Neural network reconstruction of single-photon emission computed tomography images

An artificial neural network (ANN) trained on high-quality medical tomograms or phantom images may be able to learn the planar data-to-tomographic image relationship with very high precision. As a result, a properly trained ANN can produce comparably accurate image reconstruction without the high computational cost inherent in some traditional reconstruction techniques. We have previously shown that a standard backpropagation neural network can be trained to reconstruct sections of single photon emission computed tomography (SPECT) images based on the planar image projections as inputs. In this study, we present a method of deriving activation functions for a backpropagation ANN that make it readily trainable for full SPECT image reconstruction. The activation functions used for this work are based on the estimated probability density functions (PDFs) of the ANN training set data. The statistically tailored ANN and the standard sigmoidal backpropagation ANN methods are compared both in terms of their trainability and generalization ability. The results presented show that a statistically tailored ANN can reconstruct novel tomographic images of a quality comparable with that of the images used to train the network. Ultimately, an adequately trained ANN should be able to properly compensate for physical photon transport effects, background noise, and artifacts while reconstructing the tomographic image.     

Whole-body single-photon emission computed tomography using dual, large-field-of-view scintillation cameras.

A whole-body single-photon emission computed tomography system (SPECT) consisting of two large-field-of-view scintillation cameras mounted on a rotatable gantry, a minicomputer and a display station has been designed, constructed and evaluated. In its usual mode of operation, eleven contiguous transverse sections, each 12.5 or 25 mm thick, are reconstructed from projection data acquired during a single, continuous 360 degree rotation lasting from 2 to 22 min. A generalised filtered and weighted backprojection algorithm is used to reconstruct data obtained with conventional parallel-hole collimators in the case of body scanning, or with specially designed fan beam collimators in the case of centrally positioned organs. A simple, yet effective, correction is used to compensate for the effects of gamma ray attenuation within the patient. In addition to providing transverse section images, the system is capable of simultaneous acquisition of opposed conventional scintigrams, the reconstruction of longitudinal section images, and the acquisition of gated cardiac transverse sections. Resolutions in the reconstructed images are typically 15 mm for body scans and 11 mm for brain scans, with only slight variations in sensitivity and resolution within the image. Phantoms and clinical data demonstrate that the SPECT system generates high quality section images while maintaining most of the flexibility of normal scintillation cameras, with the added advantage of dual heads.     

Iterative reconstruction techniques in emission computed tomography

In emission tomography statistically based iterative methods can improve image quality relative to analytic image reconstruction through more accurate physical and statistical modelling of high-energy photon production and detection processes. Continued exponential improvements in computing power, coupled with the development of fast algorithms, have made routine use of iterative techniques practical, resulting in their increasing popularity in both clinical and research environments. Here we review recent progress in developing statistically based iterative techniques for emission computed tomography. We describe the different formulations of the emission image reconstruction problem and their properties. We then describe the numerical algorithms that are used for optimizing these functions and illustrate their behaviour using small scale simulations.     

Critical examination of the uniformity requirements for single-photon emission computed tomography

It is generally recognized that single-photon emission computed tomography (SPECT) imposes very stringent requirements on gamma camera uniformity to prevent the occurrence of ring artifacts. The purpose of this study was to examine the relationship between nonuniformities in the planar data and the magnitude of the consequential ring artifacts in the transaxial data, and how the perception of these artifacts is influenced by factors such as reconstruction matrix size, reconstruction filter, and image noise. The study indicates that the relationship between ring artifact magnitude and image noise is essentially independent of the acquisition or reconstruction matrix sizes, but is strongly dependent upon the type of smoothing filter applied during the reconstruction process. Furthermore, the degree to which a ring artifact can be perceived above image noise is dependent on the size and location of the nonuniformity in the planar data, with small nonuniformities (1-2 pixels wide) close to the center of rotation being less perceptible than those further out (8-20 pixels). Small defects or nonuniformities close to the center of rotation are thought to cause the greatest potential corruption to tomographic data. The study indicates that such may not be the case. Hence the uniformity requirements for SPECT may be less demanding than was previously thought.     

Iterative concurrent reconstruction algorithms for emission computed tomography

Direct reconstruction techniques, such as those based on filtered backprojection, are typically used for emission computed tomography (ECT), even though it has been argued that iterative reconstruction methods may produce better clinical images. The major disadvantage of iterative reconstruction algorithms, and a significant reason for their lack of clinical acceptance, is their computational burden. We outline a new class of 'concurrent' iterative reconstruction techniques for ECT in which the reconstruction process is reorganized such that a significant fraction of the computational processing occurs concurrently with the acquisition of ECT projection data. These new algorithms use the 10-30 min required for acquisition of a typical SPECT scan to iteratively process the available projection data, significantly reducing the requirements for post-acquisition processing. These algorithms are tested on SPECT projection data from a Hoffman brain phantom acquired with a 2 x 10(5) counts in 64 views each having 64 projections. The SPECT images are reconstructed as 64 x 64 tomograms, starting with six angular views. Other angular views are added to the reconstruction process sequentially, in a manner that reflects their availability for a typical acquisition protocol. The results suggest that if T s of concurrent processing are used, the reconstruction processing time required after completion of the data acquisition can be reduced by at least 1/3T s.     

Simultaneous iterative reconstruction for emission and attenuation images in positron emission tomography

The quality of the attenuation correction strongly influences the outcome of the reconstructed emission scan in positron emission tomography. Usually the attenuation correction factors are calculated from the transmission and blank scan and thereafter applied during the reconstruction on the emission data. However, this is not an optimal treatment of the available data, because the emission data themselves contain additional information about attenuation: The optimal treatment must use this information for the determination of the attenuation correction factors. Therefore, our purpose is to investigate a simultaneous emission and attenuation image reconstruction using a maximum likelihood estimator, which takes the attenuation information in the emission data into account. The total maximum likelihood function for emission and transmission is used to derive a one-dimensional Newton-like algorithm for the calculation of the emission and attenuation image. Log-likelihood convergence, mean differences, and the mean of squared differences for the emission image and the attenuation correction factors of a mathematical thorax phantom were determined and compared. As a result we obtain images improved with respect to log likelihood in all cases and with respect to our figures of merit in most cases. We conclude that the simultaneous reconstruction can improve the performance of image reconstruction.     

Convolutional image reconstruction for quantitative single photon emission computed tomography

Two image reconstruction algorithms have been investigated. They are based on filtered backprojection, and are useful when the tissue attenuation is considered to be uniform in the object. The first method uses a weighted backprojection, the weighting factor being determined in such a way that the photon attenuation is compensated with low noise propagation. The parameters involved in the convolution kernel and the correction function were determined by a computer iteration program. The second method, which is a simplified version of the first, uses conventional backprojection, and takes a shorter computation time than the first method. The statistical noise of an image can be minimised by suitable positioning of the coordinate origin for the reconstruction. The theory of the two methods, their performance on statistical noise and some results of mathematical and experimental phantom studies are described.     

Accelerated median root prior reconstruction for pinhole single-photon emission tomography (SPET)

Pinhole collimation can be used to improve spatial resolution in SPET. However, the resolution improvement is achieved at the cost of reduced sensitivity, which leads to projection images with poor statistics. Images reconstructed from these projections using the maximum likelihood expectation maximization (ML-EM) algorithms, which have been used to reduce the artefacts generated by the filtered backprojection (FBP) based reconstruction, suffer from noise/bias trade-off: noise contaminates the images at high iteration numbers, whereas early abortion of the algorithm produces images that are excessively smooth and biased towards the initial estimate of the algorithm. To limit the noise accumulation we propose the use of the pinhole median root prior (PH-MRP) reconstruction algorithm. MRP is a Bayesian reconstruction method that has already been used in PET imaging and shown to possess good noise reduction and edge preservation properties. In this study the PH-MRP algorithm was accelerated with the ordered subsets (OS) procedure and compared to the FBP, OS-EM and conventional Bayesian reconstruction methods in terms of noise reduction, quantitative accuracy, edge preservation and visual quality. The results showed that the accelerated PH-MRP algorithm was very robust. It provided visually pleasing images with lower noise level than the FBP or OS-EM and with smaller bias and sharper edges than the conventional Bayesian methods.     

The peanut orbit: a modified elliptical orbit for single-photon emission computed tomography imaging

The front collimator surface of a conventional single-photon emission computed tomography (SPECT) detector system transcribing an elliptical orbit to approximate body contour could, under certain circumstances, penetrate the ellipse and make patient contact. The problem is associated both with the large front surface dimensions of tomographic cameras and the need to maintain the camera perpendicular to a radius through the axis of rotation. To aid in the development of an improved body contour orbit, software has been developed to simulate SPECT imaging systems. A major feature of the algorithms is a spatially calibrated graphic representation of the pallet, patient and orbit of the camera head. Based on computer simulations performed with this software, a modified elliptical orbit has been proposed for patient contour SPECT scanning on two different types of tomographic acquisition systems: a dual camera, rotating ring system and a single camera, multimotion stand system. This peanut orbit has the advantage of minimizing collimator patient distance, as does the elliptical orbit, but also compensates for camera motion that could potentially produce patient contact. Versions of the peanut orbit algorithms have been installed and are operational on clinical systems similar to those that were simulated. The ring-based gantry system (with cameras mounted on cantilevered arms) requires additional software to correct for translational shift in the field of view as a function of radial distance from the center of the system. This is done during uniformity correction. Standard unmodified backprojection software is used for reconstruction. The peanut orbit may represent an improved approach to body contour imaging in rotational SPECT.     

Breast cancer detection using neutron stimulated emission computed tomography: prominent elements and dose requirements

Neutron stimulated emission computed tomography (NSECT) is being developed to noninvasively determine concentrations of trace elements in biological tissue. Studies have shown prominent differences in the trace element concentration of normal and malignant breast tissue. NSECT has the potential to detect these differences and diagnose malignancy with high accuracy with dose comparable to that of a single mammogram. In this study, NSECT imaging was simulated for normal and malignant human breast tissue samples to determine the significance of individual elements in determining malignancy. The normal and malignant models were designed with different elemental compositions, and each was scanned spectroscopically using a simulated 2.5 MeV neutron beam. The number of incident neutrons was varied from 0.5 million to 10 million neutrons. The resulting gamma spectra were evaluated through receiver operating characteristic (ROC) analysis to determine which trace elements were prominent enough to be considered markers for breast cancer detection. Four elemental isotopes (133Cs, 81Br, 79Br, and 87Rb) at five energy levels were shown to be promising features for breast cancer detection with an area under the ROC curve (A(Z)) above 0.85. One of these elements--87Rb at 1338 keV--achieved perfect classification at 10 million incident neutrons and could be detected with as low as 3 million incident neutrons. Patient dose was calculated for each gamma spectrum obtained and was found to range from between 0.05 and 0.112 mSv depending on the number of neutrons. This simulation demonstrates that NSECT has the potential to noninvasively detect breast cancer through five prominent trace element energy levels, at dose levels comparable to other breast cancer screening techniques.     

Convergence of the maximum likelihood reconstruction algorithm for emission computed tomography

Convergence properties of the maximum likelihood estimator (MLE) for emission computed tomographic (ECT) image reconstruction are evaluated as a function of Poisson noise, precision of the assumed system resolution model and iteration number up to 10,000 iterations. In the ECT reconstruction problem, the photon-emitting source distribution is to be estimated from measurements of projections of the emitted photon flux. The MLE algorithm seeks a source distribution which will maximise the maximum likelihood function relating the estimated and the measured projections. A Monte Carlo model of the system transfer function of a single photon emission computed tomographic (SPECT) system allowed realistic projection data to be simulated from a known source distribution. Poisson noise was added to the Monte Carlo simulations. By using projection data from a known source distribution generated through a known system transfer function, we were able to simultaneously evaluate the convergence of both the projection estimations as well as the source distribution estimations. As predicted by theory, the estimates of the projections did continue to improve (or remain the same) for all combinations of Poisson noise (up to 10% RMS) and system resolution (+/- 10% of true value) tested. Convergence of source distribution estimates to the true value was found for up to 10,000 iterations only for low noise (0.1% RMS) with the correct resolution function. For all other combinations, there was some optimum iteration (between 30 and 400) after which the source estimate was degraded even though the estimate of the projections was improved.     

Modeling of polychromatic attenuation using computed tomography reconstructed images

This paper presents a procedure for estimating an accurate model of the CT imaging process including spectral effects. As raw projection data are typically unavailable to the end-user, we adopt a post-processing approach that utilizes the reconstructed images themselves. This approach includes errors from x-ray scatter and the nonidealities of the built-in soft tissue correction into the beam characteristics, which is crucial to beam hardening correction algorithms that are designed to be applied directly to CT reconstructed images. We formulate this approach as a quadratic programming problem and propose two different methods, dimension reduction and regularization, to overcome ill conditioning in the model. For the regularization method we use a statistical procedure, Cross Validation, to select the regularization parameter. We have constructed step-wedge phantoms to estimate the effective beam spectrum of a GE CT-I scanner. Using the derived spectrum, we computed the attenuation ratios for the wedge phantoms and found that the worst case modeling error is less than 3% of the corresponding attenuation ratio. We have also built two test (hybrid) phantoms to evaluate the effective spectrum. Based on these test phantoms, we have shown that the effective beam spectrum provides an accurate model for the CT imaging process. Last, we used a simple beam hardening correction experiment to demonstrate the effectiveness of the estimated beam profile for removing beam hardening artifacts. We hope that this estimation procedure will encourage more independent research on beam hardening corrections and will lead to the development of application-specific beam hardening correction algorithms.     

Simultaneous iterative reconstruction of emission and attenuation images in positron emission tomography from emission data only

For quantitative image reconstruction in positron emission tomography attenuation correction is mandatory. In case that no data are available for the calculation of the attenuation correction factors one can try to determine them from the emission data alone. However, it is not clear if the information content is sufficient to yield an adequate attenuation correction together with a satisfactory activity distribution. Therefore, we determined the log likelihood distribution for a thorax phantom depending on the choice of attenuation and activity pixel values to measure the crosstalk between both. In addition an iterative image reconstruction (one-dimensional Newton-type algorithm with a maximum likelihood estimator), which simultaneously reconstructs the images of the activity distribution and the attenuation coefficients is used to demonstrate the problems and possibilities of such a reconstruction. As result we show that for a change of the log likelihood in the range of statistical noise, the associated change in the activity value of a structure is between 6% and 263%. In addition, we show that it is not possible to choose the best maximum on the basis of the log likelihood when a regularization is used, because the coupling between different structures mediated by the (smoothing) regularization prevents an adequate solution due to crosstalk. We conclude that taking into account the attenuation information in the emission data improves the performance of image reconstruction with respect to the bias of the activities, however, the reconstruction still is not quantitative.     

Increased single-photon emission computed tomography image processing speed achieved in personal computers with memory-intensive algorithms

Recent dramatic reductions in the cost of computer random access memory (RAM) and the ability of newer microprocessors and associated personal computer operating systems to address large amounts of memory make novel strategies for high-speed image processing possible. We developed image processing algorithms that use this newly available memory to achieve increases in effective processing speed. These algorithms rely on the use of precomputed lookup tables to avoid repeated use of relatively expensive machine instructions, such as multiplications and divisions. Programs using this strategy to perform single photon emission computer tomography (SPECT) analysis were written in C and assembly language and tested on a Macintosh Quadra 950 (Apple Computer, Cupertino, CA) having 64 megabytes of RAM. The measured processing times are competitive with most dedicated nuclear medicine computers. A general implementation of such programs will allow personal computers to compete with dedicated imaging systems, at a substantial reduction in cost.     

An approximate reconstruction method for helical cone-beam differential phase-contrast computed tomography images

Differential phase-contrast computed tomography (DPC-CT) is a novel x-ray inspection method. Currently, DPC-CT reconstruction problems are solved by using parallel-beam, fan-beam and cone-beam algorithms. The above algorithms cannot show the internal structures of rod-shaped objects well enough for only reconstructing a few slices. The helical cone-beam algorithms have significant advantages for rod-shaped objects over other algorithms. Along with our numerical evaluation and verification, we report a PI-line-based approximate algorithm for helical cone-beam DPC-CT, which can be applied to reconstruct the refractive index decrement distribution of the samples directly from phase-contrast projection images. Simulations using a 3D Shepp-Logan phantom are carried out to verify the proposed algorithm. Reconstruction results show that the proposed algorithm can provide higher quality performance compared with the existing interpolation-based reconstruction algorithm.     

Statistical reconstruction for x-ray computed tomography using energy-integrating detectors

Statistical image reconstruction (SR) algorithms have the potential to significantly reduce x-ray CT image artefacts because they use a more accurate model than conventional filtered backprojection and can incorporate effects such as noise, incomplete data and nonlinear detector response. Most SR algorithms assume that the CT detectors are photon-counting devices and generate Poisson-distributed signals. However, actual CT detectors integrate energy from the x-ray beam and exhibit compound Poisson-distributed signal statistics. This study presents the first assessment of the impact on image quality of the resultant mismatch between the detector and signal statistics models assumed by the sinogram data model and the reconstruction algorithm. A 2D CT projection simulator was created to generate synthetic polyenergetic transmission data assuming (i) photon-counting with simple Poisson-distributed signals and (ii) energy-weighted detection with compound Poisson-distributed signals. An alternating minimization (AM) algorithm was used to reconstruct images from the data models (i) and (ii) for a typical abdominal scan protocol with incident particle fluence levels ranging from 10(5) to 1.6 x 10(6) photons/detector. The images reconstructed from data models (i) and (ii) were compared by visual inspection and image-quality figures of merit. The reconstructed image quality degraded significantly when the means were mismatched from the assumed model. However, if the signal means are appropriately modified, images from data models (i) and (ii) did not differ significantly even when SNR is very low. While data-mean mismatches characteristic of the difference between particle-fluence and energy-fluence transmission can cause significant streaking and cupping artefacts, the mismatch between the actual and assumed CT detector signal statistics did not significantly degrade image quality once systematic data means mismatches were corrected.     

Use of fuzzy edge single-photon emission computed tomography analysis in definite Alzheimer's disease - a retrospective study

Background  

Definite Alzheimer's disease (AD) requires neuropathological confirmation. Single-photon emission computed tomography (SPECT) may enhance diagnostic accuracy, but due to restricted sensitivity and specificity, the role of SPECT is largely limited with regard to this purpose.     

Three-dimensional imaging of xenograft tumors using optical computed and emission tomography

The physical basis and preliminary applications of optical computed tomography (optical-CT) and optical emission computed tomography (optical-ECT) are introduced, as new techniques with potential to provide unique 3D information on a variety of aspects of tumor structure and function. A particular focus here is imaging tumor micro-vasculature, and the spatial distribution of viable tumor cells, although the techniques have the potential for much wider application. The principle attractiveness of optical-CT and optical-ECT are that high resolution (<20 microm) and high contrast co-registered 3D images of structure and function can be acquired for relatively large intact samples. The unique combination of high contrast and resolution offers advantages over micro-CT and micro-MRI, and the lack of requirement for sectioning offers advantages over confocal microscopy, conventional microscopy, and histological sectioning techniques. Optical-CT/ECT are implemented using in-house custom apparatus and a commercial dissecting microscope capable of both transmission and fluorescence imaging. Basic studies to characterize imaging performance are presented. Negligible geometrical distortion and accurate reconstruction of relative attenuation coefficients was observed. Optical-CT and optical-ECT are investigated here by application to high resolution imaging of HCT116 xenograft tumors, about 1 cc in dimension, which were transfected with constitutive red fluorescent protein (RFP). Tumor microvasculature was stained in vivo by tail vein injection of either passive absorbing dyes or active fluorescent markers (FITC conjugated lectin). Prior to imaging, the tumors were removed (ex vivo) and optically cleared in a key process to make the samples amenable to light transmission. The cleared tumors were imaged in three modes (i) optical-CT to image the 3D distribution of microvasculature as indicated by absorbing dye, (ii) optical-ECT using the FITC excitation and emission filter set, to determine microvasculature as indicated by lectin-endothelial binding, and (iii) optical-ECT using the DSRed2 filter set to determine the 3D distribution of viable tumor as indicated by RFP emission. A clear correlation was observed between the independent vasculature imaging modes (i) and (ii) and postimaging histological sections, providing substantial validation of the optical-CT and optical-ECT techniques. Strong correlation was also observed between the RFP imaging of mode iii, and modes i and ii, supporting the intuitive conclusion that well-perfused regions contain significant viable tumor. In summary, optical-CT and optical-ECT, when combined with new optical clearing techniques, represent powerful new imaging modalities with potential for providing unique information on the structure and function of tumors.