Correction of multiplexing artefacts in multi-pinhole SPECT through temporal shuttering, de-multiplexing of projections, and alternating reconstruction.

Objective.Single-photon emission computed tomography (SPECT) with pinhole collimators can provide high-resolution imaging, but is often limited by low sensitivity. Acquiring projections simultaneously through multiple pinholes affords both high resolution and high sensitivity. However, the overlap of projections from different pinholes on detectors, known as multiplexing, has been shown to cause artefacts which degrade reconstructed images.Approach.Multiplexed projection sets were considered here using an analytic simulation model of AdaptiSPECT-C-a brain-dedicated multi-pinhole SPECT system. AdaptiSPECT-C has fully adaptable aperture shutters, so can acquire projections with a combination of multiplexed and non-multiplexed frames using temporal shuttering. Two strategies for reducing multiplex artefacts were considered: an algorithm to de-multiplex projections, and an alternating reconstruction strategy for projections acquired with a combination of multiplexed and non-multiplexed frames. Geometric and anthropomorphic digital phantoms were used to assess a number of metrics.Main results.Both de-multiplexing strategies showed a significant reduction in image artefacts and improved fidelity, image uniformity, contrast recovery and activity recovery (AR). In all cases, the two de-multiplexing strategies resulted in superior metrics to those from images acquired with only mux-free frames. The de-multiplexing algorithm provided reduced image noise and superior uniformity, whereas the alternating strategy improved contrast and AR.Significance.The use of these de-multiplexing algorithms means that multi-pinhole SPECT systems can acquire projections with more multiplexing without degradation of images.

Respiratory signal estimation for cardiac perfusion SPECT using deep learning.

BACKGROUND: Respiratory motion induces artifacts in reconstructed cardiac perfusion SPECT images. Correction for respiratory motion often relies on a respiratory signal describing the heart displacements during breathing. However, using external tracking devices to estimate respiratory signals can add cost and operational complications in a clinical setting. PURPOSE: We aim to develop a deep learning (DL) approach that uses only SPECT projection data for respiratory signal estimation. METHODS: A modified U-Net was implemented that takes temporally finely sampled SPECT sub-projection data (100 ms) as input. These sub-projections are obtained by reframing the 20-s list-mode data, resulting in 200 sub-projections, at each projection angle for each SPECT camera head. The network outputs a 200-time-point motion signal for each projection angle, which was later aggregated over all angles to give a full respiratory signal. The target signal for DL model training was from an external stereo-camera visual tracking system (VTS). In addition to comparing DL and VTS, we also included a data-driven approach based on the center-of-mass (CoM) strategy. This CoM method estimates respiratory signals by monitoring the axial changes of CoM for counts in the heart region of the sub-projections. We utilized 900 subjects with stress cardiac perfusion SPECT studies, with 302 subjects for testing and the remaining 598 subjects for training and validation. RESULTS: The Pearson's correlation coefficient between the DL respiratory signal and the reference VTS signal was 0.90, compared to 0.70 between the CoM signal and the reference. For respiratory motion correction on SPECT images, all VTS, DL, and CoM approaches partially de-blured the heart wall, resulting in a thinner wall thickness and increased recovered maximal image intensity within the wall, with VTS reducing blurring the most followed by the DL approach. Uptake quantification for the combined anterior and inferior segments of polar maps showed a mean absolute difference from the reference VTS of 1.7% for the DL method for patients with motion >12 mm, compared to 2.6% for the CoM method and 8.5% for no correction. CONCLUSION: We demonstrate the capability of a DL approach to estimate respiratory signal from SPECT projection data for cardiac perfusion imaging. Our results show that the DL based respiratory motion correction reduces artefacts and achieves similar regional quantification to that obtained using the stereo-camera VTS signals. This may enable fully automatic data-driven respiratory motion correction without relying on external motion tracking devices.

Observer studies of image quality of denoising reduced-count cardiac single photon emission computed tomography myocardial perfusion imaging by three-dimensional Gaussian post-reconstruction filtering and deep learning.

BACKGROUND: The aim of this research was to asses perfusion-defect detection-accuracy by human observers as a function of reduced-counts for 3D Gaussian post-reconstruction filtering vs deep learning (DL) denoising to determine if there was improved performance with DL. METHODS: SPECT projection data of 156 normally interpreted patients were used for these studies. Half were altered to include hybrid perfusion defects with defect presence and location known. Ordered-subset expectation-maximization (OSEM) reconstruction was employed with the optional correction of attenuation (AC) and scatter (SC) in addition to distance-dependent resolution (RC). Count levels varied from full-counts (100%) to 6.25% of full-counts. The denoising strategies were previously optimized for defect detection using total perfusion deficit (TPD). Four medical physicist (PhD) and six physician (MD) observers rated the slices using a graphical user interface. Observer ratings were analyzed using the LABMRMC multi-reader, multi-case receiver-operating-characteristic (ROC) software to calculate and compare statistically the area-under-the-ROC-curves (AUCs). RESULTS: For the same count-level no statistically significant increase in AUCs for DL over Gaussian denoising was determined when counts were reduced to either the 25% or 12.5% of full-counts. The average AUC for full-count OSEM with solely RC and Gaussian filtering was lower than for the strategies with AC and SC, except for a reduction to 6.25% of full-counts, thus verifying the utility of employing AC and SC with RC. CONCLUSION: We did not find any indication that at the dose levels investigated and with the DL network employed, that DL denoising was superior in AUC to optimized 3D post-reconstruction Gaussian filtering.

Mesh modeling of system geometry and anatomy phantoms for realistic GATE simulations and their inclusion in SPECT reconstruction.

Objective.Monte-Carlo simulation studies have been essential for advancing various developments in single photon emission computed tomography (SPECT) imaging, such as system design and accurate image reconstruction. Among the simulation software available, Geant4 application for tomographic emission (GATE) is one of the most used simulation toolkits in nuclear medicine, which allows building systems and attenuation phantom geometries based on the combination of idealized volumes. However, these idealized volumes are inadequate for modeling free-form shape components of such geometries. Recent GATE versions alleviate these major limitations by allowing users to import triangulated surface meshes.Approach.In this study, we describe our mesh-based simulations of a next-generation multi-pinhole SPECT system dedicated to clinical brain imaging, called AdaptiSPECT-C. To simulate realistic imaging data, we incorporated in our simulation the XCAT phantom, which provides an advanced anatomical description of the human body. An additional challenge with the AdaptiSPECT-C geometry is that the default voxelized XCAT attenuation phantom was not usable in our simulation due to intersection of objects of dissimilar materials caused by overlap of the air containing regions of the XCAT beyond the surface of the phantom and the components of the imaging system.Main results.We validated our mesh-based modeling against the one constructed by idealized volumes for a simplified single vertex configuration of AdaptiSPECT-C through simulated projection data of123I-activity distributions. We resolved the overlap conflict by creating and incorporating a mesh-based attenuation phantom following a volume hierarchy. We then evaluated our reconstructions with attenuation and scatter correction for projections obtained from simulation consisting of mesh-based modeling of the system and the attenuation phantom for brain imaging. Our approach demonstrated similar performance as the reference scheme simulated in air for uniform and clinical-like123I-IMP brain perfusion source distributions.Significance.This work enables the simulation of complex SPECT acquisitions and reconstructions for emulating realistic imaging data close to those of actual patients.

A rapid in vitro assay for evaluating the effects of acetylcholinesterase inhibitors and reactivators in the rat basolateral amygdala.

We established a novel brain slice assay to test the ability of acetylcholinesterase (AChE) reactivators to prevent ACh-induced M1 muscarinic acetylcholine receptor (mAChR) dependent hyperexcitability observed after exposure to the organophosphate (OP)-based AChE inhibitor and sarin surrogate 4-nitrophenyl isopropyl methylphosphonate (NIMP). Whole-cell patch clamp recordings were used to evaluate the response of pyramidal neurons in the rat basolateral amygdala (BLA) to brief (1 min) bath application of ACh (100 µM), either in control conditions, or after exposure to NIMP ± an AChE reactivator. Bath application of ACh produced atropine- and pirenzepine-sensitive inward currents in voltage clamped BLA pyramidal neurons, and increased the frequency of spontaneous EPSCs, suggesting robust activation of M1 mAChRs. Responses to ACh were increased ~3-5 fold in slices that had been preincubated in NIMP, and these effects were reversed in a concentration dependent manner by exposure to a commercially available AChE reactivator. The current work outlines a simple assay that can be used to evaluate the efficacy of both known and novel AChE reactivators in an area of the limbic system that likely contributes to seizures after acute exposure to OP-based AChE inhibitors.

Dual gating myocardial perfusion SPECT denoising using a conditional generative adversarial network.

PURPOSE: Dual respiratory-cardiac gating reduces respiratory and cardiac motion blur in myocardial perfusion single-photon emission computed tomography (MP-SPECT). However, image noise is increased as detected counts are reduced in each dual gate (DG). We aim to develop a denoising method for dual gating MP-SPECT images using a 3D conditional generative adversarial network (cGAN). METHODS: Twenty extended cardiac-torso phantoms with various 99m Tc-sestamibi distributions, defect characteristics, and body and organ sizes were used in the simulation, modeling six respiratory and eight cardiac gates (CGs), that is, 48 DGs for ordered subset expectation maximization reconstruction. Twenty clinical 99m Tc-sestamibi SPECT/CT datasets were re-binned into 7 respiratory gates and 8 CGs, that is, 56 DGs for maximum likelihood expectation maximization reconstruction. We evaluated the use of (i) phantoms' own datasets (patient-specific denoising [PD]) or other phantoms' datasets (cross-patient denoising) for training; (ii) the CG or the static (non-gated [NG]) data as the training references for cGAN; and (iii) cGAN as compared to conventional 3D post-reconstruction filtering, cardiac gating methods, and convolutional neural network. Normalized mean squared error, noise as assessed by normalized standard deviation, spatial blurring measured as the full-width-at-half-maximum of left ventricular wall, ejection fraction, joint correlation histogram, and defect size were analyzed as metrics of image quality. RESULTS: Training using patients' own dataset is superior to conventional training based on other patients' data. Using CG image as training reference provides a better trade-off in terms of noise and image blur as compared to the use of NG. cGAN-CG-PD provides superior performance as compared to other denoising methods for all physical and diagnostic indices evaluated in both simulation and clinical studies. CONCLUSIONS: cGAN denoising is promising for dual gating MP-SPECT based on the metrics mentioned earlier.

Cross-vender, cross-tracer, and cross-protocol deep transfer learning for attenuation map generation of cardiac SPECT.

It has been proved feasible to generate attenuation maps (µ-maps) from cardiac SPECT using deep learning. However, this assumed that the training and testing datasets were acquired using the same scanner, tracer, and protocol. We investigated a robust generation of CT-derived µ-maps from cardiac SPECT acquired by different scanners, tracers, and protocols from the training data. We first pre-trained a network using 120 studies injected with 99mTc-tetrofosmin acquired from a GE 850 SPECT/CT with 360-degree gantry rotation, which was then fine-tuned and tested using 80 studies injected with 99mTc-sestamibi acquired from a Philips BrightView SPECT/CT with 180-degree gantry rotation. The error between ground-truth and predicted µ-maps by transfer learning was 5.13 ± 7.02%, as compared to 8.24 ± 5.01% by direct transition without fine-tuning and 6.45 ± 5.75% by limited-sample training. The error between ground-truth and reconstructed images with predicted µ-maps by transfer learning was 1.11 ± 1.57%, as compared to 1.72 ± 1.63% by direct transition and 1.68 ± 1.21% by limited-sample training. It is feasible to apply a network pre-trained by a large amount of data from one scanner to data acquired by another scanner using different tracers and protocols, with proper transfer learning.

Data-driven respiratory signal estimation from temporally finely sampled projection data in conventional cardiac perfusion SPECT imaging.

PURPOSE: The aim of this work was to revisit the data-driven approach of axial center-of-mass (COM) measurements to recover a surrogate respiratory signal from finely sampled (100 ms) single photon emission computed tomography (SPECT) projection data derived from list-mode acquisitions. METHODS: For our initial evaluation, we acquired list-mode projection data from an anthropomorphic cardiac phantom mounted on a Quasar respiratory motion platform simulating 15 mm amplitude respiratory motion. We also selected 302 consecutive patients (138 males, 164 females) with list-mode acquisitions, external respiratory motion tracking, and written consent to evaluate the clinical efficacy of our data-driven approach. Linear regression, Pearson's correlation coefficient (r), and standard error of the estimates (SEE) between the respiratory signals obtained with a visual tracking system (VTS) and COM measurements were calculated for individual projection data sets and for the patient group as a whole. Both the VTS- and COM-derived respiratory signals were used to estimate and correct respiratory motion. The reconstruction for six-degree of freedom rigid-body motion estimation was done in two ways: (1) using three iterations of ordered-subsets expectation-maximization (OSEM) with four subsets (16 projection angles per subset), or 12 iterations of maximum-likelihood expectation-maximization (MLEM). Respiratory motion compensation was done employing either OSEM with 16 subsets (four projection angles per subset) and five iterations or MLEM and 80 iterations, using the two respiratory estimates, respectively. Polar map quantification was also performed, calculating the percentage count difference (%Diff) between polar maps without and with respiratory motion included. Average % Diff was calculated in 17 segments (defined according to ASNC Guidelines). Paired t-tests were used to determine significance (p-values). RESULTS: The r-value calculated when comparing the VTS and COM respiratory signals varied widely between -0.01 and 0.96 with an average of 0.70, while the SEE varied between 0.80 and 6.48 mm with an average of 2.05 mm for our patient set, while the same values for the one anthropomorphic phantom acquisition are 0.91 and 1.11 mm, respectively. A comparison between the respiratory motion estimates for VTS and COM in the S-I direction yielded an r = 0.90 (0.94), and an SEE of 1.56 mm (1.20 mm) for OSEM (MLEM), respectively. Bland-Altman plots and calculated intraclass correlation coefficients also showed excellent agreement between the VTS and COM respiratory motion estimates. Average S-I respiratory estimates for the VTS (COM) were 9.04 (9.2 mm) and 9.01 mm (9.14 mm) for the OSEM and MLEM, respectively. The paired t-test approached significance when comparing VTS and COM estimated respiratory signals with p-values of 0.069 and 0.051 for OSEM and MLEM. The respiratory estimates from the anthropomorphic cardiac phantom experiment using the VTS (COM) were 12.62 (14.10 mm) and 12.55 mm (14.29 mm) for OSEM and MLEM, respectively. Polar map quantification yielded average % Diff consistently better when employing VTS-derived respiratory estimates to correct for respiration compared to the COM-derived estimates. CONCLUSIONS: The results indicate that our COM method has the potential to provide an automated data-driven correction of cardiac respiratory motion without the drawbacks of our VTS methodology. However, it is not generally equivalent to the VTS method in extent of correction.

Improving detection accuracy of perfusion defect in standard dose SPECT-myocardial perfusion imaging by deep-learning denoising.

BACKGROUND: We previously developed a deep-learning (DL) network for image denoising in SPECT-myocardial perfusion imaging (MPI). Here we investigate whether this DL network can be utilized for improving detection of perfusion defects in standard-dose clinical acquisitions. METHODS: To quantify perfusion-defect detection accuracy, we conducted a receiver-operating characteristic (ROC) analysis on reconstructed images with and without processing by the DL network using a set of clinical SPECT-MPI data from 190 subjects. For perfusion-defect detection hybrid studies were used as ground truth, which were created from clinically normal studies with simulated realistic lesions inserted. We considered ordered-subset expectation-maximization (OSEM) reconstruction with corrections for attenuation, resolution, and scatter and with 3D Gaussian post-filtering. Total perfusion deficit (TPD) scores, computed by Quantitative Perfusion SPECT (QPS) software, were used to evaluate the reconstructed images. RESULTS: Compared to reconstruction with optimal Gaussian post-filtering (sigma = 1.2 voxels), further DL denoising increased the area under the ROC curve (AUC) from 0.80 to 0.88 (P-value < 10-4). For reconstruction with less Gaussian post-filtering (sigma = 0.8 voxels), thus better spatial resolution, DL denoising increased the AUC value from 0.78 to 0.86 (P-value < 10-4) and achieved better spatial resolution in reconstruction. CONCLUSIONS: DL denoising can effectively improve the detection of abnormal defects in standard-dose SPECT-MPI images over conventional reconstruction.

Improved Performance of a Multipinhole SPECT for DAT Imaging by Increasing Number of Pinholes at the Expense of Increased Multiplexing.

SPECT imaging of dopamine transporters (DAT) in the brain is a widely utilized study to improve the diagnosis of Parkinsonian syndromes, where conventional (parallel-hole and fan-beam) collimators on dual-head scanners are commonly employed. We have designed a multi-pinhole (MPH) collimator to improve the performance of DAT imaging. The MPH collimator focuses on the striatum and hence offers a better trade-off for sensitivity and spatial resolution than the conventional collimators within this clinically most relevant region for DAT imaging. Our original MPH design consisted of 9 pinholes with a background-to-striatal (Bkg/Str) projection multiplexing of 1% only. In this simulation study, we investigated whether further improvements in the performance of MPH imaging could be obtained by increasing the number of pinholes, hence by enhancing the sensitivity and sampling, despite the ambiguity in reconstructing images due to increased multiplexing. We performed analytic simulations of the MPH configurations with 9, 13, and 16 pinholes (aperture diameters: 4-6mm) using a digital phantom modeling DAT imaging. Our quantitative analyses indicated that using 13 (Bkg/Str: 12%) and 16 (Bkg/Str: 22%) pinholes provided better performance than the original 9-pinhole configuration for the acquisition with 2 or 4 angular views, but a similar performance with 8 and 16 views.

Cerebral SPECT imaging with different acquisition schemes using varying levels of multiplexing versus sensitivity in an adaptive multi-pinhole brain-dedicated scanner.

Application of multi-pinhole collimator in pinhole-based SPECT increases detection sensitivity. The presence of multiplexing in projection images due to the usage of multiple pinholes can further improve the sensitivity at the cost of adding data ambiguity. We are developing a next-generation adaptive brain-dedicated SPECT system -AdaptiSPECT-C. The AdaptiSPECT-C can adapt the multiplexing level and system sensitivity using adaptable pinhole modules. In this study, we investigated the performance of 4 data acquisition schemes with different multiplexing levels and sensitivities on cerebral SPECT imaging. Schemes #1, #2, and #3 have <1%, 67%, and 31% overall multiplexing, respectively, while the 4th scheme without multiplexing is considered as ground truth. The ground-truth and schemes #1-3 have 1.0, 1.7, 5.1, and 4.0 times higher sensitivity, respectively, compared to a dual-headed parallel-hole SPECT system at matched spatial resolution. A customized XCAT brain perfusion digital phantom emulating the distribution of I-123 N-isopropyl iodoamphetamine (IMP) in a 99th percentile size male was used for simulations. Data acquisition for each scheme was performed at two count levels (low-count and high-count relative to the recommended clinical count level). The normalized root-mean-square error (NRMSE) for schemes #1, #2, and #3 with the low-count (high-count) scenario showed 11%, 4%, and 5% (10%, 5%, and 6%) deviation, respectively, from that of the multiplex-free ground truth. For both the low-count and high-count scenarios, scheme #1 resulted in the least accurate activity ratio (AR) for almost all the analyzed gray-matter brain regions. Further schemes #2 or #3 led to the most accurate AR values with both low-count and high-count scenarios for all the analyzed gray-matter regions. It was thus observed that even with this large head size which leads to significant multiplexing levels, the higher sensitivity from multiplexing could to some extent mitigate the data ambiguity and be translated into reconstructed images of higher quality.

A novel deep-learning-based approach for automatic reorientation of 3D cardiac SPECT images.

PURPOSE: Reconstructed transaxial cardiac SPECT images need to be reoriented into standard short-axis slices for subsequent accurate processing and analysis. We proposed a novel deep-learning-based method for fully automatic reorientation of cardiac SPECT images and evaluated its performance on data from two clinical centers. METHODS: We used a convolutional neural network to predict the 6 rigid-body transformation parameters and a spatial transformation network was then implemented to apply these parameters on the input images for image reorientation. A novel compound loss function which balanced the parametric similarity and penalized discrepancy of the prediction and training dataset was utilized in the training stage. Data from a set of 322 patients underwent data augmentation to 6440 groups of images for the network training, and a dataset of 52 patients from the same center and 23 patients from another center were used for evaluation. Similarity of the 6 parameters was analyzed between the proposed and the manual methods. Polar maps were generated from the output images and the averaged count values of the 17 segments were computed from polar maps to evaluate the quantitative accuracy of the proposed method. RESULTS: All the testing patients achieved automatic reorientation successfully. Linear regression results showed the 6 predicted rigid parameters and the average count value of the 17 segments having good agreement with the reference manual method. No significant difference by paired t-test was noticed between the rigid parameters of our method and the manual method (p > 0.05). Average count values of the 17 segments show a smaller difference of the proposed and manual methods than those between the existing and manual methods. CONCLUSION: The results strongly indicate the feasibility of our method in accurate automatic cardiac SPECT reorientation. This deep-learning-based reorientation method has great promise for clinical application and warrants further investigation.

Improvement in sampling and modulation of multiplexing with temporal shuttering of adaptable apertures in a brain-dedicated multi-pinhole SPECT system.

We are developing a multi-detector pinhole-based stationary brain-dedicated SPECT system: AdaptiSPECT-C. In this work, we introduced a new design prototype with multiple adaptable pinhole apertures for each detector to modulate the multiplexing by employing temporal shuttering of apertures. Temporal shuttering of apertures over the scan time provides the AdaptiSPECT-C with the capability of multiple-frame acquisition. We investigated, through analytic simulation, the impact of projection multiplexing on image quality using several digital phantoms and a customized anthropomorphic phantom emulating brain perfusion clinical distribution. The 105 pinholes in the collimator of the system were categorized into central, axial, and lateral apertures. We generated, through simulation, collimators of different multiplexing levels. Several data acquisition schemes were also created by changing the imaging time share of the acquisition frames. Sensitivity increased by 35% compared to the single-pinhole-per-detector base configuration of the AdaptiSPECT-C when using the central, axial, and lateral apertures with equal acquisition time shares within a triple-frame scheme with a high multiplexing scenario. Axial and angular sampling of the base configuration was enhanced by adding the axial and lateral apertures. We showed that the temporal shuttering of apertures can be exploited, trading the sensitivity, to modulate the multiplexing and to acquire a set of non-multiplexed non-truncated projections. Our results suggested that reconstruction benefited from utilizing both non-multiplexed projections and projections with modulated multiplexing resulting in a noticeably reduction in the multiplexing-induced image artefacts. Contrast recovery factor improved by 20% (9%) compared to the base configuration for a Defrise (hot-rod) phantom study when the central and axial (lateral) apertures with equal time shares were combined. The results revealed that, as an overall trend at each simulated multiplexing level, lowest normalized root-mean-square errors for the brain gray-matter regions were achieved with the combined usage of the central apertures and axial/lateral apertures.

Retrospective fractional dose reduction in Tc-99m cardiac perfusion SPECT/CT patients: A human and model observer study.

BACKGROUND: In the ongoing efforts to reduce cardiac perfusion dose (injected radioactivity) for conventional SPECT/CT systems, we performed a human observer study to confirm our clinical model observer findings that iterative reconstruction employing OSEM (ordered-subset expectation-maximization) at 25% of the full dose (quarter-dose) has a similar performance for detection of hybrid cardiac perfusion defects as FBP at full dose. METHODS: One hundred and sixty-six patients, who underwent routine rest-stress Tc-99m sestamibi cardiac perfusion SPECT/CT imaging and clinically read as normally perfused, were included in the study. Ground truth was established by the normal read and the insertion of hybrid defects. In addition to the reconstruction of the 25% of full-dose data using OSEM with attenuation (AC), scatter (SC), and spatial resolution correction (RC), FBP and OSEM (with AC, SC, and RC) both at full dose (100%) were done. Both human observer and clinical model observer confidence scores were obtained to generate receiver operating characteristics (ROC) curves in a task-based image quality assessment. RESULTS: Average human observer AUC (area under the ROC curve) values of 0.725, 0.876, and 0.890 were obtained for FBP at full dose, OSEM at 25% of full dose, and OSEM at full dose, respectively. Both OSEM strategies were significantly better than FBP with P values of 0.003 and 0.01 respectively, while no significant difference was recorded between OSEM methods (P = 0.48). The clinical model observer results were 0.791, 0.822, and 0.879, respectively, for the same patient cases and processing strategies used in the human observer study. CONCLUSIONS: Cardiac perfusion SPECT/CT using OSEM reconstruction at 25% of full dose has AUCs larger than FBP and closer to those of full-dose OSEM when read by human observers, potentially replacing the higher dose studies during clinical reading.

Deep learning with noise-to-noise training for denoising in SPECT myocardial perfusion imaging.

PURPOSE: Post-reconstruction filtering is often applied for noise suppression due to limited data counts in myocardial perfusion imaging (MPI) with single-photon emission computed tomography (SPECT). We study a deep learning (DL) approach for denoising in conventional SPECT-MPI acquisitions, and investigate whether it can be more effective for improving the detectability of perfusion defects compared to traditional postfiltering. METHODS: Owing to the lack of ground truth in clinical studies, we adopt a noise-to-noise (N2N) training approach for denoising in SPECT-MPI images. We consider a coupled U-Net (CU-Net) structure which is designed to improve learning efficiency through feature map reuse. For network training we employ a bootstrap procedure to generate multiple noise realizations from list-mode clinical acquisitions. In the experiments we demonstrated the proposed approach on a set of 895 clinical studies, where the iterative OSEM algorithm with three-dimensional (3D) Gaussian postfiltering was used to reconstruct the images. We investigated the detection performance of perfusion defects in the reconstructed images using the non-prewhitening matched filter (NPWMF), evaluated the uniformity of left ventricular (LV) wall in terms of image intensity, and quantified the effect of smoothing on the spatial resolution of the reconstructed LV wall by using its full-width at half-maximum (FWHM). RESULTS: Compared to OSEM with Gaussian postfiltering, the DL denoised images with CU-Net significantly improved the detection performance of perfusion defects at all contrast levels (65%, 50%, 35%, and 20%). The signal-to-noise ratio (SNRD ) in the NPWMF output was increased on average by 8% over optimal Gaussian smoothing (P < 10-4 , paired t-test), while the inter-subject variability was greatly reduced. The CU-Net also outperformed a 3D nonlocal means (NLM) filter and a convolutional autoencoder (CAE) denoising network in terms of SNRD . In addition, the FWHM of the LV wall in the reconstructed images was varied by less than 1%. Furthermore, CU-Net also improved the detection performance when the images were processed with less post-reconstruction smoothing (a trade-off of increased noise for better LV resolution), with SNRD improved on average by 23%. CONCLUSIONS: The proposed DL with N2N training approach can yield additional noise suppression in SPECT-MPI images over conventional postfiltering. For perfusion defect detection, DL with CU-Net could outperform conventional 3D Gaussian filtering with optimal setting as well as NLM and CAE.

Inclusion of quasi-vertex views in a brain-dedicated multi-pinhole SPECT system for improved imaging performance.

With brain-dedicated multi-detector systems employing pinhole apertures the usage of detectors facing the top of the patient's head (i.e. quasi-vertex (QV) views) can provide the advantage of additional viewing from close to the brain for improved detector coverage. In this paper, we report the results of simulation and reconstruction studies to investigate the impact of the QV views on the imaging performance of AdaptiSPECT-C, a brain-dedicated stationary SPECT system under development. In this design, both primary and scatter photons from regions located inferior to the brain can contribute to SPECT projections acquired by the QV views, and thus degrade AdaptiSPECT-C imaging performance. In this work, we determined the proportion, origin, and nature (i.e. primary, scatter, and multiple-scatter) of counts emitted from structures within the head and throughout the body contributing to projections from the different AdaptiSPECT-C detector rings, as well as from a true vertex view detector. We simulated phantoms used to assess different aspects of image quality (i.e. uniform activity concentration sphere, and Derenzo), as well as anthropomorphic phantoms with different count levels emulating clinical 123I activity distributions (i.e. DaTscan and perfusion). We determined that attenuation and scatter in the patient's body greatly diminish the probability of the photons emitted outside the volume of interest reaching to detectors and being recorded within the 15% photopeak energy window. In addition, we demonstrated that the inclusion of the residual of such counts in the system acquisition does not degrade visual interpretation or quantitative analysis. The addition of the QV detectors improves volumetric sensitivity, angular sampling, and spatial resolution leading to significant enhancement in image quality, especially in the striato-thalamic and superior regions of the brain. Besides, the use of QV detectors improves the recovery of clinically relevant metrics such as the striatal binding ratio and mean activity in selected cerebral structures. Our findings proving the usefulness of the QV ring for brain imaging with 123I agents can be generalized to other commonly used SPECT imaging agents labelled with isotopes, such as 99mTc and likely 111In.

The clinical utilities of multi-pinhole single photon emission computed tomography.

Single photon emission computed tomography (SPECT) is an important imaging modality for various applications in nuclear medicine. The use of multi-pinhole (MPH) collimators can provide superior resolution-sensitivity trade-off when imaging small field-of-view compared to conventional parallel-hole and fan-beam collimators. Besides the very successful application in small animal imaging, there has been a resurgence of the use of MPH collimators for clinical cardiac and brain studies, as well as other small field-of-view applications. This article reviews the basic principles of MPH collimators and introduces currently available and proposed clinical MPH SPECT systems.

Investigation of Axial and Angular Sampling in Multi-Detector Pinhole-SPECT Brain Imaging.

We designed a dedicated multi-detector multi-pinhole brain SPECT scanner to generate images of higher quality compared to general-purpose systems. The system, AdaptiSPECT-C, is intended to adapt its sensitivity-resolution trade-off by varying its aperture configurations allowing both high-sensitivity dynamic and high-spatial-resolution static imaging. The current system design consists of 23 detector heads arranged in a truncated spherical geometry. In this work, we investigated the axial and angular sampling capability of the current stationary system design. Two data acquisition schemes using limited rotation of the gantry and two others using axial translation of the imaging bed were also evaluated concerning their impact on image quality through improved sampling. Increasing both angular and axial sampling in the current prototype system resulted in quantitative improvements in image quality metrics and qualitative appearance of the images as determined in studies with specifically selected phantoms. Visual improvements for the brain phantoms with clinical distributions were less pronounced but presented quantitative improvements in the fidelity (normalized root-mean-square error (NRMSE)) and striatal specific binding ratio (SBR) for a dopamine transporter (DAT) distribution, and in NRMSE and activity recovery for a brain perfusion distribution. More pronounced improvements with increased sampling were seen in contrast recovery coefficient, bias, and coefficient of variation for a lesion in the brain perfusion distribution. The negligible impact of the most cranial ring of detectors on axial sampling, but its significant impact on sensitivity and angular sampling in the cranial portion of the imaging volume-of-interest were also determined.

Pharmacological targeting of striatal indirect pathway neurons improves subthalamic nucleus dysfunction and reduces repetitive behaviors in C58 mice.

Repetitive behaviors (e.g., stereotypic movements, compulsions, rituals) are common features of a number of neurodevelopmental disorders. Clinical and animal model studies point to the importance of cortical-basal ganglia circuitry in the mediation of repetitive behaviors. In the current study, we tested whether a drug cocktail (dopamine D2 receptor antagonist + adenosine A2A receptor agonist + glutamate mGlu5 positive allosteric modulator) designed to activate the indirect basal ganglia pathway would reduce repetitive behavior in C58 mice after both acute and sub-chronic administration. In addition, we hypothesized that sub-chronic administration (i.e. 7 days of twice-daily injections) would increase the functional activation of the subthalamic nucleus (STN), a key node of the indirect pathway. Functional activation of STN was indexed by dendritic spine density, analysis of GABA, glutamate, and synaptic plasticity genes, and cytochrome oxidase activity. The drug cocktail used significantly reduced repetitive motor behavior in C58 mice after one night as well as seven nights of twice-nightly injections. These effects did not reflect generalized motor behavior suppression as non-repetitive motor behaviors such as grooming, digging and eating were not reduced relative to vehicle. Sub-chronic drug treatment targeting striatopallidal neurons resulted in significant changes in the STN, including a four-fold increase in brain-derived neurotrophic factor (BDNF) mRNA expression as well as a significant increase in dendritic spine density. The present findings are consistent with, and extend, our prior work linking decreased functioning of the indirect basal ganglia pathway to expression of repetitive motor behavior in C58 mice and suggest novel therapeutic targets.

Improving Diagnostic Accuracy in Low-Dose SPECT Myocardial Perfusion Imaging With Convolutional Denoising Networks.

Lowering the administered dose in SPECT myocardial perfusion imaging (MPI) has become an important clinical problem. In this study we investigate the potential benefit of applying a deep learning (DL) approach for suppressing the elevated imaging noise in low-dose SPECT-MPI studies. We adopt a supervised learning approach to train a neural network by using image pairs obtained from full-dose (target) and low-dose (input) acquisitions of the same patients. In the experiments, we made use of acquisitions from 1,052 subjects and demonstrated the approach for two commonly used reconstruction methods in clinical SPECT-MPI: 1) filtered backprojection (FBP), and 2) ordered-subsets expectation-maximization (OSEM) with corrections for attenuation, scatter and resolution. We evaluated the DL output for the clinical task of perfusion-defect detection at a number of successively reduced dose levels (1/2, 1/4, 1/8, 1/16 of full dose). The results indicate that the proposed DL approach can achieve substantial noise reduction and lead to improvement in the diagnostic accuracy of low-dose data. In particular, at 1/2 dose, DL yielded an area-under-the-ROC-curve (AUC) of 0.799, which is nearly identical to the AUC = 0.801 obtained by OSEM at full-dose ( p -value = 0.73); similar results were also obtained for FBP reconstruction. Moreover, even at 1/8 dose, DL achieved AUC = 0.770 for OSEM, which is above the AUC = 0.755 obtained at full-dose by FBP. These results indicate that, compared to conventional reconstruction filtering, DL denoising can allow for additional dose reduction without sacrificing the diagnostic accuracy in SPECT-MPI.