Search PubMed⌕ Search

Biomedical subjects

B M Tsui

Publications and source records attributed to B M Tsui.

At least 37 records · Page 2Linked to original sources
PubMed1995

Quantitative SPECT imaging: a review and recommendations by the Focus Committee of the Society of Nuclear Medicine Computer and Instrumentation Council.

This article is a review of the physics principles, instrumentation and reconstruction methods behind SPECT imaging. Particular attention is paid to the mechanisms that can significantly affect the accuracy of a SPECT image. We describe instrumentation advances and reconstruction methods used to correct images to improve image quality and produce quantitative images. The clinical importance of improved image quality and quantitation are also reviewed.

Algorithms↗
PubMed1994

Quantitative single-photon emission computed tomography: basics and clinical considerations.

Although quantitative single-photon emission computed tomography (SPECT) has been the goal of much research effort for a number of years, only recently has it received wide interest, especially for clinical applications. It has been increasingly recognized that the achievement of quantitative SPECT will increase the accuracy of measurements, such as dimensions of specific regions of interest, absolute amount of radioactivity, and dosimetry calculations, and substantially reduce reconstruction image artifacts and distortions, thus, greatly improving clinical diagnosis. This article provides a review of the definition of terms, major factors affecting SPECT quantitation and their degrading effects on SPECT image quality, and methods to compensate for these effects. Compensation methods include those that make certain approximations for ease of implementation and those that provide more accurate compensation by modeling the imaging process more exactly, usually at the cost of increased complexity and computational requirements. Different reconstruction and compensation methods may be compared through the use of phantom cardiac and brain SPECT studies. The clinical efficacy of the methods may be demonstrated by applying them to a clinical thallium-201 myocardial perfusion SPECT study. The results clearly demonstrate that, by modeling the imaging process and/or image degrading factors three-dimensionally, quantitative reconstruction and compensation methods provide the best image quality and quantitative accuracy. Important research efforts and developmental work being conducted currently to bring quantitative SPECT into routine clinical use are also discussed.

Humans↗
PubMed1994

An analytical reconstruction algorithm for multifocal converging-beam SPECT.

Multifocal converging-beam collimation has been suggested for cardiac SPECT imaging to increase sensitivity over the heart without truncation of the activity distribution in the chest. In this study, an analytical reconstruction algorithm is derived for multifocal fan-beam and multifocal cone-beam tomography. In the algorithm, the projection data are differently weighted and filtered, depending on the distance from the detector. For a given image point, the set of filtered data corresponding to the distance between this point and detector is backprojected to determine the pixel value of the point. Thus, the backprojection is done only once at each projection view. To evaluate the algorithm, simulation studies are performed using a 3D Defrise slab phantom without considering photon attenuation and scatter, detector response, and statistical noise. Reconstructed images demonstrate that reasonable quality can be achieved with a modest focal-length variation rate and with a small radius of rotation. A collimator with a focal length increasing quickly near its centre provides better quality in the image region distant from the central plane of the cone geometry, but produces more severe artifacts at the centre of the reconstructed image, compared to a collimator with an initially slowly varying focal length.

Algorithms↗
PubMed1994

The importance and implementation of accurate 3D compensation methods for quantitative SPECT.

The purpose of this study was to investigate the importance of 2D versus 3D compensation methods in SPECT. The compensation methods included in the study addressed two important degrading factors, namely attenuating and collimator-detector response in SPET. They can be divided into two general categories. The conventional methods are based on the filtered backprojection algorithm, the Chang algorithm for attenuation compensation and the Metz filter for detector response compensation. These methods, which were computationally efficient, could only achieve approximate compensation due to the assumptions made. The quantitative compensation methods provide accurate compensation by modelling the degrading effects at the expense of large computational requirements. Both types of compensation methods were implemented in 2D and 3D reconstructions. The 2D and 3D reconstruction/compensation methods were evaluated using data from simulation of brain and heart, and patient thallium SPECT studies. Our results demonstrate the importance of compensation methods in improving the quality and quantitative accuracy of SPECT images and the relative effectiveness of the different 2D and 3D reconstruction/compensation methods. We concluded that 3D implementation of the quantitative compensation methods provides the best SPECT image in terms of quantitative accuracy, spatial resolution, and noise at a cost of high computational requirements.

Algorithms↗
PubMed1994

Noise properties of the EM algorithm: I. Theory.

The expectation-maximization (EM) algorithm is an important tool for maximum-likelihood (ML) estimation and image reconstruction, especially in medical imaging. It is a non-linear iterative algorithm that attempts to find the ML estimate of the object that produced a data set. The convergence of the algorithm and other deterministic properties are well established, but relatively little is known about how noise in the data influences noise in the final reconstructed image. In this paper we present a detailed treatment of these statistical properties. The specific application we have in mind is image reconstruction in emission tomography, but the results are valid for any application of the EM algorithm in which the data set can be described by Poisson statistics. We show that the probability density function for the grey level at a pixel in the image is well approximated by a log-normal law. An expression is derived for the variance of the grey level and for pixel-to-pixel covariance. The variance increases rapidly with iteration number at first, but eventually saturates as the ML estimate is approached. Moreover, the variance at any iteration number has a factor proportional to the square of the mean image (though other factors may also depend on the mean image), so a map of the standard deviation resembles the object itself. Thus low-intensity regions of the image tend to have low noise. By contrast, linear reconstruction methods, such as filtered back-projection in tomography, show a much more global noise pattern, with high-intensity regions of the object contributing to noise at rather distant low-intensity regions. The theoretical results of this paper depend on two approximations, but in the second paper in this series we demonstrate through Monte Carlo simulation that the approximations are justified over a wide range of conditions in emission tomography. The theory can, therefore, be used as a basis for calculation of objective figures of merit for image quality.

Algorithms↗
PubMed1994

Noise properties of the EM algorithm: II. Monte Carlo simulations.

In an earlier paper we derived a theoretical formulation for estimating the statistical properties of images reconstructed using the iterative ML-EM algorithm. To gain insight into this complex problem, two levels of approximation were considered in the theory. These techniques revealed the dependence of the variance and covariance of the reconstructed image noise on the source distribution, imaging system transfer function, and iteration number. In this paper a Monte Carlo approach was taken to study the noise properties of the ML-EM algorithm and to test the predictions of the theory. The study also served to evaluate the approximations used in the theory. Simulated data from phantoms were used in the Monte Carlo experiments. The ML-EM statistical properties were calculated from sample averages of a large number of images with different noise realizations. The agreement between the more exact form of the theoretical formulation and the Monte Carlo formulation was better than 10% in most cases examined, and for many situations the agreement was within the expected error of the Monte Carlo experiments. Results from the studies provide valuable information about the noise characteristics of ML-EM reconstructed images. Furthermore, the studies demonstrate the power of the theoretical and Monte Carlo approaches for investigating noise properties of statistical reconstruction algorithms.

Algorithms↗
PubMed1994

A filtering technique to compensate for detector response in converging-beam SPECT reconstruction.

In this study a non-stationary filtering technique was developed to compensate for spatially variant detector response in fan- and cone-beam SPECT. First the frequency-distance relation for fan- and cone-beam geometries was derived from their parallel-beam counterpart, and then used to compute the Fourier transform of the detector response that was used as the MTF of a Metz filter. The filter was applied to the blurred data before reconstruction. The unique feature of the derived MTF is that it depends on source location, so the filtering is spatially variant. Results of evaluation studies demonstrate that the ratio between the tangential and radial FWHMS of the reconstructed point response function obtained from our filtering technique is closer to one and varies less with source location as compared to the ratio obtained without compensation or using Metz filtering with a stationary detector response. This improvement is mainly because the tangential FWHM is increased to match the radial FWHM better. With a higher order of Metz filter, both tangential and radial FWHMS are slightly smaller but their ratio deviates more from one compared to those of lower orders. In conclusion, our filtering technique provides an effective means to compensate for detector response mainly in terms of reducing anisotropy of the point response function.

Algorithms↗
PubMed1994

Improving the convergence of iterative filtered backprojection algorithms.

Several authors have proposed variations of the iterative filtered backprojection (IFBP) reconstruction algorithms claiming fast initial convergence rates. We have found that these algorithms are trying to minimize an unusual squared-error criterion in a suboptimal way. As a result, existing IFBP algorithms are inefficient in the minimization of the criterion, and may become unstable at higher iteration numbers. We show that existing IFBP algorithms can be modified to use the steepest descent technique by simply optimizing the step size at each iteration. Further gains in convergence rates can be achieved with conjugate gradient IFBP algorithms derived from the same criterion. The steepest descent and conjugate gradient IFBP algorithms are guaranteed to converge, unlike some IFBP algorithms, and will do so in fewer iterations than existing IFBP algorithms.

Algorithms↗
PubMed1993

Direct cone beam SPECT reconstruction with camera tilt.

A filtered backprojection (FBP) algorithm is derived to perform cone beam (CB) single-photon emission computed tomography (SPECT) reconstruction with camera tilt using circular orbits. This algorithm reconstructs the tilted angle CB projection data directly by incorporating the tilt angle into it. When the tilt angle becomes zero, this algorithm reduces to that of Feldkamp. Experimentally acquired phantom studies using both a two-point source and the three-dimensional Hoffman brain phantom have been performed. The transaxial tilted cone beam brain images and profiles obtained using the new algorithm are compared with those without camera tilt. For those slices which have approximately the same distance from the detector in both tilt and non-tilt set-ups, the two transaxial reconstructions have similar profiles. The two-point source images reconstructed from this new algorithm and the tilted cone beam brain images are also compared with those reconstructed from the existing tilted cone beam algorithm which requires rebinning the projection data. These comparisons demonstrate that the new algorithm, compared with the existing tilted cone beam algorithm, can provide better image contrast and improved image resolution.

Algorithms↗
PubMed1993

A generalized Gibbs prior for maximum a posteriori reconstruction in SPECT.

In this paper, we introduce a Gibbs prior for use in MAP (maximum a posteriori) reconstruction in SPECT. This new prior, with three parameters, is able to approximate the results of previously-proposed priors with two parameters, as well as a continuum of others. Also, it allows the user increased flexibility in selecting the properties to be emphasized in the final reconstructed image estimate. The additional flexibility offered by the new prior is important in addressing the problem of selecting a prior and its associated parameters in a clinical situation. The paper demonstrates the importance of the derivative potential function (DPF) of the Gibbs distribution in determining which properties will be emphasized in the iterated image estimates. The effects of each of the three parameters are demonstrated on reconstructions from acquired SPECT data. We conclude that the parameters must be chosen carefully with consideration for the object distribution and the relative requirements for low-contrast detail, smoothing and edge sharpness in the reconstructed image.

Algorithms↗
PubMed1993

Evaluation of cardiac cone-beam single photon emission computed tomography using observer performance experiments and receiver operating characteristic analysis.

RATIONALE AND OBJECTIVES: Single photon emission computed tomography (SPECT) with a cone-beam collimator improves the trade-off between detection efficiency and spatial resolution for cardiac imaging. However, acquisitions using orbits where the focus remains in a plane do not provide sufficient data for exact reconstruction. In the current study the authors evaluate the clinical utility of planar-orbit cone-beam SPECT in detecting a simple myocardial defect. METHODS: Observer performance experiments compared high-resolution cone-beam with same-resolution parallel-hole and fan-beam collimator designs in myocardial defect detection using a computer-simulated cardiac model. The uptake of Thallium-201 in the myocardium and other tissue organs was modeled by a mathematical three-dimensional upper torso phantom from which physically realistic projections were simulated. Eight observers viewed reconstructed transaxial images from the three collimator designs and indicated the certainty with which they detected a Gaussian-shaped defect at a specified location. RESULTS: The area under the receiver operating characteristic curve indicated that the cone-beam design, regardless of slice position, was superior to the fan-beam, which in turn was superior to the parallel-hole design for the specified detection task. CONCLUSIONS: The observer study demonstrated that reconstruction artifacts resulting from insufficient data sampling do not hinder obtaining improved diagnostic information from planar-orbit cone-beam cardiac SPECT images compared to conventional cardiac SPECT using parallel-hole and fan-beam collimators.

Adult↗
PubMed1992

Simultaneous acquisition of emission and transmission data for improved thallium-201 cardiac SPECT imaging using a technetium-99m transmission source.

Transmission computed tomography (TCT) data provides useful complementary information to single-photon emission computed tomography (SPECT) reconstructions, especially for cardiac studies. In particular, TCT data has been used to correct for nonuniform attenuation in the chest. Typically the transmission data are acquired in a separate acquisition, but simultaneous acquisition is preferable both to save time and to avoid difficulties involved with registration. In this work, we present a technique for simultaneously acquiring 201Tl SPECT and TCT data using a 99mTc sheet source that requires only minor equipment modifications. The use of these isotopes results in cross-contamination of the emission and transmission data. We present a practical technique to compensate for this contamination using post-acquisition image processing. This technique was evaluated by performing phantom and patient studies. The resulting images compare well with data obtained from separate emission and transmission studies.

Heart↗
PubMed1992

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.

Algorithms↗
PubMed1990

The geometric transfer function for cone and fan beam collimators.

Geometric response functions are derived for both cone and fan beam collimators for the scintillation camera. The formulation is based on an effective response function which is determined by the geometric response of a single hole. The technique provides an accurate description of the spatial resolution by characterising the complete geometric response function which includes the effects of the shape and orientation of the collimator holes. The theoretical formulation was used to design a fan beam collimator for SPECT imaging and was shown to agree well with the experimental results.

Equipment Design↗
PubMed1990

SPECT dual-energy-window Compton correction: scatter multiplier required for quantification.

The dual-energy window Compton-scattering correction technique is defined here especially for accurate quantification of focal regions having higher than average uptake. The quantification is relative to a known-activity reference source. The scatter multiplier ("k" value) is determined for a radioactive 99mTc sphere on or off the axis of a cylinder containing water with or without background. Both maximum likelihood and filtered-backprojection reconstruction are employed. Either projections or tomograms are corrected. With tight regions of interest, there is a tendency for the requisite "k" value to be slightly lower as the diameter of the cylinder is increased. Neither sphere location nor background perturbs "k", however, so a constant value is a good, first approximation. Then a two-sphere validation test yields an accuracy of 8% with subtracted-tomograms ("k" = 1.30) and 2% with subtracted-projections ("k" = 1.20). With a reference-source region of interest which is four times larger, "k" is reduced and also now depends on background. Although equivalent quantitatively, maximum likelihood is preferable to filtered backprojection with Chang attenuation correction since it produces a less-noisy image.

Algorithms↗
PubMed1989

Multiparameter extrapolation of biodistribution data between species.

Values of an inaccessible biological parameter in man may be predicted from values measured in animals by correlating with a parameter accessible in both species, such as body weight, energy production, excretion rate, etc. Predicting toxic effects, from environmental chemicals, of therapeutic doses for drug administration and of radiation absorbed dose from medical and environmental radioactivity depends on the rationalization of relationships between concentration and time when scaling to humans from animal data. For example, the retention of 99mTc, injected intravenously as pertechnetate, reaches 10% in the mouse at about 1 d, but this level occurs in humans at about 7 d. Making a simultaneous transformation between two species for the concentration and time variables by using a method of least-squares fitting, we have derived a series of transformation factors for several species. When correlated with a biological parameter such as body weight, these factors can be used to yield predicted values that are in good agreement with measured values. This system may be used with any related variables, making it useful for predicting other types of biological data.

Animals↗