Peter M. Joseph

Noise considerations in dual energy CT scanning

Dual energy CT scanning (tomochemistry) has been proposed as a method for determining various parameters relating to the elemental composition of the tissues. In this paper, our aim is to study the relative noise inherent in two proposed techniques for dual energy scanning; a two crystal technique and a two kV technique. In the two crystal technique, a split crystal detector is used to simultaneously obtain the high and low energy data during one scan at high kV. The two kV technique requires two scans taken with widely different kV settings. We first review three commonly used approaches for utilizing the scan data to compute the relevant parameters. A theoretical formalism is constructed which aids in understanding these methods. Then this formalism is used to study the influence of CT image noise on measurement precision in the case where the unknown parameters are densities. It is shown that, (1) the unavoidable overlap in the spectral data obtained by the two crystal technique results in a much lower signal-to-noise ratio than can be obtained by using the two kV technique, (2) the necessity for hard filtration of the high energy beam in the two kV technique has not heretofore been appreciated, and (3) the dose for a given x-ray tube heat load is also lower with the two kV technique.

Computed tomographic measurement of the xenon brain-blood partition coefficient and implications for regional cerebral blood flow: a preliminary report.

The calculation of regional cerebral blood flow requires, in addition to the measurement of the clearance, a knowledge of the regional brain-blood partition coefficient. The usual 133Xe washout techniques do not measure this latter parameter but use published values for normal brain tissue. This may lead to large errors in pathological tissue because the partition coefficient changes significantly in brain tumors. Investigations have begun into the use of CT and stable xenon to produce a cross sectional view of the brain in terms of its brain-blood partition coefficients. Results of experiments using an iodine phantom and xenon inhalation in animals are presented.

View sampling requirements in fan beam computed tomography

It is known that good CT reconstructions require that many views of the scanned object be obtained, where each view consists of many rays which traverse the patient. It is known that the required number of views depends on the diameter of the region reconstructed and the spatial resolution achieved. We present a new analysis of this question, valid for reconstruction from fan beam views, which shows that larger fan angles will require more views. Computer simulations, using a fan beam reconstruction algorithm, confirmed the predicted fan angle dependence and demonstrated that streak artifacts will result if too few views are used. The effect is also demonstrated experimentally by a phantom scan on a clinical CT scanner. Implications for clinical CT scanning are discussed.

The effects of sampling on CT images

Abstract Among the most important parameters that characterize CT data collection are the linear sample spacing, and the nominal beam-width. In this paper, we study the effects of these parameters on CT imaging using both a theoretical model and simulated images. An exact treatment of the point spread function in our model, which properly accounts for aliasing and reconstruction algorithm effects, is given for the case of parallel beam geometry. It is shown that it is always desirable to make the sample spacing as small as possible. In particular, aliasing artifacts are shown to decrease as sample spacing is decreased, both in the theoretical point spread function and in simulated CT images. Furthermore, abasing is shown to destroy the spatial invariance of the point spread function, whose shape will then depend on the exact location of the point in the scan field. Also investigated is the “volcano effect”, in which the point spread function develops a central depression when small sample spacing is used. It is shown that this effect produces no unusual distortions in the reconstruction of objects larger in diameter than the beam width; and further, that an attempt to avoid the volcano effect by means of inadequate sampling can have disastrous consequences for the reconstructed image.

Frontal and lateral views of the brain reconstructed from EMI axial slices.

A computer program has been written to produce lateral and frontal views of the brain from transverse slices generated by the EMI scanner. The resulting images have markedly facilitated the planning of surgical exploration of intracranial lesions by graphically indicating their locations in a manner consistent with the familiar lateral and frontal radiographs. Our image is formed from a distribution of the average EMI density number along the directio, of projection. Incorporated in the program is a bone-searching algorithm which eliminated density from the cranium and extracranial air and water so that only brain is seen in the projection. Several enhancement algorithms are provided to enhance pathology. The images are displayed on the existing EMI viewer, where they are available for inspection and photography.

Mathematical method for determining kVp from x‐ray attenuation measurements

A mathematical analysis of the shape of x-ray attenuation curves is presented that suggests a new method of determining the kVp of bremsstrahlung beams. The main result is a simple analytic formula which gives an excellent approximation to the attenuation curve for large thickness of high-Z absorber. From this, a graphical procedure is derived which yields a straight line whose slope is related to the mass attenuation coefficient evaluated at the peak energy. Included are corrections for aluminum filtration, ion-chamber energy dependence, and coherent scattering. The method is limited to beams generated by a constant potential and with negligible characteristic photons in the filtered spectrum. Numerical data are given for using copper absorber below 70 kVcp and tin above 90 kVcp.

Image Noise And Smoothing In Computed Tomography (CT) Scanners

The question of what spatial frequency response gives optimum image quality in computed tomography (CT) images is studied. We have developed a technique for smoothing these images by using a modified back projection filter with attenuated high frequency response. The results are images with a blurred appearance but dramatically reduced image noise. On processing raw data from the EMI head scanner in this way, grey and white matter distributions are clearly visualized with no increase in radiation dose. These results are analyzed in terms of the Weiner spectrum and correlation properties of CT noise. In particular, a noise equivalent aperture is defined in a way analogous to that of Shade. It is shown that the prediction of the simple Rose model of noise sampling can not explain the improvement produced by smoothing, even when the correlated character of the noise is taken into account.

The Influence of Gantry Geometry on Aliasing and Other Geometry Dependent Errors

At least three gantry geometries are widely used in medical CT scanners: (1) rotate-translate, (2) rotating detectors, (3) stationary detectors. There are significant geometrical differences between these designs, especially regarding (a) the region of space scanned by any given detector and (b) the sample density of rays which scan the patient. It is imperative to distinguish between views and rays in analyzing this situation. In particular, views are defined by the x-ray source in type 2 and by the detector in type 3 gantries. It is known that ray dependent errors are generally much more important than view dependent errors. It is shown that spatial resolution is primarily limited by the spacing between rays in any view, while the number of ray samples per beam width determines the extent of aliasing artifacts. Rotating detector gantries are especially susceptible to aliasing effects. It is shown that aliasing effects can distort the point spread function in a way that is highly dependent on the position of the point in the scanned field. Such effects can cause anomalies in the MTF functions as derived from points in machines with significant aliasing problems.

The Role Of Hospital Physicists In Performance Specifications

The concept of the radiological consumer writing specifications for x-ray equipment prior to purchase is both very new and well established. It is new in that many medical centers, who are now employing full time professional physicists and engineers to evaluate their existing equipment, are quite logically beginning to generate pre purchase specifications. However, at least one federal agency, the Veterans Administration, has for years required that all x-ray equipment purchased for their hospitals meet a long list of fairly demanding qyccifications. These requirements, published as Federal Specifications GG-X-635 , are an example of the kind of thing which a customer can demand and the industry can provide. If anything, the VA specifications err in the direction of being too specific and detailed, and most hospitals would be foolishi to adopt them in a wholesale manner. For example, they require that the MA meter shall have double or triple scales, that the unit fit in a room of between 9 and 10-1/2 feet, that an electrical steroscopic shift be available, etc. Obviously, also, any set of specifications must be upgraded to allow for new technologies as they come along. However, the VA has demonstrated that the x-ray industry can meet these standards of quality. By contrast, it has been our experience that, in the absence of any hospital based program of quality evaluation, even brand new x-ray installations can be seriously deficient in some way or other. For example, we have not only suffered from installation errors in such areas as TV systems, KV calibration, or tube protection circuits, but also from faults of design wherein the equipment could not possibly meet performance standards as good as those in GG-X-635. For example, we have a generator used in tomographic examinations at low KVP, where the KVP changes by 10 when switching MA stations. Other examples include timers with incorrect resistors values used in the design, KVP values which change by as much as 10 KVP during a given exposure (see figure 1) and between successive angiographic exposures, and phototimers that simply do not work. I emphasize that these faults were apparently not due to faulty installation or failure of components, but represent designed limitations in the original equipment.

Energy Dependence Effects On Performance Of Phototimers

The purpose of this paper is to communicate some observations on a some-what subtle aspect of the performance of x-ray phototimers: viz, the way in which their response depends on the x-ray energy spectrum and on the presence of scattered radiation. We specifically avoid the more basic question of sta-bility, reliability, and speed of response. Much of the phenomena which we have studied would, in the jargon of the radiological engineer, fall under the heading of KVP dependence. However, from the point of view of the physicist, it is the effective energy, n, of the beam which emerges from the patient, ra-ther than the KVP, which is the relevant variable. The point is that E depends on both KVP and the amount and nature of the absorbing substance in the beam. For example, we would expect the beam emerging from a skull study at 80 KVP to have higher R than that of a chest film at the same KVP. If this difference were significant to the phototimer, then it would be meaningless to speak of a KVP correction.