Design of a digital phantom population for myocardial perfusion SPECT imaging research

Method for comparison of data driven gating algorithms in emission tomography

Objective.Multiple algorithms have been proposed for data driven gating (DDG) in single photon emission computed tomography (SPECT) and have successfully been applied to myocardial perfusion imaging (MPI). Application of DDG to acquisition types other than SPECT MPI has not been demonstrated so far, as limitations and pitfalls of current methods are unknown.Approach.We create a comprehensive set of phantoms simulating the influence of different motion artifacts, view angles, moving objects, contrast, and count levels in SPECT. We perform Monte Carlo simulation of the phantoms, allowing the characterization of DDG algorithms using quantitative metrics derived from the data and evaluate the Center of Light (COL) and Laplacian Eigenmaps methods as sample DDG algorithms.Main results.View angle, object size, count rate density, and contrast influence the accuracy of both DDG methods. Moreover, the ability to extract the respiratory motion in the phantom was shown to correlate with the contrast of the moving feature to the background, the signal to noise ratio, and the noise in the data.Significance.We showed that reporting the average correlation to an external physical reference signal per acquisition is not sufficient to characterize DDG methods. Assessing DDG methods on a view-by-view basis using the simulations and metrics from this work could enable the identification of pitfalls of current methods, and extend their application to acquisitions beyond SPECT MPI.

SPECT and CT misregistration reduction in [99mTc]Tc-MAA SPECT/CT for precision liver radioembolization treatment planning

PURPOSE

Respiration and body movement induce misregistration between static [99mTc]Tc-MAA SPECT and CT, causing lung shunting fraction (LSF) and tumor-to-normal liver ratio (TNR) errors for 90Y radioembolization planning. We aim to alleviate the misregistration between [99mTc]Tc-MAA SPECT and CT using two registration schemes on simulation and clinical data.

METHODS

In the simulation study, 70 XCAT phantoms were modeled. The SIMIND Monte Carlo program and OS-EM algorithm were used for projection generation and reconstruction, respectively. Low-dose CT (LDCT) at end-inspiration was simulated for attenuation correction (AC), lungs and liver segmentation, while contrast-enhanced CT (CECT) was simulated for tumor and perfused liver segmentation. In the clinical study, 16 patient data including [99mTc]Tc-MAA SPECT/LDCT and CECT with observed SPECT and CT mismatch were analyzed. Two liver-based registration schemes were studied: SPECT registered to LDCT/CECT and vice versa. Mean count density (MCD) of different volumes-of-interest (VOIs), normalized mutual information (NMI), LSF, TNR, and maximum injected activity (MIA) based on the partition model before and after registration were compared. Wilcoxon signed-rank test was performed.

RESULTS

In the simulation study, compared to before registration, registrations significantly reduced estimation errors of MCD of all VOIs, LSF (Scheme 1: - 100.28%, Scheme 2: - 101.59%), and TNR (Scheme 1: - 7.00%, Scheme 2: - 5.67%), as well as MIA (Scheme 1: - 3.22%, Scheme 2: - 2.40%). In the clinical study, Scheme 1 reduced 33.68% LSF and increased 14.75% TNR, while Scheme 2 reduced 38.88% LSF and increased 6.28% TNR compared to before registration. One patient may change from 90Y radioembolization untreatable to treatable and other patients may change the MIA up to 25% after registration. NMI between SPECT and CT was significantly increased after registrations in both studies.

CONCLUSION

Registration between static [99mTc]Tc-MAA SPECT and corresponding CTs is feasible to reduce their spatial mismatch and improve dosimetric estimation. The improvement of LSF is larger than TNR. Our method can potentially improve patient selection and personalized treatment planning for liver radioembolization.

Need for objective task-based evaluation of deep learning-based denoising methods: A study in the context of myocardial perfusion SPECT

BACKGROUND

Artificial intelligence-based methods have generated substantial interest in nuclear medicine. An area of significant interest has been the use of deep-learning (DL)-based approaches for denoising images acquired with lower doses, shorter acquisition times, or both. Objective evaluation of these approaches is essential for clinical application.

PURPOSE

DL-based approaches for denoising nuclear-medicine images have typically been evaluated using fidelity-based figures of merit (FoMs) such as root mean squared error (RMSE) and structural similarity index measure (SSIM). However, these images are acquired for clinical tasks and thus should be evaluated based on their performance in these tasks. Our objectives were to: (1) investigate whether evaluation with these FoMs is consistent with objective clinical-task-based evaluation; (2) provide a theoretical analysis for determining the impact of denoising on signal-detection tasks; and (3) demonstrate the utility of virtual imaging trials (VITs) to evaluate DL-based methods.

METHODS

A VIT to evaluate a DL-based method for denoising myocardial perfusion SPECT (MPS) images was conducted. To conduct this evaluation study, we followed the recently published best practices for the evaluation of AI algorithms for nuclear medicine (the RELAINCE guidelines). An anthropomorphic patient population modeling clinically relevant variability was simulated. Projection data for this patient population at normal and low-dose count levels (20%, 15%, 10%, 5%) were generated using well-validated Monte Carlo-based simulations. The images were reconstructed using a 3-D ordered-subsets expectation maximization-based approach. Next, the low-dose images were denoised using a commonly used convolutional neural network-based approach. The impact of DL-based denoising was evaluated using both fidelity-based FoMs and area under the receiver operating characteristic curve (AUC), which quantified performance on the clinical task of detecting perfusion defects in MPS images as obtained using a model observer with anthropomorphic channels. We then provide a mathematical treatment to probe the impact of post-processing operations on signal-detection tasks and use this treatment to analyze the findings of this study.

RESULTS

Based on fidelity-based FoMs, denoising using the considered DL-based method led to significantly superior performance. However, based on ROC analysis, denoising did not improve, and in fact, often degraded detection-task performance. This discordance between fidelity-based FoMs and task-based evaluation was observed at all the low-dose levels and for different cardiac-defect types. Our theoretical analysis revealed that the major reason for this degraded performance was that the denoising method reduced the difference in the means of the reconstructed images and of the channel operator-extracted feature vectors between the defect-absent and defect-present cases.

CONCLUSIONS

The results show the discrepancy between the evaluation of DL-based methods with fidelity-based metrics versus the evaluation on clinical tasks. This motivates the need for objective task-based evaluation of DL-based denoising approaches. Further, this study shows how VITs provide a mechanism to conduct such evaluations computationally, in a time and resource-efficient setting, and avoid risks such as radiation dose to the patient. Finally, our theoretical treatment reveals insights into the reasons for the limited performance of the denoising approach and may be used to probe the effect of other post-processing operations on signal-detection tasks.

Deep learning-based denoising in projection-domain and reconstruction-domain for low-dose myocardial perfusion SPECT

BACKGROUND

Low-dose (LD) myocardial perfusion (MP) SPECT suffers from high noise level, leading to compromised diagnostic accuracy. Here we investigated the denoising performance for MP-SPECT using a conditional generative adversarial network (cGAN) in projection-domain (cGAN-prj) and reconstruction-domain (cGAN-recon).

METHODS

Sixty-four noisy SPECT projections were simulated for a population of 100 XCAT phantoms with different anatomical variations and 99mTc-sestamibi distributions. Series of LD projections were obtained by scaling the full dose (FD) count rate to be 1/20 to 1/2 of the original. Twenty patients with 99mTc-sestamibi stress SPECT/CT scans were retrospectively analyzed. For each patient, LD SPECT images (7/10 to 1/10 of FD) were generated from the FD list mode data. All projections were reconstructed by the quantitative OS-EM method. A 3D cGAN was implemented to predict FD images from their corresponding LD images in the projection- and reconstruction-domain. The denoised projections were reconstructed for analysis in various quantitative indices along with cGAN-recon, Gaussian, and Butterworth-filtered images.

RESULTS

cGAN denoising improves image quality as compared to LD and conventional post-reconstruction filtering. cGAN-prj can further reduce the dose level as compared to cGAN-recon without compromising the image quality.

CONCLUSIONS

Denoising based on cGAN-prj is superior to cGAN-recon for MP-SPECT.

Deep-learning-based estimation of attenuation map improves attenuation correction performance over direct attenuation estimation for myocardial perfusion SPECT

BACKGROUND

Deep learning (DL)-based attenuation correction (AC) is promising to improve myocardial perfusion (MP) SPECT. We aimed to optimize and compare the DL-based direct and indirect AC methods, with and without SPECT and CT mismatch.

METHODS

One hundred patients with different 99mTc-sestamibi activity distributions and anatomical variations were simulated by a population of XCAT phantoms. Additionally, 34 patients 99mTc-sestamibi stress/rest SPECT/CT scans were retrospectively recruited. Projections were reconstructed by OS-EM method with or without AC. Mismatch between SPECT and CT images was modeled. A 3D conditional generative adversarial network (cGAN) was optimized for two DL-based AC methods: (i) indirect approach, i.e., non-attenuation corrected (NAC) SPECT paired with the corresponding attenuation map for training. The projections were reconstructed with the DL-generated attenuation map for AC; (ii) direct approach, i.e., NAC SPECT paired with the corresponding AC SPECT for training to perform direct AC.

RESULTS

Mismatch between SPECT and CT degraded DL-based AC performance. The indirect approach is superior to direct approach for various physical and clinical indices, even with mismatch modeled.

CONCLUSION

DL-based estimation of attenuation map for AC is superior and more robust to direct generation of AC SPECT.

Pix2Pix generative adversarial network for low dose myocardial perfusion SPECT denoising

BACKGROUND

Myocardial perfusion (MP) SPECT is a well-established method for diagnosing cardiac disease, yet its radiation risk poses safety concern. This study aims to apply and evaluate the use of Pix2Pix generative adversarial network (Pix2Pix GAN) in denoising low dose MP SPECT images.

METHODS

One hundred male and female patients with different 99mTc-sestamibi activity distributions, organ and body sizes were simulated by a population of digital 4D Extended Cardiac Torso (XCAT) phantoms. Realistic noisy SPECT projections of full dose of 987 MBq injection and 16 min acquisition, and low dose ranged from 1/20 to 1/2 of the full dose, were generated by an analytical projector from the right anterior oblique (RAO) to the left posterior oblique (LPO) positions. Additionally, twenty patients underwent ~1,184 MBq 99mTc-sestamibi stress SPECT/CT scan were also retrospectively recruited for the study. For each patient, low dose SPECT images (7/10 to 1/10 of full dose) were generated from the full dose list mode data. Our Pix2Pix GAN model was trained with full dose and low dose reconstructed SPECT image pairs. Normalized mean square error (NMSE), structural similarity index (SSIM), coefficient of variation (CV), full-width-at-half-maximum (FWHM) and relative defect size differences (RSD) of Pix2Pix GAN processed images were evaluated along with a reference convolutional auto encoder (CAE) network and post-reconstruction filters.

RESULTS

NMSE values of 0.0233±0.004 vs. 0.0249±0.004 and 0.0313±0.007 vs. 0.0579±0.016 were obtained on 1/2 and 1/20 dose level for Pix2Pix GAN and CAE in the simulation study, while they were 0.0376±0.010 vs. 0.0433±0.010 and 0.0907±0.020 vs. 0.1186±0.025 on 7/10 and 1/10 dose level in the clinical study. Similar results were also obtained from the SSIM, CV, FWHM and RSD values. Overall, the use of Pix2Pix GAN was superior to other denoising methods in all physical indices, particular in the lower dose levels in the simulation and clinical study.

CONCLUSIONS

The Pix2Pix GAN method is effective to reduce the noise level of low dose MP SPECT. Further studies on clinical performance are warranted to demonstrate its full clinical effectiveness.

The Use of Digital Coronary Phantoms for the Validation of Arterial Geometry Reconstruction and Computation of Virtual FFR

We present computational fluid dynamics (CFD) results of virtual fractional flow reserve (vFFR) calculations, performed on reconstructed arterial geometries derived from a digital phantom (DP). The latter provides a convenient and parsimonious description of the main vessels of the left and right coronary arterial trees, which, crucially, is CFD-compatible. Using our DP, we investigate the reconstruction error in what we deem to be the most relevant way—by evaluating the change in the computed value of vFFR, which results from varying (within representative clinical bounds) the selection of the virtual angiogram pair (defined by their viewing angles) used to segment the artery, the eccentricity and severity of the stenosis, and thereby, the CFD simulation’s luminal boundary. The DP is used to quantify reconstruction and computed haemodynamic error within the VIRTUheartTM software suite. However, our method and the associated digital phantom tool are readily transferable to equivalent, clinically oriented workflows. While we are able to conclude that error within the VIRTUheartTM workflow is suitably controlled, the principal outcomes of the work reported here are the demonstration and provision of a practical tool along with an exemplar methodology for evaluating error in a coronary segmentation process.

Technical Note: Respiratory impacts on static and respiratory gated 99m Tc-MAA SPECT/CT for liver radioembolization- A simulation study

PURPOSE

We aimed to evaluate respiratory impacts on static and respiratory gated (RG) ^99m Tc-MAA SPECT in terms of respiratory motion (RM) blur, attenuation correction (AC) and volume-of-interest (VOI) segmentation on lung shunt faction (LSF) and tumor-to-normal liver ratio (TNR) estimation for liver radioembolization therapy planning.

METHODS

The XCAT phantom was used to simulate a population of 300 phantoms, modelling various anatomical variations, tumor characteristics, respiratory motion amplitudes, LSFs and TNRs. One hundred and twenty noisy projections of average activity maps near end-expiration (End-EX) and whole respiratory cycle were simulated analytically, modeling attenuation and geometric collimator-detector-response (GCDR). The OS-EM algorithm was employed for reconstruction, modeling AC and GCDR. RM effect was evaluated for static SPECT, while AC and VOI mismatch effects were investigated independently and together for static and RG SPECT utilizing one gate, i.e., End-EX. LSF and TNR errors were measured based on the ground truth. Lesions with different characteristics were also investigated for static and RG SPECT.

RESULTS

RM overestimates LSF and underestimates TNR. The VOI mismatch caused the largest errors in both RG and static SPECT for LSF and TNR estimation, reaching 160% and -52% correspondingly with extremely mismatched VOIs for RG SPECT, even larger than those for static SPECT. With matched AC and VOIs, RG SPECT has better performance than static SPECT. Larger TNR errors are associated with tumors of smaller sizes and higher TNR for static SPECT.

CONCLUSIONS

The VOI segmentation mismatch has a stronger impact, followed by RM and AC in static ^99m Tc-MAA SPECT/CT. This effect is more pronounced for RG SPECT. When VOI masks are derived from a matched CT, RG SPECT is generally superior to static SPECT for LSF and TNR estimation. The performance of RG SPECT could be worse than static SPECT when a mismatched CT is used for segmentation. This article is protected by copyright. All rights reserved.

Ideal-Observer Computation with anthropomorphic phantoms using Markov chain Monte Carlo

In medical imaging, it is widely recognized that image quality should be objectively evaluated based on performance in clinical tasks. To evaluate performance in signal-detection tasks, the ideal observer (IO) is optimal but also challenging to compute in clinically realistic settings. Markov Chain Monte Carlo (MCMC)-based strategies have demonstrated the ability to compute the IO using pre-computed projections of an anatomical database. To evaluate image quality in clinically realistic scenarios, the observer performance should be measured for a realistic patient distribution. This implies that the anatomical database should also be derived from a realistic population. In this manuscript, we propose to advance the MCMC-based approach towards achieving these goals. We then use the proposed approach to study the effect of anatomical database size on IO computation for the task of detecting perfusion defects in simulated myocardial perfusion SPECT images. Our preliminary results provide evidence that the size of the anatomical database affects the computation of the IO.

An update on computational anthropomorphic anatomical models

The prevalent availability of high-performance computing coupled with validated computerized simulation platforms as open-source packages have motivated progress in the development of realistic anthropomorphic computational models of the human anatomy. The main application of these advanced tools focused on imaging physics and computational internal/external radiation dosimetry research. This paper provides an updated review of state-of-the-art developments and recent advances in the design of sophisticated computational models of the human anatomy with a particular focus on their use in radiation dosimetry calculations. The consolidation of flexible and realistic computational models with biological data and accurate radiation transport modeling tools enables the capability to produce dosimetric data reflecting actual setup in clinical setting. These simulation methodologies and results are helpful resources for the medical physics and medical imaging communities and are expected to impact the fields of medical imaging and dosimetry calculations profoundly.

Evaluation of different CT maps for attenuation correction and segmentation in static 99m Tc-MAA SPECT/CT for 90 Y radioembolization treatment planning: A simulation study

PURPOSE

Conventional ^99m Tc-macroaggregated albumin (^99m Tc-MAA) planar scintigraphy overestimates lung shunt fraction (LSF) compared to SPECT/CT. However, the respiratory motion artifact due to the temporal mismatch between static SPECT and helical CT (HCT) may compromise the SPECT quantitation accuracy by incorrect attenuation correction (AC) and volume-of-interest (VOI) segmentation. This study aims to evaluate AC and VOI segmentation effects systematically and to propose a CT map for LSF and tumor-to-normal liver ratio (TNR) estimation in static ^99m Tc-MAA SPECT/CT.

METHODS

The 4D XCAT phantom was used to simulate a phantom population of 120 phantoms, modeling 10 different anatomical variations, nine TNRs (2-13.2), nine tumor sizes (2-6.7 cm diameter), eight tumor locations, three axial motion amplitudes of 1, 1.5, and 2 (cm), and four LSFs of 5%, 10%, 15%, and 20%. An analytical projector for low-energy high-resolution parallel-hole collimator was used to simulate 60 noisy projections over 360°, modeling attenuation and geometric collimator-detector response (GCDR). AC and VOI mismatch effects were investigated independently and together, using cine average CT (CACT), HCT at end-inspiration (HCT-IN), mid-respiration (HCT-MID), and end-expiration (HCT-EX) respectively as attenuation and segmentation maps. SPECT images without motion, AC, and VOI errors were also generated as reference. LSF and TNR errors were measured as compared to the ground truth.

RESULTS

HCT-MID has slightly better performance for AC effect compared with other CT maps in LSF and TNR estimation, while HCT-EX and HCT-MID perform better for VOI effect. For a respiratory motion amplitude of 1.5 cm and a LSF of 5%, the LSF errors are 19.56 ± 4.58%, -6.79 ± 1.74%, 77.29 ± 14.74%, and 111.25 ± 18.29% corresponding to HCT-MID, HCT-EX, HCT-IN, and CACT in static SPECT. The TNR errors are -12.38 ± 6.42%, -20.55 ± 11.25%, -20.89 ± 9.98%, and -22.89 ± 14.38% respectively. HCT-MID has the best performance for LSF estimation for LSF > 10% and TNR estimation, followed by HCT-EX, HCT-IN, and CACT.

CONCLUSIONS

The HCT-MID is recommended for AC and segmentation to alleviate respiratory artifacts and improve quantitation accuracy in ⁹⁰ Y radioembolization treatment planning. HCT-EX would also be a recommended choice if HCT-MID is not available.

A physics and learning-based transmission-less attenuation compensation method for SPECT

Attenuation compensation (AC) is a pre-requisite for reliable quantification and beneficial for visual interpretation tasks in single-photon emission computed tomography (SPECT). Typical AC methods require the availability of an attenuation map, which is obtained using a transmission scan, such as a CT scan. This has several disadvantages such as increased radiation dose, higher costs, and possible misalignment between SPECT and CT scans. Also, often a CT scan is unavailable. In this context, we and others are showing that scattered photons in SPECT contain information to estimate the attenuation distribution. To exploit this observation, we propose a physics and learning-based method that uses the SPECT emission data in the photopeak and scatter windows to perform transmission-less AC in SPECT. The proposed method uses data acquired in the scatter window to reconstruct an initial estimate of the attenuation map using a physics-based approach. A convolutional neural network is then trained to segment this initial estimate into different regions. Pre-defined attenuation coefficients are assigned to these regions, yielding the reconstructed attenuation map, which is then used to reconstruct the activity distribution using an ordered subsets expectation maximization (OSEM)-based reconstruction approach. We objectively evaluated the performance of this method using highly realistic simulation studies conducted on the clinically relevant task of detecting perfusion defects in myocardial perfusion SPECT. Our results showed no statistically significant differences between the performance achieved using the proposed method and that with the true attenuation maps. Visually, the images reconstructed using the proposed method looked similar to those with the true attenuation map. Overall, these results provide evidence of the capability of the proposed method to perform transmission-less AC and motivate further evaluation.

Evaluation of different respiratory gating schemes for cardiac SPECT

BACKGROUND

Respiratory gating reduces motion blurring in cardiac SPECT. Here we aim to evaluate the performance of three respiratory gating strategies using a population of digital phantoms with known truth and clinical data.

METHODS

We analytically simulated 60 projections for 10 XCAT phantoms with 99mTc-sestamibi distributions using three gating schemes: equal amplitude gating (AG), equal count gating (CG), and equal time gating (TG). Clinical list-mode data for 10 patients who underwent 99mTc-sestamibi scans were also processed using the 3 gating schemes. Reconstructed images in each gate were registered to a reference gate, averaged and reoriented to generate the polar plots. For simulations, image noise, relative difference (RD) of averaged count for each of the 17 segment, and relative defect size difference (RSD) were analyzed. For clinical data, image intensity profile and FWHM were measured across the left ventricle wall.

RESULTS

For simulations, AG and CG methods showed significantly lower RD and RSD compared to TG, while noise variation was more non-uniform through different gates for AG. In the clinical study, AG and CG had smaller FWHM than TG.

CONCLUSIONS

AG and CG methods show better performance for motion reduction and are recommended for clinical respiratory gating SPECT implementation.

Recent advances in cardiac SPECT instrumentation and imaging methods

Cardiac SPECT continues to play a critical role in detecting and managing cardiovascular disease, in particularly coronary artery disease (CAD) (Jaarsma et al 2012 J. Am. Coll. Cardiol. 59 1719-28), (Agostini et al 2016 Eur. J. Nucl. Med. Mol. Imaging 43 2423-32). While conventional dual-head SPECT scanners using parallel-hole collimators and scintillation crystals with photomultiplier tubes are still the workhorse of cardiac SPECT, they have the limitations of low photon sensitivity (~130 count s^-1 MBq^-1), poor image resolution (~15 mm) (Imbert et al 2012 J. Nucl. Med. 53 1897-903), relatively long acquisition time, inefficient use of the detector, high radiation dose, etc. Recently our field observed an exciting growth of new developments of dedicated cardiac scanners and collimators, as well as novel imaging algorithms for quantitative cardiac SPECT. These developments have opened doors to new applications with potential clinical impact, including ultra-low-dose imaging, absolute quantification of myocardial blood flow (MBF) and coronary flow reserve (CFR), multi-radionuclide imaging, and improved image quality as a result of attenuation, scatter, motion, and partial volume corrections (PVCs). In this article, we review the recent advances in cardiac SPECT instrumentation and imaging methods. This review mainly focuses on the most recent developments published since 2012 and points to the future of cardiac SPECT from an imaging physics perspective.

Joint reconstruction of rest/stress myocardial perfusion SPECT

Myocardial perfusion imaging (MPI) using rest/stress single photon emission computed tomography (SPECT) allows non-invasive assessment of reversible cardiac perfusion defects. Conventionally, reversible defects are identified using a difference image, called reversible map, obtained by subtracting the stress image from the rest image after registration and normalization of the two images. The identification of reversible defects using the conventional subtraction method is however limited by noise. We propose to jointly reconstruct rest and stress projection data to directly obtain the reversible map in a single reconstruction framework to improve the detectability of reversible defects. To evaluate the performance of the proposed method, we performed phantom studies to mimic reversible defects with different levels of severity and doses. As compared to the conventional subtraction method, the joint method yielded reversible maps with much lower noise and improved defect detectability. At a normal clinical dose level, the joint method improved the signal to noise ratio (SNR) of defect contrast in reversible maps from 13.2 to 66.4, 9.7 to 35.0, 6.1 to 13.2, and 3.1 to 6.5, for defect to normal myocardium concentration ratios of 0%, 25%, 50%, and 75%, respectively. The SNRs obtained using the joint method were improved from 6.1 to 13.2, 3.9 to 9.4, 3.0 to 8.0, and 2.1 to 7.1, for 100%, 75%, 50%, and 25% of the normal clinical dose as compared to the conventional subtraction method. To access clinical feasibility, we applied the joint method to a rest/stress SPECT MPI patient study. The joint method yielded a reversible map with much lower noise, translating into a much higher defect detectability as compared to the conventional subtraction method. Our results indicate that the joint method has the potential to improve radiologists' performance for assessing defects in rest/stress SPECT MPI. In addition, the joint method can be used to reduce dose or imaging time.

A comparison of resampling schemes for estimating model observer performance with small ensembles

In objective assessment of image quality, an ensemble of images is used to compute the 1st and 2nd order statistics of the data. Often, only a finite number of images is available, leading to the issue of statistical variability in numerical observer performance. Resampling-based strategies can help overcome this issue. In this paper, we compared different combinations of resampling schemes (the leave-one-out (LOO) and the half-train/half-test (HT/HT)) and model observers (the conventional channelized Hotelling observer (CHO), channelized linear discriminant (CLD) and channelized quadratic discriminant). Observer performance was quantified by the area under the ROC curve (AUC). For a binary classification task and for each observer, the AUC value for an ensemble size of 2000 samples per class served as a gold standard for that observer. Results indicated that each observer yielded a different performance depending on the ensemble size and the resampling scheme. For a small ensemble size, the combination [CHO, HT/HT] had more accurate rankings than the combination [CHO, LOO]. Using the LOO scheme, the CLD and CHO had similar performance for large ensembles. However, the CLD outperformed the CHO and gave more accurate rankings for smaller ensembles. As the ensemble size decreased, the performance of the [CHO, LOO] combination seriously deteriorated as opposed to the [CLD, LOO] combination. Thus, it might be desirable to use the CLD with the LOO scheme when smaller ensemble size is available.

Use of Sub-Ensembles and Multi-Template Observers to Evaluate Detection Task Performance for Data That are Not Multivariate Normal

The Hotelling Observer (HO) is widely used to evaluate image quality in medical imaging. However, applying it to data that are not multivariate-normally (MVN) distributed is not optimal. In this paper, we apply two multi-template linear observer strategies to handle such data. First, the entire data ensemble is divided into sub-ensembles that are exactly or approximately MVN and homoscedastic. Next, a different linear observer template is estimated for and applied to each sub-ensemble. The first multi-template strategy, adapted from previous work, applies the HO to each sub-ensemble, calculates the area under the receiver operating characteristics curve (AUC) for each sub-ensemble, and averages the AUCs from all the sub-ensembles. The second strategy applies the Linear Discriminant (LD) to estimate test statistics for each sub-ensemble and calculates a single global AUC using the pooled test statistics from all the sub-ensembles. We show that this second strategy produces the maximum AUC when only shifting of the HO test statistics is allowed. We compared these strategies to the use of a single HO template for the entire data ensemble by applying them to the non-MVN data obtained from reconstructed images of a realistic simulated population of myocardial perfusion SPECT studies with the goal of optimizing the reconstruction parameters. Of the strategies investigated, the multi-template LD strategy yielded the highest AUC for any given set of reconstruction parameters. The optimal reconstruction parameters obtained by the two multi-template strategies were comparable and produced higher AUCs for each sub-ensemble than the single-template HO strategy.

Simplifying volumes-of-interest (VOIs) definition in quantitative SPECT: Beyond manual definition of 3D whole-organ VOIs

PURPOSE

We investigated the feasibility of using simpler methods than manual whole-organ volume-of-interest (VOI) definition to estimate the organ activity concentration in single photon emission computed tomography (SPECT) in cases where the activity in the organ can be assumed to be uniformly distributed on the scale of the voxel size. In particular, we investigated an anatomic region-of-interest (ROI) defined in a single transaxial slice, and a single sphere placed inside the organ boundaries.

METHODS

The evaluation was carried out using Monte Carlo simulations based on patient indium ¹¹¹ In pentetreotide SPECT and computed tomography (CT) images. We modeled constant activity concentrations in each organ, validating this assumption by comparing the distribution of voxel values inside the organ VOIs of the simulated data with the patient data. We simulated projection data corresponding to 100, 50, and 25% of the clinical count level to study the effects of noise level due to shortened acquisition time. Images were reconstructed using a previously validated quantitative SPECT reconstruction method. The evaluation was performed in terms of the accuracy and precision of the activity concentration estimates.

RESULTS

The results demonstrated that the non-uniform image intensity observed in the reconstructed images in the organs with normal uptake was consistent with uniform activity concentration in the organs on the scale of the voxel size; observed non-uniformities in image intensity were due to a combination of partial-volume effects at the boundaries of the organ, artifacts in the reconstructed image due to collimator-detector response compensation, and noise. Using an ROI defined in a single transaxial slice produced similar biases compared to the three-dimensional (3D) whole-organ VOIs, provided that the transaxial slice was near the central plane of the organ and that the pixels from the organ boundaries were not included in the ROI. Although this slice method was sensitive to noise, biases were less than 10% for all the noise levels studied. The use of spherical VOIs was more sensitive to noise. The method was more accurate for larger spheres and larger organs such as the liver in comparison to the kidneys. Biases lower than 7% were found in the liver when using large enough spheres (radius ≥ 28 mm), regardless of the position, of the VOI inside the organ even with shortened acquisition times. The biases were more position-dependent for smaller organs.

CONCLUSIONS

Both of the simpler methods provided suitable surrogates in terms of accuracy and precision. The results suggested that a spherical VOI was more appropriate for estimating the activity concentration in larger organs such as the liver, regardless of the position of the sphere inside the organ. Larger spheres resulted in better estimates. A single-slice ROI was more suitable for activity estimation in smaller organs such as the kidneys, providing that the transaxial slice selected was near the central plane of the organ and that voxels from the organ boundaries were excluded. Under those conditions, activity concentrations with biases lower than 5% were observed for all the studied count levels and coefficients of variation were less than 9% and 5% for the 25% and 100% count levels, respectively.

Collimator optimization and collimator-detector response compensation in myocardial perfusion SPECT using the ideal observer with and without model mismatch and an anthropomorphic model observer

The collimator is the primary factor that determines the spatial resolution and noise tradeoff in myocardial perfusion SPECT images. In this paper, the goal was to find the collimator that optimizes the image quality in terms of a perfusion defect detection task. Since the optimal collimator could depend on the level of approximation of the collimator-detector response (CDR) compensation modeled in reconstruction, we performed this optimization for the cases of modeling the full CDR (including geometric, septal penetration and septal scatter responses), the geometric CDR, or no model of the CDR. We evaluated the performance on the detection task using three model observers. Two observers operated on data in the projection domain: the Ideal Observer (IO) and IO with Model-Mismatch (IO-MM). The third observer was an anthropomorphic Channelized Hotelling Observer (CHO), which operated on reconstructed images. The projection-domain observers have the advantage that they are computationally less intensive. The IO has perfect knowledge of the image formation process, i.e. it has a perfect model of the CDR. The IO-MM takes into account the mismatch between the true (complete and accurate) model and an approximate model, e.g. one that might be used in reconstruction. We evaluated the utility of these projection domain observers in optimizing instrumentation parameters. We investigated a family of 8 parallel-hole collimators, spanning a wide range of resolution and sensitivity tradeoffs, using a population of simulated projection (for the IO and IO-MM) and reconstructed (for the CHO) images that included background variability. We simulated anterolateral and inferior perfusion defects with variable extents and severities. The area under the ROC curve was estimated from the IO, IO-MM, and CHO test statistics and served as the figure-of-merit. The optimal collimator for the IO had a resolution of 9-11 mm FWHM at 10 cm, which is poorer resolution than typical collimators used for MPS. When the IO-MM and CHO used a geometric or no model of the CDR, the optimal collimator shifted toward higher resolution than that obtained using the IO and the CHO with full CDR modeling. With the optimal collimator, the IO-MM and CHO using geometric modeling gave similar performance to full CDR modeling. Collimators with poorer resolution were optimal when CDR modeling was used. The agreement of rankings between the IO-MM and CHO confirmed that the IO-MM is useful for optimization tasks when model mismatch is present due to its substantially reduced computational burden compared to the CHO.

Collimator optimization in myocardial perfusion SPECT using the ideal observer and realistic background variability for lesion detection and joint detection and localization tasks

In SPECT imaging, collimators are a major factor limiting image quality and largely determine the noise and resolution of SPECT images. In this paper, we seek the collimator with the optimal tradeoff between image noise and resolution with respect to performance on two tasks related to myocardial perfusion SPECT: perfusion defect detection and joint detection and localization. We used the Ideal Observer (IO) operating on realistic background-known-statistically (BKS) and signal-known-exactly (SKE) data. The areas under the receiver operating characteristic (ROC) and localization ROC (LROC) curves (AUCd, AUCd+l), respectively, were used as the figures of merit for both tasks. We used a previously developed population of 54 phantoms based on the eXtended Cardiac Torso Phantom (XCAT) that included variations in gender, body size, heart size and subcutaneous adipose tissue level. For each phantom, organ uptakes were varied randomly based on distributions observed in patient data. We simulated perfusion defects at six different locations with extents and severities of 10% and 25%, respectively, which represented challenging but clinically relevant defects. The extent and severity are, respectively, the perfusion defect's fraction of the myocardial volume and reduction of uptake relative to the normal myocardium. Projection data were generated using an analytical projector that modeled attenuation, scatter, and collimator-detector response effects, a 9% energy resolution at 140 keV, and a 4 mm full-width at half maximum (FWHM) intrinsic spatial resolution. We investigated a family of eight parallel-hole collimators that spanned a large range of sensitivity-resolution tradeoffs. For each collimator and defect location, the IO test statistics were computed using a Markov Chain Monte Carlo (MCMC) method for an ensemble of 540 pairs of defect-present and -absent images that included the aforementioned anatomical and uptake variability. Sets of test statistics were computed for both tasks and analyzed using ROC and LROC analysis methodologies. The results of this study suggest that collimators with somewhat poorer resolution and higher sensitivity than those of a typical low-energy high-resolution (LEHR) collimator were optimal for both defect detection and joint detection and localization tasks in myocardial perfusion SPECT for the range of defect sizes investigated. This study also indicates that optimizing instrumentation for a detection task may provide near-optimal performance on the more challenging detection-localization task.

Factors affecting the normality of channel outputs of channelized model observers: an investigation using realistic myocardial perfusion SPECT images

The channelized Hotelling observer (CHO) uses the first- and second-order statistics of channel outputs under both hypotheses to compute test statistics used in binary classification tasks. If these input data deviate from a multivariate normal (MVN) distribution, the classification performance will be suboptimal compared to an ideal observer operating on the same channel outputs. We conducted a comprehensive investigation to rigorously study the validity of the MVN assumption under various kinds of background and signal variability in a realistic population of phantoms. The study was performed in the context of myocardial perfusion SPECT imaging; anatomical, uptake (intensity), and signal variability were simulated. Quantitative measures and graphical approaches applied to the outputs of each channel were used to investigate the amount and type of deviation from normality. For some types of background and signal variations, the channel outputs, under both hypotheses, were non-normal (i.e., skewed or multimodal). This indicates that, for realistic medical images in cases where there is signal or background variability, the normality of the channel outputs should be evaluated before applying a CHO. Finally, the different degrees of departure from normality of the various channels are explained in terms of violations of the central limit theorem.

Optimization and comparison of simultaneous and separate acquisition protocols for dual isotope myocardial perfusion SPECT

Dual-isotope simultaneous-acquisition (DISA) rest-stress myocardial perfusion SPECT (MPS) protocols offer a number of advantages over separate acquisition. However, crosstalk contamination due to scatter in the patient and interactions in the collimator degrade image quality. Compensation can reduce the effects of crosstalk, but does not entirely eliminate image degradations. Optimizing acquisition parameters could further reduce the impact of crosstalk. In this paper we investigate the optimization of the rest Tl-201 energy window width and relative injected activities using the ideal observer (IO), a realistic digital phantom population and Monte Carlo (MC) simulated Tc-99m and Tl-201 projections as a means to improve image quality. We compared performance on a perfusion defect detection task for Tl-201 acquisition energy window widths varying from 4 to 40 keV centered at 72 keV for a camera with a 9% energy resolution. We also investigated 7 different relative injected activities, defined as the ratio of Tc-99m and Tl-201 activities, while keeping the total effective dose constant at 13.5 mSv. For each energy window and relative injected activity, we computed the IO test statistics using a Markov chain Monte Carlo (MCMC) method for an ensemble of 1,620 triplets of fixed and reversible defect-present, and defect-absent noisy images modeling realistic background variations. The volume under the 3-class receiver operating characteristic (ROC) surface (VUS) was estimated and served as the figure of merit. For simultaneous acquisition, the IO suggested that relative Tc-to-Tl injected activity ratios of 2.6-5 and acquisition energy window widths of 16-22% were optimal. For separate acquisition, we observed a broad range of optimal relative injected activities from 2.6 to 12.1 and acquisition energy window of widths 16-22%. A negative correlation between Tl-201 injected activity and the width of the Tl-201 energy window was observed in these ranges. The results also suggested that DISA methods could potentially provide image quality as good as that obtained with separate acquisition protocols. We compared observer performance for the optimized protocols and the current clinical protocol using separate acquisition. The current clinical protocols provided better performance at a cost of injecting the patient with approximately double the injected activity of Tc-99m and Tl-201, resulting in substantially increased radiation dose.

The development and initial evaluation of a realistic simulated SPECT dataset with simultaneous respiratory and cardiac motion for gated myocardial perfusion SPECT

We developed a realistic simulation dataset for simultaneous respiratory and cardiac (R&C) gated SPECT/CT using the 4D NURBS-based Cardiac-Torso (NCAT) Phantom and Monte Carlo simulation methods, and evaluated it for a sample application study. The 4D NCAT phantom included realistic respiratory motion and beating heart motion based on respiratory gated CT and cardiac tagged MRI data of normal human subjects. To model the respiratory motion, a set of 24 separate 3D NCAT phantoms excluding the heart was generated over a respiratory cycle. The beating heart motion was modeled separately with 48 frames per cardiac cycle for each of the 24 respiratory phases. The resultant set of 24  ×  48 3D NCAT phantoms provides a realistic model of a normal human subject at different phases of combined R&C motions. An almost noise-free SPECT projection dataset for each of the 1152 3D NCAT phantoms was generated using Monte Carlo simulation techniques and the radioactivity uptake distribution of (99m)Tc sestamibi in different organs. By grouping and summing the separate projection datasets, separate or simultaneous R&C gated acquired data with different gating schemes could be simulated. In the initial evaluation, we combined the projection datasets into ungated, 6 respiratory-gates only, 8 cardiac-gates only, and combined 6 respiratory-gates & 8 cardiac-gates projection datasets. Each dataset was reconstructed using 3D OS-EM without and with attenuation correction using the averaged and respiratory-gated attenuation maps, and the resulting reconstructed images were compared. These results were used to demonstrate the effects of R&C motions and the reduction of image artifact due to R&C motions by gating and attenuation corrections. We concluded that the realistic 4D NCAT phantom and Monte Carlo simulated SPECT projection datasets with R&C motions are powerful tools in the study of the effects of R&C motions, as well as in the development of R&C gating schemes and motion correction methods for improved SPECT/CT imaging.

Optimization and evaluation of reconstruction-based compensation methods and reconstruction parameters for Tc-99m MIBI parathyroid SPECT

The value of Tc-99m MIBI parathyroid SPECT for localizing parathyroid hyperplasia in chronic renal failure patients remains inconclusive due to limited image quality. Advanced reconstruction methods to improve image quality have been developed but require optimization and evaluation. The goal of this study was to optimize and evaluate compensation methods and reconstruction parameters for Tc-99m MIBI parathyroid SPECT. A phantom population and projection data that modelled clinically realistic variations found in patients were simulated. The 3D OS-EM reconstruction with compensation for attenuation, detector response and scatter in various combinations were studied. For each compensation, the number of updates for OS-EM and the cutoff frequency of a 3D Butterworth filter were optimized and evaluated using anthropomorphic model observer. With optimal parameters, the method with compensation for attenuation and detector response, with or without the addition of scatter compensation, provided the highest lesion detectability for Tc-99m MIBI parathyroid SPECT.

Optimization of energy window and evaluation of scatter compensation methods in myocardial perfusion SPECT using the ideal observer with and without model mismatch and an anthropomorphic model observer

We used the ideal observer (IO) and IO with model mismatch (IO-MM) applied in the projection domain and an anthropomorphic channelized Hotelling observer (CHO) applied to reconstructed images to optimize the acquisition energy window width and to evaluate various scatter compensation methods in the context of a myocardial perfusion single-photon emission computed tomography (SPECT) defect detection task. The IO has perfect knowledge of the image formation process and thus reflects the performance with perfect compensation for image-degrading factors. Thus, using the IO to optimize imaging systems could lead to suboptimal parameters compared with those optimized for humans interpreting SPECT images reconstructed with imperfect or no compensation. The IO-MM allows incorporating imperfect system models into the IO optimization process. We found that with near-perfect scatter compensation, the optimal energy window for the IO and CHO was similar; in its absence, the IO-MM gave a better prediction of the optimal energy window for the CHO using different scatter compensation methods. These data suggest that the IO-MM may be useful for projectiondomain optimization when MM is significant and that the IO is useful when followed by reconstruction with good models of the image formation process.