Motion estimation for cardiac emission tomography by optical flow methods.

This paper describes a new method for estimating the 3D, non-rigid object motion in a time sequence of images. The method is a generalization of a standard optical flow algorithm that is incorporated into a successive quadratic approximation framework. The method was evaluated for gated cardiac emission tomography using images obtained from a mathematical, 4D phantom and a physical, dynamic phantom. The results showed that the proposed method offers improved motion estimation accuracy relative to the standard optical flow method. Convergence of the proposed algorithm was evidenced with a monotonically decreasing objective function value with iteration. Practical application of the motion estimation method in cardiac emission tomography includes quantitative myocardial motion estimation and 4D, motion-compensated image reconstruction.

Modeling the axial extension of a transmission line source within iterative reconstruction via multiple transmission sources.

Reconstruction algorithms for transmission tomography have generally assumed that the photons reaching a particular detector bin at a particular angle originate from a single point source. In this paper, we highlight several cases of extended transmission sources, in which it may be useful to approach the estimation of attenuation coefficients as a problem involving multiple transmission point sources. Examined in detail is the case of a fixed transmission line source with a fan-beam collimator. This geometry can result in attenuation images that have significant axial blur. Herein it is also shown, empirically, that extended transmission sources can result in biased estimates of the average attenuation, and an explanation is proposed. The finite axial resolution of the transmission line source configuration is modeled within iterative reconstruction using an expectation-maximization algorithm that was previously derived for estimating attenuation coefficients from single photon emission computed tomography (SPECT) emission data. The same algorithm is applicable to both problems because both can be thought of as involving multiple transmission sources. It is shown that modeling axial blur within reconstruction removes the bias in the average estimated attenuation and substantially improves the axial resolution of attenuation images.

Transmission scanning system for a gamma camera coincidence scanner.

UNLABELLED: The goal of this research was to develop and evaluate a practical transmission scanning system for attenuation correction on a 2-head gamma camera coincidence scanner. METHODS: The transmission system operates in singles mode and uses point sources of 137Cs that emit 662-keV gamma-radiation. Each point source is inserted between existing septa that are normally used to provide an approximately 2-dimensional emission acquisition geometry. The sources are placed along a line parallel to the axis of rotation near the edge of 1 camera. Data are acquired with the opposing camera. The septa provide axial collimation for the sources so that the transmission system operates in a 2-dimensional offset fanbeam geometry. Camera energy and spatial resolution were measured at 511 and 662 keV. Sensitivity was measured at 662 keV. The effects on axial resolution of adding supplemental collimation to the septa were shown. The system was calibrated and tested using a resolution (rod) phantom and a uniformity phantom. Torso phantom data were acquired. Patient transmission and emission scans were obtained. Postinjection transmission data were used to correct patient emission data. RESULTS: The camera resolution at postinjection counting rates was 11.7% full width at half maximum (FWHM) for 662-keV gamma-rays. Intrinsic spatial resolution was 2.7 mm (FWHM) at 662 keV. The sensitivity of the system was 280 Hz/MBq using five 74-MBq sources of 137Cs in the transmission geometry, with supplemental collimation added to the septa to improve axial resolution. The transaxial resolution of the system was such that the smallest rods (6-mm diameter and 12-mm spacing) were well resolved in a reconstructed resolution-phantom image. The corrected patient emission scans were free of attenuation-induced artifacts. CONCLUSION: An easily implemented transmission system for a 2-head gamma camera coincidence scanner that can be used for postinjection transmission scanning has been developed.

Quantitative pulmonary single photon emission computed tomography for radiotherapy applications.

Pulmonary imaging using single photon emission computed tomography (SPECT) is the focus of current radiotherapy research, including dose-response analysis and three-dimensional (3D) radiation treatment planning. Improvement in the quantitative capability of SPECT may help establish its potential role in this application as well as others requiring accurate knowledge of pulmonary blood flow. The purposes of this study were to quantitatively evaluate SPECT filtered backprojection (FBP) and ordered subset-expectation maximization (OS-EM) reconstruction implementations for measuring absolute activity concentration in lung phantom experiments, and to incorporate quantitative SPECT techniques in 3D-RTP for lung cancer. Quantitative FBP (nonuniform iterative Chang attenuation compensation, scatter correction, and 3D postreconstruction Metz filtering) and OS-EM implementations were compared with a "clinical" implementation of FBP (uniform multiplicative Chang attenuation compensation and post-reconstruction von Hann filtering), for their ability to improve quantification of inactive and active spherical defects in the lungs of an anthropomorphic torso phantom. Activity concentration estimates were found to depend on many factors, such as region of interest size, scatter subtraction constant (k), postreconstruction deconvolution filtering and, in the case of OS-EM, total number of iterations. In general, reconstruction implementations incorporating compensation for nonuniform attenuation and scatter provided reduced bias relative to the clinical implementation. Potential applications to lung radiotherapy, including dose-functional histograms and treatment planning are also discussed. SPECT has the potential to provide accurate estimates of lung activity distributions that, together with improved image quality, may be useful for the study and prediction of therapeutic response.

Multimodality nuclear medicine imaging in three-dimensional radiation treatment planning for lung cancer: challenges and prospects.

The purpose of this study was to determine the utility of quantitative single photon emission computed tomography (SPECT) lung perfusion scans and F-18 fluorodeoxyglucose positron emission computed tomography (PET) during X-ray computed tomography (CT)-based treatment planning for patients with lung cancer. Pre-radiotherapy SPECT (n = 104) and PET (n = 35) images were available to the clinician to assist in radiation field design for patients with bronchogenic cancer. The SPECT and PET scans were registered with anatomic information derived from CT. The information from SPECT and PET provides the treatment planner with functional data not seen with CT. SPECT yields three-dimensional (3D) lung perfusion maps. PET provides 3D metabolic images that assist in tumor localization. The impact of the nuclear medicine images on the treatment planning process was assessed by determining the frequency, type, and extent of changes to plans. Pre-radiotherapy SPECT scans were used to modify 11 (11%) treatment plans; primarily altering beam angles to avoid highly functioning tissue. Fifty (48%) SPECT datasets were judged to be 'potentially useful' due to the detection of hypoperfused regions of the lungs, but were not used during treatment planning. PET data influenced 34% (12 of 35) of the treatment plans examined, and resulted in enlarging portions of the beam aperture (margins) up to 15 mm. Challenges associated with image quality and registration arise when utilizing nuclear medicine data in the treatment planning process. Initial implementation of advanced SPECT image reconstruction techniques that are not typically used in the clinic suggests that the reconstruction method may influence dose response data derived from the SPECT images and improve image registration with CT. The use of nuclear medicine transmission computed tomography (TCT) for both SPECT and PET is presented as a possible tool to reconstruct more accurate emission images and to aid in the registration of emission data with the planning CT. Nuclear medicine imaging techniques appear to be a potentially valuable tool during radiotherapy treatment planning for patients with lung cancer. The utilization of accurate nuclear medicine image reconstruction techniques and TCT may improve the treatment planning process.

Transmission imaging for nonuniform attenuation correction using a three-headed SPECT camera.

UNLABELLED: Our objective was to build and test a new system for transmission CT (TCT) imaging on a three-headed SPECT camera. The TCT images are intended for use in nonuniform attenuation correction of cardiac SPECT data. METHODS: The system consists of a transmission line source mounted to the camera gantry at the focal line of a long focal length, asymmetric fanbeam collimator. The focal line is 114 cm from the collimator surface and shifted 20 cm from the detector midline. This asymmetric fanbeam geometry is used to reduce truncation artifacts in the reconstructed TCT image. The line source fixture accommodates a 25-cm long source and contains removable, variable thickness attenuator plates (copper or lead) to modulate the photon flux density and a slat collimator to collimate the TCT source beam in the axial direction. For the TCT reconstruction, an iterative maximum likelihood-expectation maximization algorithm is used that models the asymmetric fanbeam geometry. Our initial studies with this system used a 1850 MBq (50 mCi) 123mTe line source. The evaluation included TCT scans of a resolution phantom, an anthropomorphic thorax phantom and a human subject. For the thorax phantom and human subject, short (2-min) and long (14-min) scans were performed. The SPECT imaging performance of the fanbeam collimator was also characterized. RESULTS: For both phantom and human data, high quality TCT reconstructions were obtained with linear attenuation coefficients closely matching narrow beam values. In the images of the resolution phantom, the smallest rods (4.8-mm diam) were resolved. The long scan images of the thorax phantom and human subject demonstrated the high resolution nature of the system and contained no evidence of truncation artifacts. With smoothing to control noise, the short scan images generally retained the attenuation features of the lung and of soft tissue and may provide a practical approach for clinical application. The fanbeam collimator demonstrated high resolution SPECT performance. CONCLUSION: These results suggest this system may provide an effective and practical approach to TCT imaging for nonuniform attenuation correction on a three-headed SPECT camera.

Quantitative imaging of iodine-131 distributions in brain tumors with pinhole SPECT: a phantom study.

UNLABELLED: A method of quantitatively imaging 131I distributions in brain tumors from intratumoral administration of activity was developed and investigated using pinhole SPECT of brain tumor phantoms. METHODS: Pinhole SPECT sensitivity and resolution were characterized using 131I point-source acquisitions with high-resolution lead (1.4-mm diameter aperture) and tungsten (1.0-mm diameter aperture) pinhole inserts. SPECT scans were obtained from brain tumor phantoms in a water-filled cylinder. The tumor phantoms consisted of spheres filled with an 131I solution to model intratumoral administration of radiolabeled monoclonal antibodies. Two spheres were 20.5 and 97 ml, and two other concentric spheres modeled a tumor with a high-activity shell (71.5 ml) and a low-activity core (21 ml). The collimator focal length was 16 cm and the distance from the pinhole to the center of rotation was 13 cm. The filtered backprojection reconstruction algorithm incorporated scatter and attenuation compensation. SPECT tumor activities and concentrations were estimated using scaling factors from reference point-source scans. RESULTS: System sensitivities for point sources at the center of rotation were 28.4 cts/sec(-1) MBq(-1) (lead insert) and 13.6 cts/sec(-1) MBq(-1) (tungsten insert). SPECT resolutions (FWHM) at the center of rotation were 8.1-11.9 mm (lead) and 6.7-10.3 mm (tungsten). Total tumor activity estimates from SPECT were within 17% of the true activities. SPECT activity concentration estimates in small regions of interest (ROIs) averaged -20% for the 20.5-ml sphere, -11% for the 97-ml sphere, -39% for the shell and +20% for the core of the shell-core phantom. Activity spillover due to limited spatial resolution and the tails of the system response functions biased the estimates. The shell-to-core activity concentration ratio of 4.1 was better estimated with the tungsten insert (2.3) than with the lead insert (1.9) due to better resolution. CONCLUSION: Pinhole SPECT is a promising technique for imaging and quantifying total 131I activity in regions the size of brain tumors. Relative errors were greater for activity concentration estimates in small ROIs than for total activity estimates.

Effect of registration errors between transmission and emission scans on a SPECT system using sequential scanning.

UNLABELLED: The purpose of this study was to evaluate the effects of patient motion on nonuniform attenuation correction of cardiac SPECT when the transmission and emission scans were performed sequentially. By using a sequential protocol rather than doing the scans simultaneously, contamination from the emission scan into the transmission scan could be eliminated, but registration of the two scans become a concern. METHODS: Transmission and emission scans were acquired using both an anthropomorphic thorax phantom containing a cardiac insert and a human volunteer. The types of motion considered were transverse shifts, axial shifts and rotations that occur in the time period between the transmission and emission scan. For this study, the various types of motion were simulated in the data. Both the transmission and emission data were reconstructed using filtered backprojection. A single-iteration Chang algorithm, modified for nonuniform attenuation correction, was used to further process the emission data. To evaluate the effects of motion errors, circumferential profiles, all normalized to the same scale, were generated for each case. The cardiac images reconstructed using registered data were considered references. Error profiles were generated by subtracting misaligned images from the reference and then normalizing the difference by the reference. For comparison purposes, an error profile was generated for the case in which no attenuation correction was performed. RESULTS: It was found that, for transverse and axial shifts of 2.9 cm, the normalized myocardial SPECT activity was decreased in certain regions of the heart by 20%-35%. For a 12 degrees rotational shift, the error was on the order of 10%-20%, compared to a normalized variation of 20%-25% in the image with no attenuation correction. CONCLUSION: The results indicate that registration errors of 2-3 cm can seriously affect image quality in both the phantom and human images.

Fully Bayesian estimation of Gibbs hyperparameters for emission computed tomography data.

In recent years, many investigators have proposed Gibbs prior models to regularize images reconstructed from emission computed tomography data. Unfortunately, hyperparameters used to specify Gibbs priors can greatly influence the degree of regularity imposed by such priors and, as a result, numerous procedures have been proposed to estimate hyperparameter values from observed image data. Many of these procedures attempt to maximize the joint posterior distribution on the image scene. To implement these methods, approximations to the joint posterior densities are required, because the dependence of the Gibbs partition function on the hyperparameter values is unknown. In this paper, we use recent results in Markov chain Monte Carlo (MCMC) sampling to estimate the relative values of Gibbs partition functions and using these values, sample from joint posterior distributions on image scenes. This allows for a fully Bayesian procedure which does not fix the hyperparameters at some estimated or specified value, but enables uncertainty about these values to be propagated through to the estimated intensities. We utilize realizations from the posterior distribution for determining credible regions for the intensity of the emission source. We consider two different Markov random field (MRF) models-the power model and a line-site model. As applications we estimate the posterior distribution of source intensities from computer simulated data as well as data collected from a physical single photon emission computed tomography (SPECT) phantom.

Approximate 3D iterative reconstruction for SPECT.

Compared with slice-by-slice approaches for SPECT reconstruction, three-dimensional iterative methods provide a more accurate physical model and an improved SPECT image. Clinical application of these methods, however, is limited primarily to their computational demands. This paper investigates the methods for approximate 3D iterative reconstruction that greatly reduce this demand by excluding from the reconstruction the smaller magnitude elements of the system matrix. A new method is described which is designed to control the resulting bias in the SPECT image for a given reduction in computation. The approximate methods were compared to fully 3D iterative reconstruction in terms of SPECT image bias and visual quality. All methods were incorporated into the ML-EM algorithm and applied to data from 3D mathematical and experimental brain phantoms. The SPECT images reconstructed by the approximate methods exhibited a positive bias throughout the image that was in general smaller with the new method (in the rage of 2%-6%). The bias was smallest in locally hot regions and largest in locally cold regions. The high quality brain phantom images demonstrated the capability of the new method in realistic imaging contexts. The time per iteration for an entire 3D brain phantom on a modern workstation using the approximate 3D method was 7.0 s.

Evaluation of SPECT quantification of radiopharmaceutical distribution in canine myocardium.

UNLABELLED: This study evaluates the quantitative accuracy of SPECT for in vivo distributions of 99mTc radiopharmaceuticals using fanbeam (FB) and parallel-beam (PB) collimators and compares uniform and nonuniform attenuation correction methods in terms of quantitative accuracy. METHODS: SPECT quantification of canine myocardial radioactivity was performed followed by well counter measurements of extracted myocardial tissue samples. Transmission scans using a line source and an FB collimator were performed to generate nonuniform attenuation maps of the canine thorax. Emission scans with two energy windows were acquired. Images were reconstructed using a filtered backprojection algorithm, with a dual-window scatter subtraction combined with either no attenuation compensation or single iteration Chang attenuation compensation based on an uniform attenuation map (mu = 0.152 cm-1) or the nonuniform transmission map. RESULTS: The measured mean counts from the SPECT images were converted to radionuclide concentrations (MBq/g) using a standard source calibration and were compared with those obtained using the well counter. CONCLUSION: The experimental results demonstrate that, compared with well counter values, the in vivo distributions of 99mTc were most accurately determined in FB and PB SPECT reconstructions with nonuniform attenuation compensation, under-estimated without attenuation compensation and overestimated with uniform attenuation compensation.

Quantitation of 211At in small volumes for evaluation of targeted radiotherapy in animal models.

We have evaluated SPECT and two planar imaging methods, geometric mean (GM) and buildup factor (BF), for their potential to quantitate in vivo 211At distributions in rat spinal subarachnoid spaces using phantom studies. The use of medium-energy collimators and the small diameter (3 mm) of the subarachnoid space complicate quantitation. Net activities from distributions in various backgrounds were obtained using a large region of interest with background subtraction. Results showed quantitation accuracy within 10% for SPECT and BF in low backgrounds increasing to 25% at higher background levels while GM errors ranged from 20 to 45%. We have also obtained images of [211At]astatide distributions, administered intrathecally, in rats.

Volume and activity quantitation with iodine-123 SPECT.

UNLABELLED: The goals of this study were to investigate the effect of septal penetration on 123I SPECT activity quantitation using low-energy, high-resolution collimators, and to evaluate a semi-automatic method for measuring volume and activity of 123I distribution with SPECT. METHODS: Data were acquired from experimental phantoms containing spheres filled with a high-purity 123I solution. The penetration study compared the reconstructed activity of a 3.4-cm diameter sphere with and without the presence of surrounding activity. In the study of volume and activity quantitation, three different size spheres (diameters of 1.8 cm, 2.8 cm and 3.4 cm) were imaged in three different sphere-to-background (S:B) 123I concentration ratios (2.5, 5 and 10) with low-energy collimators. The filtered backprojection reconstruction method was used with compensation for scatter, attenuation and detector response. Volume and activity measurements were obtained from the SPECT image using a semiautomatic gradient technique which estimates the location of the sphere/boundary in three dimensions. RESULTS: With the low-energy collimator, there was only a small (< 2%) increase in the measured activity of the sphere when surrounding activity was present. The measured volume for the two largest spheres was within 5% of the true volume for all S:B ratios. The activity measurement of these spheres was consistently underestimated by 20%-25% but suggested that the accuracy could be improved with calibration. For the smallest sphere, the volume was grossly overestimated and only at the 10 S:B ratio was the activity measured reasonably accurately (< 20%). CONCLUSIONS: The low-energy collimators used in this study are suitable for quantitative 123I SPECT. Accurate SPECT volume and activity quantitation of 123I distribution can be achieved by semiautomatic means at clinical count densities for objects as small as 2.8 cm in diameter and reasonable activity quantitation is possible for smaller objects with an S:B ratio of at least 10.

A 3D model of non-uniform attenuation and detector response for efficient iterative reconstruction in SPECT.

A 3D physical model for iterative reconstruction in SPECT has been developed and applied to experimental data. The model incorporates non-uniform attenuation using reconstructed transmission CT data and distance-dependent detector response based on response function measurements over a range of distances from the detector. The 3D model has been implemented in a computationally efficient manner with practical memory requirements. The features of the model that provide efficiency are described including a new region-dependent reconstruction (RDR) technique. With RDR, filtered backprojection is used to reconstruct areas of the image of minimal clinical importance, and the result is used to supplement the iterative reconstruction of the clinically important areas of the image. The 3D model was incorporated into the maximum likelihood-expectation maximization (ML-EM) reconstruction algorithm and tested in three phantom studies--a point source, a uniform cylinder, and an anthropomorphic thorax--and a patient 9Tc(m) sestamibi study. Reconstructed images with the 3D method exhibited excellent noise and resolution characteristics. With the sestamibi data, the RDR technique produced essentially the conventional ML-EM estimate in the cardiac region with substantial time savings.

An evaluation of lesion detectability with cone-beam, fanbeam and parallel-beam collimation in SPECT by continuous ROC study.

UNLABELLED: To evaluate lesion detectability for clinical evaluation of cone-beam (CB), fanbeam (FB) and parallel-beam (PB) collimator sensitivity, experimentally acquired phantom data were used to assess the advantage of CB collimation over conventional collimation. METHODS: Lesion detectability with CB, FB and PB collimation in SPECT was compared using a three-dimensional brain phantom and continuous receiver operating characteristic (CROC) analysis. A simulated cold lesion was located near the posterior portion of the thalamus. High count density scans of this phantom were acquired with CB, FB and PB collimators with similar resolution. These projections were scaled to count levels which reflected the measured sensitivities of the three collimators. Computer-generated Poisson noise was added to the projections to produce uncorrelated data sets. Images were reconstructed using a filtered backprojection algorithm. All reconstructions used a Hann filter with multiplicative attenuation correction. Each of seven trained observers viewed 288 sets of images and indicated the certainty of perceiving a cold lesion at a specified location by a rating of 0-100. Each image set contained four adjacent slices centered on the lesion to minimize partial volume effects. The program LABROC4 was used to fit CROC curves to individual observers' ratings. A t-test for paired data was performed on the individual areas. RESULTS: The average areas (standard deviations) under CROC curves for CB, FB and PB were 0.89 (0.03), 0.83 (0.05) and 0.76 (0.04), respectively. The differences of the areas were statistically significant with all two-tailed p values < 0.02. CONCLUSION: These results demonstrate that cold lesions in the posterior portion of the thalamus are best detected by images obtained using CB followed by FB and PB collimation.

Fast transmission CT for determining attenuation maps using a collimated line source, rotatable air-copper-lead attenuators and fan-beam collimation.

We describe a technique using a line source and a rotatable air-copper-lead assembly to acquire gamma transmission computed tomographic (TCT) data for determining attenuation maps to compensate SPECT emission scans. The technique minimizes problems associated with discriminating 99mTc transmission and 201Tl emission photons and requires only a modest increase in total study time. A 99mTc line source and a stacked foil ("multislat") collimator are placed near the focal line of a fan-beam collimator (114 cm focal length) mounted on one detector of a triple-camera SPECT system. We acquired TCT data of plastic rod and anthropomorphic thorax phantoms to investigate the capability of the line source and rotatable air-copper-lead attenuators to determine attenuation maps. The data were acquired with and without 5.4 MBq (145 microCi) of 201Tl placed in the myocardial chamber of the thorax phantom. Phantoms also were scanned using a curved transmission slab source mounted to a parallel-hole collimator. Fan-beam TCT images have improved resolution compared with parallel-beam TCT images. Two patient scans also were performed to evaluate the clinical usefulness of fan-beam TCT. The rotatable air-copper-lead attenuator method eliminates contamination of emission data by transmission photons and reduces spill-over of emission data into the transmission energy window for some cases. Results show the feasibility of using fast, sequential or interlaced transmission scans of a line source within a rotatable air-copper-lead attenuator assembly to obtain accurate attenuation maps for SPECT attenuation compensation.

Simultaneous compensation for attenuation, scatter and detector response for SPECT reconstruction in three dimensions.

A three-dimensional reconstruction method for simultaneous compensation of attenuation, scatter and distance-dependent detector response for single photon emission computed tomography is described and tested by experimental studies. The method determines the attenuation factors recursively along each projection ray starting at the intersected source voxel closest to the detector. The method substracts the scatter energy window data from the primary energy window data for scatter compensation. The detector response is modelled to be spatially invariant at a constant distance from the detector. The method convolves source distribution with the modelled response function to compensate for the smoothed by use of a non-uniform entropy prior to searching for the maximum a posteriori probability solution. The method was tested using projections acquired from a chest phantom by a three-headed detector system with parallel hole collimators. An improvement was shown in image noise, recognition of object sizes and shapes, and quantification of concentration ratios.

An evaluation of maximum likelihood-expectation maximization reconstruction for SPECT by ROC analysis.

A ROC study was performed in order to evaluate whether the maximum likelihood expectation maximization (ML-EM) reconstruction algorithm improves diagnostic performance compared to the conventional filtered backprojection method in SPECT. Several implementations of the algorithm were tested including 25 and 50 iteration stopping points, with and without nonuniform attenuation compensation, and with and without Metz filtering. Filtered backprojection was with Metz filter and without attenuation compensation. The test data were computer simulated to model cardiac 201Tl SPECT. The data incorporated the effects of nonuniform attenuation, distance-dependent collimator response, and scatter. Patient CT images provided realistic anatomy and attenuation information for the data simulation. Four observers each viewed 120 images for each of the reconstruction methods. Lesion detectability with ML-EM increased with Metz filtering and decreased with nonuniform attenuation compensation. The best MIL-EM implementation, 50 iterations with Metz filtering and without attenuation compensation, was not statistically better than filtered backprojection.

Quantitative SPECT reconstruction of iodine-123 data.

Many clinical and research studies in nuclear medicine require quantitation of iodine-123 (123I) distribution for the determination of kinetics or localization. The objective of this study was to implement several reconstruction methods designed for single-photon emission computed tomography (SPECT) using 123I and to evaluate their performance in terms of quantitative accuracy, image artifacts, and noise. The methods consisted of four attenuation and scatter compensation schemes incorporated into both the filtered backprojection/Chang (FBP) and maximum likelihood-expectation maximization (ML-EM) reconstruction algorithms. The methods were evaluated on data acquired of a phantom containing a hot sphere of 123I activity in a lower level background 123I distribution and nonuniform density media. For both reconstruction algorithms, nonuniform attenuation compensation combined with either scatter subtraction or Metz filtering produced images that were quantitatively accurate to within 15% of the true value. The ML-EM algorithm demonstrated quantitative accuracy comparable to FBP and smaller relative noise magnitude for all compensation schemes.

Determination of the optimum filter function for SPECT imaging.

An observer study was performed in order to evaluate several filters used in SPECT imaging. The filters were applied to the simulated projection data of a uniform activity density cylinder which contained a cold, spherical lesion, 2 cm in diameter. The data incorporated the effects of the detector and scatter response functions, photon attenuation, and noise. Reconstructed transaxial images were used in 2AFC and ROC observer studies testing lesion detectability. In the 2AFC experiment, the Hanning filter scored lowest and did not show a optimum cutoff frequency. The Butterworth filter performed better and showed a well-defined optimum cutoff frequency at 0.15 cycles/pixel. The Metz filter performed as well as the optimum Butterworth but did not show an optimum power factor. In the ROC study, a high power Metz filter demonstrated an ROC curve of lower Az index and different shape from a lower power Metz filter and the optimum Butterworth filter.