demonstration that the MAP reconstruction of a low-count FDG PET dataset using the proposed joint entropy prior can lead to comparable image quality to a conventional ML reconstruction with up to 5 times more counts. In conclusion, multi-parametric anato-functional priors provide a solution to address the pitfalls of the conventional priors and are therefore likely to increase the diagnostic confidence in MR-guided PET image reconstructions.

Assessment of metal artifact reduction methods in pelvic CT.

PURPOSE: Metal artifact reduction (MAR) produces images with improved quality potentially leading to confident and reliable clinical diagnosis and therapy planning. In this work, the authors evaluate the performance of five MAR techniques for the assessment of computed tomography images of patients with hip prostheses. METHODS: Five MAR algorithms were evaluated using simulation and clinical studies. The algorithms included one-dimensional linear interpolation (LI) of the corrupted projection bins in the sinogram, two-dimensional interpolation (2D), a normalized metal artifact reduction (NMAR) technique, a metal deletion technique, and a maximum a posteriori completion (MAPC) approach. The algorithms were applied to ten simulated datasets as well as 30 clinical studies of patients with metallic hip implants. Qualitative evaluations were performed by two blinded experienced radiologists who ranked overall artifact severity and pelvic organ recognition for each algorithm by assigning scores from zero to five (zero indicating totally obscured organs with no structures identifiable and five indicating recognition with high confidence). RESULTS: Simulation studies revealed that 2D, NMAR, and MAPC techniques performed almost equally well in all regions. LI falls behind the other approaches in terms of reducing dark streaking artifacts as well as preserving unaffected regions (p < 0.05). Visual assessment of clinical datasets revealed the superiority of NMAR and MAPC in the evaluated pelvic organs and in terms of overall image quality. CONCLUSIONS: Overall, all methods, except LI, performed equally well in artifact-free regions. Considering both clinical and simulation studies, 2D, NMAR, and MAPC seem to outperform the other techniques.

Vision 20/20: Magnetic resonance imaging-guided attenuation correction in PET/MRI: Challenges, solutions, and opportunities.

Attenuation correction is an essential component of the long chain of data correction techniques required to achieve the full potential of quantitative positron emission tomography (PET) imaging. The development of combined PET/magnetic resonance imaging (MRI) systems mandated the widespread interest in developing novel strategies for deriving accurate attenuation maps with the aim to improve the quantitative accuracy of these emerging hybrid imaging systems. The attenuation map in PET/MRI should ideally be derived from anatomical MR images; however, MRI intensities reflect proton density and relaxation time properties of biological tissues rather than their electron density and photon attenuation properties. Therefore, in contrast to PET/computed tomography, there is a lack of standardized global mapping between the intensities of MRI signal and linear attenuation coefficients at 511 keV. Moreover, in standard MRI sequences, bones and lung tissues do not produce measurable signals owing to their low proton density and short transverse relaxation times. MR images are also inevitably subject to artifacts that degrade their quality, thus compromising their applicability for the task of attenuation correction in PET/MRI. MRI-guided attenuation correction strategies can be classified in three broad categories: (i) segmentation-based approaches, (ii) atlas-registration and machine learning methods, and (iii) emission/transmission-based approaches. This paper summarizes past and current state-of-the-art developments and latest advances in PET/MRI attenuation correction. The advantages and drawbacks of each approach for addressing the challenges of MR-based attenuation correction are comprehensively described. The opportunities brought by both MRI and PET imaging modalities for deriving accurate attenuation maps and improving PET quantification will be elaborated. Future prospects and potential clinical applications of these techniques and their integration in commercial systems will also be discussed.

Quantitative analysis of MRI-guided attenuation correction techniques in time-of-flight brain PET/MRI.

PURPOSE: In quantitative PET/MR imaging, attenuation correction (AC) of PET data is markedly challenged by the need of deriving accurate attenuation maps from MR images. A number of strategies have been developed for MRI-guided attenuation correction with different degrees of success. In this work, we compare the quantitative performance of three generic AC methods, including standard 3-class MR segmentation-based, advanced atlas-registration-based and emission-based approaches in the context of brain time-of-flight (TOF) PET/MRI. MATERIALS AND METHODS: Fourteen patients referred for diagnostic MRI and (18)F-FDG PET/CT brain scans were included in this comparative study. For each study, PET images were reconstructed using four different attenuation maps derived from CT-based AC (CTAC) serving as reference, standard 3-class MR-segmentation, atlas-registration and emission-based AC methods. To generate 3-class attenuation maps, T1-weighted MRI images were segmented into background air, fat and soft-tissue classes followed by assignment of constant linear attenuation coefficients of 0, 0.0864 and 0.0975 cm(-1) to each class, respectively. A robust atlas-registration based AC method was developed for pseudo-CT generation using local weighted fusion of atlases based on their morphological similarity to target MR images. Our recently proposed MRI-guided maximum likelihood reconstruction of activity and attenuation (MLAA) algorithm was employed to estimate the attenuation map from TOF emission data. The performance of the different AC algorithms in terms of prediction of bones and quantification of PET tracer uptake was objectively evaluated with respect to reference CTAC maps and CTAC-PET images. RESULTS: Qualitative evaluation showed that the MLAA-AC method could sparsely estimate bones and accurately differentiate them from air cavities. It was found that the atlas-AC method can accurately predict bones with variable errors in defining air cavities. Quantitative assessment of bone extraction accuracy based on Dice similarity coefficient (DSC) showed that MLAA-AC and atlas-AC resulted in DSC mean values of 0.79 and 0.92, respectively, in all patients. The MLAA-AC and atlas-AC methods predicted mean linear attenuation coefficients of 0.107 and 0.134 cm(-1), respectively, for the skull compared to reference CTAC mean value of 0.138cm(-1). The evaluation of the relative change in tracer uptake within 32 distinct regions of the brain with respect to CTAC PET images showed that the 3-class MRAC, MLAA-AC and atlas-AC methods resulted in quantification errors of -16.2 ± 3.6%, -13.3 ± 3.3% and 1.0 ± 3.4%, respectively. Linear regression and Bland-Altman concordance plots showed that both 3-class MRAC and MLAA-AC methods result in a significant systematic bias in PET tracer uptake, while the atlas-AC method results in a negligible bias. CONCLUSION: The standard 3-class MRAC method significantly underestimated cerebral PET tracer uptake. While current state-of-the-art MLAA-AC methods look promising, they were unable to noticeably reduce quantification errors in the context of brain imaging. Conversely, the proposed atlas-AC method provided the most accurate attenuation maps, and thus the lowest quantification bias.

Accelerated time-of-flight (TOF) PET image reconstruction using TOF bin subsetization and TOF weighting matrix pre-computation.

Time-of-flight (TOF) positron emission tomography (PET) technology has recently regained popularity in clinical PET studies for improving image quality and lesion detectability. Using TOF information, the spatial location of annihilation events is confined to a number of image voxels along each line of response, thereby the cross-dependencies of image voxels are reduced, which in turns results in improved signal-to-noise ratio and convergence rate. In this work, we propose a novel approach to further improve the convergence of the expectation maximization (EM)-based TOF PET image reconstruction algorithm through subsetization of emission data over TOF bins as well as azimuthal bins. Given the prevalence of TOF PET, we elaborated the practical and efficient implementation of TOF PET image reconstruction through the pre-computation of TOF weighting coefficients while exploiting the same in-plane and axial symmetries used in pre-computation of geometric system matrix. In the proposed subsetization approach, TOF PET data were partitioned into a number of interleaved TOF subsets, with the aim of reducing the spatial coupling of TOF bins and therefore to improve the convergence of the standard maximum likelihood expectation maximization (MLEM) and ordered subsets EM (OSEM) algorithms. The comparison of on-the-fly and pre-computed TOF projections showed that the pre-computation of the TOF weighting coefficients can considerably reduce the computation time of TOF PET image reconstruction. The convergence rate and bias-variance performance of the proposed TOF subsetization scheme were evaluated using simulated, experimental phantom and clinical studies. Simulations demonstrated that as the number of TOF subsets is increased, the convergence rate of MLEM and OSEM algorithms is improved. It was also found that for the same computation time, the proposed subsetization gives rise to further convergence. The bias-variance analysis of the experimental NEMA phantom and a clinical FDG-PET study also revealed that for the same noise level, a higher contrast recovery can be obtained by increasing the number of TOF subsets. It can be concluded that the proposed TOF weighting matrix pre-computation and subsetization approaches enable to further accelerate and improve the convergence properties of OSEM and MLEM algorithms, thus opening new avenues for accelerated TOF PET image reconstruction.

Emission-based estimation of lung attenuation coefficients for attenuation correction in time-of-flight PET/MR.

In standard segmentation-based MRI-guided attenuation correction (MRAC) of PET data on hybrid PET/MRI systems, the inter/intra-patient variability of linear attenuation coefficients (LACs) is ignored owing to the assignment of a constant LAC to each tissue class. This can lead to PET quantification errors, especially in the lung regions. In this work, we aim to derive continuous and patient-specific lung LACs from time-of-flight (TOF) PET emission data using the maximum likelihood reconstruction of activity and attenuation (MLAA) algorithm. The MLAA algorithm was constrained for estimation of lung LACs only in the standard 4-class MR attenuation map using Gaussian lung tissue preference and Markov random field smoothness priors. MRAC maps were derived from segmentation of CT images of 19 TOF-PET/CT clinical studies into background air, lung, soft tissue and fat tissue classes, followed by assignment of predefined LACs of 0, 0.0224, 0.0864 and 0.0975 cm(-1), respectively. The lung LACs of the resulting attenuation maps were then estimated from emission data using the proposed MLAA algorithm. PET quantification accuracy of MRAC and MLAA methods was evaluated against the reference CT-based AC method in the lungs, lesions located in/near the lungs and neighbouring tissues. The results show that the proposed MLAA algorithm is capable of retrieving lung density gradients and compensate fairly for respiratory-phase mismatch between PET and corresponding attenuation maps. It was found that the mean of the estimated lung LACs generally follow the trend of the reference CT-based attenuation correction (CTAC) method. Quantitative analysis revealed that the MRAC method resulted in average relative errors of -5.2 ± 7.1% and -6.1 ± 6.7% in the lungs and lesions, respectively. These were reduced by the MLAA algorithm to -0.8 ± 6.3% and -3.3 ± 4.7%, respectively. In conclusion, we demonstrated the potential and capability of emission-based methods in deriving patient-specific lung LACs to improve the accuracy of attenuation correction in TOF PET/MR imaging, thus paving the way for their adaptation in the clinic.

Clinical Assessment of Emission- and Segmentation-Based MR-Guided Attenuation Correction in Whole-Body Time-of-Flight PET/MR Imaging.

UNLABELLED: The joint maximum-likelihood reconstruction of activity and attenuation (MLAA) for emission-based attenuation correction has regained attention since the advent of time-of-flight PET/MR imaging. Recently, we improved the performance of the MLAA algorithm using an MR imaging-constrained gaussian mixture model (GMM). In this study, we compare the performance of our proposed algorithm with standard 4-class MR-based attenuation correction (MRAC) implemented on commercial systems. METHODS: Five head and neck (18)F-FDG patients were scanned on PET/MR imaging and PET/CT scanners. Dixon fat and water MR images were registered to CT images. MRAC maps were derived by segmenting the MR images into 4 tissue classes and assigning predefined attenuation coefficients. For MLAA-GMM, MR images were segmented into known tissue classes, including fat, soft tissue, lung, background air, and an unknown MR low-intensity class encompassing cortical bones, air cavities, and metal artifacts. A coregistered bone probability map was also included in the unknown tissue class. Finally, the GMM prior was constrained over known tissue classes of attenuation maps using unimodal gaussians parameterized over a patient population. RESULTS: The results showed that the MLAA-GMM algorithm outperformed the MRAC method by differentiating bones from air gaps and providing more accurate patient-specific attenuation coefficients of soft tissue and lungs. It was found that the MRAC and MLAA-GMM methods resulted in average standardized uptake value errors of -5.4% and -3.5% in the lungs, -7.4% and -5.0% in soft tissues/lesions, and -18.4% and -10.2% in bones, respectively. CONCLUSION: The proposed MLAA algorithm is promising for accurate derivation of attenuation maps on time-of-flight PET/MR systems.

Impact of time-of-flight PET on quantification errors in MR imaging-based attenuation correction.

UNLABELLED: Time-of-flight (TOF) PET/MR imaging is an emerging imaging technology with great capabilities offered by TOF to improve image quality and lesion detectability. We assessed, for the first time, the impact of TOF image reconstruction on PET quantification errors induced by MR imaging-based attenuation correction (MRAC) using simulation and clinical PET/CT studies. METHODS: Standard 4-class attenuation maps were derived by segmentation of CT images of 27 patients undergoing PET/CT examinations into background air, lung, soft-tissue, and fat tissue classes, followed by the assignment of predefined attenuation coefficients to each class. For each patient, 4 PET images were reconstructed: non-TOF and TOF both corrected for attenuation using reference CT-based attenuation correction and the resulting 4-class MRAC maps. The relative errors between non-TOF and TOF MRAC reconstructions were compared with their reference CT-based attenuation correction reconstructions. The bias was locally and globally evaluated using volumes of interest (VOIs) defined on lesions and normal tissues and CT-derived tissue classes containing all voxels in a given tissue, respectively. The impact of TOF on reducing the errors induced by metal-susceptibility and respiratory-phase mismatch artifacts was also evaluated using clinical and simulation studies. RESULTS: Our results show that TOF PET can remarkably reduce attenuation correction artifacts and quantification errors in the lungs and bone tissues. Using classwise analysis, it was found that the non-TOF MRAC method results in an error of -3.4% ± 11.5% in the lungs and -21.8% ± 2.9% in bones, whereas its TOF counterpart reduced the errors to -2.9% ± 7.1% and -15.3% ± 2.3%, respectively. The VOI-based analysis revealed that the non-TOF and TOF methods resulted in an average overestimation of 7.5% and 3.9% in or near lung lesions (n = 23) and underestimation of less than 5% for soft tissue and in or near bone lesions (n = 91). Simulation results showed that as TOF resolution improves, artifacts and quantification errors are substantially reduced. CONCLUSION: TOF PET substantially reduces artifacts and improves significantly the quantitative accuracy of standard MRAC methods. Therefore, MRAC should be less of a concern on future TOF PET/MR scanners with improved timing resolution.

Joint Estimation of Activity and Attenuation in Whole-Body TOF PET/MRI Using Constrained Gaussian Mixture Models.

It has recently been shown that the attenuation map can be estimated from time-of-flight (TOF) PET emission data using joint maximum likelihood reconstruction of attenuation and activity (MLAA). In this work, we propose a novel MRI-guided MLAA algorithm for emission-based attenuation correction in whole-body PET/MR imaging. The algorithm imposes MR spatial and CT statistical constraints on the MLAA estimation of attenuation maps using a constrained Gaussian mixture model (GMM) and a Markov random field smoothness prior. Dixon water and fat MR images were segmented into outside air, lung, fat and soft-tissue classes and an MR low-intensity (unknown) class corresponding to air cavities, cortical bone and susceptibility artifacts. The attenuation coefficients over the unknown class were estimated using a mixture of four Gaussians, and those over the known tissue classes using unimodal Gaussians, parameterized over a patient population. To eliminate misclassification of spongy bones with surrounding tissues, and thus include them in the unknown class, we heuristically suppressed fat in water images and also used a co-registered bone probability map. The proposed MLAA-GMM algorithm was compared with the MLAA algorithms proposed by Rezaei and Salomon using simulation and clinical studies with two different tracer distributions. The results showed that our proposed algorithm outperforms its counterparts in suppressing the cross-talk and scaling problems of activity and attenuation and thus produces PET images of improved quantitative accuracy. It can be concluded that the proposed algorithm effectively exploits the MR information and can pave the way toward accurate emission-based attenuation correction in TOF PET/MRI.