EMISSION-TRANSMISSION IMAGING SYSTEM
USING SINGLE ENERGY AND DUAL ENERGY
TRANSMISSION AND RADIONUCLIDE
EMISSION DATA 5
This invention was made by Government support under Grant No. lPOlDK-39964 awarded by the National Institute of Health. The Government has certain rights in the invention. 10
BACKGROUND OF THE INVENTION
This invention relates generally to x-ray transmission and radionuclide emission imaging systems. In a primary application the invention relates to a diagnostic 15 imaging system where dual-energy x-ray transmission data are used to derive material properties of the object imaged, which properties are then used directly to correct radionuclide emission data obtained simultaneously from the same object using the same imaging system. 20
Emission radionuclide imaging is a well-established technique for localization of lesions in diagnostic radiology and nuclear medicine. Briefly, a compound labelled with a radionuclide or a radionuclide itself is injected into a subject. The radiolabeled material concentrates 25 in an organ or lesion of interest, or can show a concentration defect. At a prescribed time following injection, the pattern of concentration of the radiolabeled material is imaged by a rectilinear scanner, gamma camera, single photon emission computed tomography 30 (SPECT) system or positron emission tomography (PET) system. The imaging procedure requires a means to define the vector path along which the emitted gamma-ray travels before striking the detector of the imaging system. In the first three systems a collimator (typi- 35 cally made of lead or other high-atomic number material is interposed between the object and the detector to define the gamma ray path; in PET the unique characteristics of positron annihilation radiation are coupled with electronic circuitry to define the vector path. In all 40 cases, the only information obtained when a gamma-ray strikes the detector is the fact that the photon originated somewhere within the object along the vector path projected back from the detector. For projection imaging systems, a two-dimensional image is formed with 45 the intensity of each picture element, or pixel, proportional to the number of photons striking the detector at that position. In SPECT and PET, the vector paths are determined for multiple projection positions, or views, of the object, and cross-sectional or tomographic im- 50 ages are reconstructed of the object using standard algorithms. Again, the intensity assigned to each vector path is proportional to the number of photons striking the detector originating along the path, and the intensity of each pixel in the reconstructed image is related to 55 these vector path intensities obtained at multiple views.
Both for projection and CT radionuclide imaging it is desireable to obtain absolute values for radionuclide concentrations (or radionuclide uptake) at each point in the image. Attenuation of the emitted photons within 60 the object, before they reach the detector, is a function of the energy of the photons and the exact composition of the material through which the photon vector path passes to reach the detector. Photons emitted deeper within the object have a higher probability of attenua- 65 tion than photons emitted near the surface. In addition, the composition of the material (in terms of effective atomic number, Z, and electron density) affects the
2
attenuation, with more attenuation if the path passes through high Z or high density regions. Thus, in order to calculate absolute uptake or concentration of a radionuclide in a region of an object it is required that the path length of each type of material or tissue (or effective Z and electron density path lengths) be known for each vector. Attenuation corrections for emitted photons are made from this knowledge, allowing accurate concentration values to be obtained.
Prior art in this field involves two approaches. The first approach uses conventional projection or CT images to estimate the material composition of the object, from which attenuation corrections can be estimated. This involves obtaining a conventional image, mapping the radionuclide distribution onto the conventional image, generating estimated material path lengths along the emission vector paths, and correcting the observed radionuclide distribution based on these path lengths. There are two fundamental limitations in this approach. First the conventional and radionuclide images are obtained sequentially using different instruments, generally in different rooms and often on different days. These factors lead to significant problems in misregistration of the conventional and radionuclide image data sets. While 3-D image matching algorithms have been applied to this problem, their inability to solve the problem has kept them from routine use. Only limited success is obtained in objects of relatively time-invariant composition such as the head, and the method is essentially useless in objects such as the chest or abdomen where motion is continuous. Secondly, conventional CT or projection images suffer from inaccuracies in determined material properties due to beam hardening and scattering among other effects. Therefore, the attenuation corrections applied to the radionuclide data for photons of often very different energy than the x-ray data are only estimates.
The second approach known in prior art is the use of a radionuclide transmission source to obtain total path length attenuation measurements, then to use these measurements to estimate attenuation corrections for the emitted photons from the administered radionuclide. Budinger and Gullberg, "Three-Dimensional Reconstruction in Nuclear Science," IEEE Transactions on Nuclear Science, June 1974, have suggested the use of multiple radionuclide photon energies to obtain material-selective path length data, and Peppier, "Combined Transmission-Emission Scanning Using Dual-Photon Absorptiometry,"Ph. D. Thesis, University of Wisconsin, 1981, has used sequential measurements, using two radionuclide energies to obtain data about bone and soft-tissue path lengths followed by acquisition of emission data which was then corrected by the transmission measurements. Some of these prior approaches use simultaneous or near simultaneous acquisition of transmission and emission data while others use sequential acquisition. However, all have the fundamental limitation that the use of a radionuclide transmission source to determine effective material path lengths limits the accuracy of these measurements due to statistical noise in the transmitted photon intensity measurements.
Material-selective imaging using x-ray sources has been described by Alvarez and Macovski in U.S. Pat. No. 4,029,963. This method or modification thereof can be used to determine effective material path lengths using two sequentially or simultaneously-acquired data sets at two different effective photon energies. Once the object being imaged is decomposed into a basis set of
two properties, a priori knowledge of the energy dependence of the basis properties can be used to reconstruct an image of the object at the exact energy of the radionuclide emission photon being imaged. From this image is derived exact attenuation corrections for the emitted 5 photons.
The present invention overcomes the fundamental limitations of the prior art. Specifically, transmitted photons from an x-ray source are acquired simultaneously with emitted photons from the contained radio- 10 nuclide source, using the same photon detector in an identical geometry, so all the problems of misregistration are overcome. The dual-energy x-ray projection data are solved exactly for material-specific properties and recombined into an effectively monoenergetic im- 15 age, eliminating inaccuracies in material property estimation due to beam hardening. The high photon flux from the x-ray tube overcomes the accuracy limitation of radionuclide transmission sources.
SUMMARY OF THE INVENTION 20
A primary object of the invention is to provide a method and instrument to obtain accurate measurement of the uptake or distribution of a radionuclide within an object using dual-energy transmission x-ray determined 25 material-specific attenuation properties to correct simultaneously-measured emission photon data.
Another object of the invention is improved localization of radionuclide emission using simultaneously acquired single energy or dual energy x-ray transmission. 30
In accordance with a feature of the invention, the accurate localized measurements so obtained can also be referenced to one or more of the materials in the object to obtain concentration information.
The invention includes a dual-energy x-ray source 35 mounted opposite to a collimated photon-counting detector and associated pulse counting electronics capable of discriminating incident photons into at least three separate photon energy windows and counting at a rate of approximately one million counts per second per 40 detector. An object of variable composition containing one or more radionuclide sources is interposed between the source and detector. Dual-energy x-ray transmission data along projections through the object are decomposed into two basis set projections of known energy 45 dependence, then recombined into an effective monoenergetic projection at the energy (or energies) of the contained radionuclide source. Radionuclide emission data along the same projection vector as the dualenergy x-ray transmission data are acquired simulta- 50 neously using the third (or more) energy windows(s) of the detector. The monoenergetic and/or original decomposed basis set projections are used to correct the detected emission photon intensity for attenuation along the path vector, based on the measured material-selec- 55 tive attenuation path lengths. The corrected emission data are used to obtain the absolute quantity or concentration of the radionuclide along the projection path. A conventional projection or CT radionuclide image is then formed from the data projections superimposed on 60 any of the images from the dual-energy x-ray data set. Alternatively, an image from a single energy x-ray data set simultaneously acquired with radionuclide emission data can be superimposed on a radionuclide data image to improve the localization of the radionuclide emis- 65 sion. The invention also has the clinical advantage of being a single instrument on which the equivalent of two conventional studies are done simultaneously, with
attendant savings in technician time and floor space in the hospital.
The invention and objects and features thereof will be more readily apparent from the following detailed description and appended claims when taken with the drawings.
BRIEF DESCRIPTION OF THE DRAWING
FIG. 1A is a schematic illustrating one embodiment of the invention.
FIG. IB shows a graph of a typical photon energy spectrum incident on the detector of FIG. 1A.
FIG. 1C shows a schematic of the detail of the data acquisition system of FIG. 1A.
FIG. 2 illustrates the method required of practicing the invention.
FIG. 3A and FIG. 3B illustrate an alternative embodiment of the invention.
DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS
An emission-transmission imaging system in accordance with one embodiment of the invention is illustrated in FIG. 1A. An x-ray tube 12 is mounted on a conventional gantry for either projection imaging or computed tomography. A rare earth filter 13 (e.g. samarium or gadolinium) is interposed between the x-ray tube and object 16 in order to produce a dual-energy x-ray spectrum incident on the object as shown by the low-energy 14 and high-energy 15 x-ray peaks in FIG. IB. The low and high energy x-ray photons pass through the object 16, through a collimator 17 focussed at the x-ray source, and are incident on an energy-sensitive detector 18. The collimator has the dual function of rejecting scatter in x-ray transmission data and localizing the radionuclide emission path. The detector can consist of one or multiple elements and is also mounted (along with the collimator) on the gantry opposed to the x-ray tube. Photons emitted from a radionuclide(s) such as Xe-133 contained within the object are emitted isotropically with an energy 19 shown in FIG. IB that is characteristic of the radionuclide. Those photons emitted in the direction of the detector pass through the collimator and are incident on the energy-sensitive detector. The detector must be capable of counting at very high counting rates with good energy resolution (in the range of 10 KeV at one million counts per second) in order to minimize crosstalk, pulse pileup and errors from scattered radiation. The high counting rate is necessary to produce low and high energy data sets of sufficient statistical accuracy in each projection to produce an accurate basis material decomposition. Such a detector is, for instance, high purity germanium, which is known in the art. Alternatively, other known high speed detectors such as cadmium telluride, lead iodide, and mercuric iodine can be used. The count from the multielement detector array 18 is passed to data acquisition and pulse counting electronics 20. As further illustrated in FIG. 1C, the electronics 20 includes preamplifiers 30 and amplifiers 31 which are optimized for these high counting rates and good energy resolution. The output voltage of the amplifiers is proportional to the photon energy spectrum incident on the detectors, and is sampled into three or more energy windows 32 using conventional high speed discriminator and logic chips. The data are counted at 33 and output through registers 34 and BUS 35 to a computer 26 and display 28.
|
