A method for determining high intensity x-ray spectra from diagnostic x-ray units is described. The method combines direct measurements of low intensity, constant potential x-ray spectra with digitized measurements of current and voltage waveforms at high intensity to determine high intensity x-ray spectra. The effect of ripple in the voltage waveform on the resulting x-ray spectra is also shown.

For calcium tungstate intensifying screens employed in film-screen imaging systems, Coltman found that approximately 1000 light photons of average energy 2.7 eV were produced for each 50 keV x-ray absorbed. Of this number, he found that only about 55% are emitted from the output side of a 109 mg/cm² screen. We have developed a method based on counting single photons to determine this number for various thicknesses of calcium tungstate screens. Monoenergetic x rays in the energy range from 17-69 keV produce upon absorption, a shower of individual photon pulses which are detected by a low noise photomultiplier. After amplification and discrimination against the noise background of the phototube, the resultant pulses are counted in a 70 MHz scaler or a 150 MHz counter. The detection system has a pulse resolving time of less than 15 ns. The data are then corrected for the quantum efficiency of the detector and normalized to the number of absorbed x rays which is determined in a separate experiment. For calcium tungstate screens with thicknesses of 30, 50, 86, and 123 mg/cm², the average numbers of light photons emitted per absorbed x ray are measured for 8 x-ray energies between 17- and 69-keV. The values for 50 keV are less than the values found by Coltman. Studies of the causes of this discrepancy are in progress.

A different method of measuring the quantum noise characteristic of image intensifier tubes is described. This method consists of measuring the transmission of primary radiation through the entire tube and, by correcting for the attenuation losses in the input and output windows, calculating the fractional number of x-rays incident on the tube that is actually detected by the input scintillating screen. The results of these measurements for several image intensifier tubes are presented and discussed. The results show that this measurement method is capable of distinguishing one tube from another as far as quantum detection is concerned. This measurement method has been found useful for measuring and spe-cifyin the noise characteristic of image intensifier tubes.

For the sensitometric evaluation of screen-film systems, consideration of reciprocity law failure (RLF) is necessary because it affects the choice of exposure modulation method. The effect of RLF on determining characteristic curves of screen-film systems is experimentally evaluated by comparing time-scale and intensity-scale methods and by comparing different exposure times in the intensity-scale method. A specially designed x-ray sensitometer permits very precise exposures for both time-scale and intensity-scale sensitometry. Results are presented for Kodak Min-R and Dupont Lo-dose/2 screen-film systems.

The use of an intensity scale sensitometer with a white light source does not simulate the discrete spectral emission by rare earth intensifying screens and as such does not truly measure the characteristic curve of rare earth screen film combinations. More traditional methods such as inverse square sensitometry are impractical in the oridinary clinical setting. The sensitometry cassette described here uses an intensifying screen with a neutral density filter to expose the x-ray film. Direct x-ray exposure of the film is eliminated by insertion of a lead glass plate between the intensifying screen and film layers. Both front and back emulsions may be exposed equally or the unequal exposure normally resulting from the absorption of x-ray photons by the front screen may be simulated. Optical spectometry of a lead glass indicates this method is valid for emissions of wavelengths greater than 400 nm which includes the rare earth phosphors. Comparison of characteristic curves obtained by this method and other more traditional methods will be shown to agree well. The system is simple, accurate, and tests can be formed on a clinically functional x-ray source and processing combination.

We present preliminary results for threshold detection of sharp-edged and "fuzzy" bars and discs superimposed on noiseless and noisy backgrounds. The luminance profiles for the "fuzzy" targets were approximately gaussian. The background quantum noise is optical density fluctuation in nominally uniformly exposed radiographs. The measurements were done using the method of adjustment. Our results for sharp-edged targets agreed with previously published work. Our "fuzzy" target results are novel and demonstrate minima in the contrast-target size curve. The minima were near S to 10 milliradians diameter for discs and 2 to 4 milli-radians width for bars. The increasing threshold slope for large targets suggests a gradient detection process.

The contrast sensitivity and patient entrance dose of a prototype computed radiography (CR) system is examined, and the results compared with those obtained from computed tomography and conventional radiography. Limiting perceptibility measurements of hole patterns are used to generate contrast-detail-dose diagrams for various contrasts, hole diameters, doses and scatter thicknesses. A perceptibility-dose factor is defined from which the relative dose efficiencies are derived for CR and conventional radiography.

Computer simulation of blood vessel images is described to study the effects of technical factors on observers' performance in detecting small details in angiographic images. The blood vessel images generated by a computer are embedded in radiographic noise of a given screen-film combination which was digitized at high resolution. The final images, which appear on film, are organized to permit the estimation of Receiver Operating Characteristic curves for detecting vessels and their correct location.

The Department of Radiology at Jackson Memorial Hospital performs 204,000 x-ray examinations per year involving 816,000 x-ray films. When it became necessary to replace a large percentage of existing screens we decided that, in view of the large potential saving in radiation dose to the community, the conversion to Rare-Earth Technology had to be considered. The size and complexity of this department makes it difficult to attempt a piecemeal conversion and therefore a combination phantom-patient study was undertaken to determine what - if any - compromises in image quality would be necessary to achieve significant dose reduction. Pelvic phantom images were produced at the same kVp, mA and identical film density (at a predetermined spot) utilizing 34 different film-screen combinations at surface exposures ranging from 55 mR to 650 mR. Radiologists in the department were asked to rank the films, blind, (including a sample of the system in use at that time, i.e. Dupont Hi-Plus Screens and 3M Type R Film) for contrast, sharpness, noise and overall subjective image quality on a scale of 1-5. The dependence of image quality on kVp was examined by taking phantom images of selected film-screen systems at 70, 90, and 110 kVp. These were ranked in a similar way. Simultaneous to the above phantom study, selected film-screen combinations were used for various routine diagnostic examinations to establish whether or not x-ray techniques could be adjusted to produce acceptable diagnostic image quality using Rare-Earth Systems in the entire department.

Recent Federal guidelines for radiation protection specify maximum exposure limits for certain radiological examinations performed in Federal institutions. For example, the maximum entrance skin exposure for a P.A. chest (23 cm) is recommended to be 30 milliroentgen. These guidelines, however, leave the criterion of image quality to the discretion of the institution. An investigation was done using an anthropomorphic chest phantom with simulated pulmoary nodules and pneumonia. A series of chest radiographs were produced at various entrance exposures by altering kilovoltage and image receptor. The exposures ranged from 5 mR to 65 mR. The variation of image quality obtainable within this range was categorized as least acceptable, acceptable, and most acceptable by three observers. The range of acceptability is correlated with entrance skin exposure.

Many new approaches and modifications of existing non-silver halide diagnostic imaging systems have been described in recent years. The basic concepts of some of these new developments are discussed, particularly with regard to their potential and limitations as future imaging systems. These systems include electroradiographic, thermographic, electronic, and solid state physical techniques. In addition, some comments on the application of modern image processing technologies are made. In spite of the fact that barely any of these systhems and ideas have reached commercial status, this review is offered to generate awareness on the present state of developmental efforts.

Obtaining hard copy images from some diagnostic medical procedures, such as Computed Tomography and Ultrasound, involves photography of video images. This is often done with cameras utilizing a negative type transparency film. Selecting a film for this application should include considerations of the characteristics of the cathode ray tube (CRT) being photographed and the properties of a video image. This paper describes CRT characteristics and their effects on film selection. The major film parameters which should be considered in selecting a film for video imaging are discussed. Our results indicate that optimum imaging is achieved with a relatively high speed, high contrast orthochromatic film.

An electrostatic radiographic receptor has been designed that allows the latent image resulting from an x-ray exposure to be reconstructed by means of a scanning laser beam. The receptor can be described as a duodielectric device making use of amorphous selenium as one of the dielectric components. The second dielectric component is a polyester sheet with a transparent gold coating on one side. Charges may be trapped at the surface of the selenium by placing an electric field across the structure while exposing it to light. A technique is described in which either subsequent polarity reversal or placing a short-circuit across the structure isolates these charges. Exposure to x-ray at this point selectively discharges the selenium surface, thus creating an electrostatic image. Further exposure of the system to a scanning light source produces a displacement current as the structure is totally discharged element by element. The resulting video signal is used to generate images of the charge pattern.

A computerized electronic radiography system is being developed for early non-invasive detection, characterization and quantification of atherosclerotic lesions. The method uses a conventional x-ray source combined with a solid state detector system which is coupled to a digital computer for processing and display of the radiographic information. The computer stores and optimizes the image for improved interpretation of the image detail. The computer assisted image enhancement, vessel localization and pattern analysis is an integral part of the instrumentation system. The system has been evaluated in preliminary studies to determine contrast sensitivity, radiation exposure, resolution and diagnostic quality. Even at this early developmental stage of this technique, very low contrasts have been visualized. The arterial images obtained with intravenous injections in dogs have demonstrated the potential of this method for the non-invasive detection and quantification of atherosclerotic disease.

A Rotating Aperature Wheel (RAW) technology is under development which involves replacing a grid for scatter elimination with an assembly of one fore and two aft lead aperature wheels which are rotated so as to maintain the alignment between the slit or aperature pattern of each. Such a design has the unique feature that its motion can be made indepen-dent of the x-ray exposure time and duration allowing for the first time the practical use of a moving slit anti-scatter technique in rapid sequence and dynamic radiographic procedures. The technology is flexible so as to lend itself toward use with varying source to image receptor distances such as is involved in fluoroscopy.

This paper describes one part of a system for automatically analysing chest x-rays. Previous publications have described the hardware, rib border detection and preliminary results for analysis of the lung fields. This paper is concerned with texture analysis and classification of the lung fields of chest x-rays. A description of the algorithms and some preliminary results are presented.

The field of mammography has changed dramatically during the past ten years. The introduction of new imaging techniques combined with the renewed interest in breast imaging has given mammography comparable status to other diagnostic radiology procedures. Mammography has been applied as a tool for mass screening for breast cancer. Several controversial reports have implied that in mass screening mammography more cancers may be induced than detected due to the radiation exposure to the breast. Such reports have put mammography on the defensive, not only for mass screening, but for conventional mammography in symptomatic women. Therefore, during the past few years great emphasis has been placed on dose reduction in breast imaging procedures. Dose reduction can be accomplished in many ways but usually at the cost of reduced image quality. In this presentation some of the factors affecting image quality will be discussed in order to provide some insight into the present status of mammography as a result of low dose techniques.

Controlled in vitro studies of breast microcalcification detectability were accomplished using a wide variety of standard and state-of-the-art mammography techniques (conventional and microfocal spot x-ray tubes, screen-film and xeroradiographic recording systems, contact and magnification techniques). Results confirm previous observations that geometric unsharpness is the limiting factor in microcalcification detectability for most conventional mammography systems. Results also indicate superior microcalcification detection for: xeroradiographic > screen-film recording systems, positive-mode > negative-mode xeroradiography, microfocal spot contact > conventional contact techniques, microfocal spot magnification > contact techniques, and 2X magnification > 1.5X magnification techniques. Clinical applications of these findings are discussed.

The possibility has been demonstrated of significant image improvement in screen-film mammography without increase in relevant patient dose, when a fast screen-film combination is used with magnification or grid, when compared with a slower screen-film combination using conventional technique (no grid use or magnification). An ideal dedi-cated mammography x-ray unit, containing a small (± 300 micron) tungsten focal spot, and a grid should result in optimal imaging of all breast sizes and tissue consistencies.

PRINCIPLE AND TECHNICS : 1) Structure and optical properties of cholesteric liquid crystals a) cholesteric structure - De Vries model (1951) b) selective reflection of light c) temperature influence d) cholesteric mixtures to he used in cutaneous thermometry 2) Technology and use of cholesteric sheets. a) description - two main properties : thickness and thermal conductibility h) specifications and gauging c) modalities of use - The Thermodetect 3) The cutaneous thermographic examination : conditions for use : environment, preparing the patient, incidency and exposures.

A totally new method of infra-red thermography is being developed using one and two dimensional arrays of Schottky barrier imaging detectors utilizing charge coupled device (CCD) readouts. This system, operating in the region of high contrast infra-red output from the human body (3.4-4.2 microns,)offers the promise of providing higher spactial resolution than existing thermographic systems. The signal output from mirror scanned 256 element linear, and 25 X 50 element square arrays can be presented as a dynamic video display in a topological view, as well as in grey scale and colorized format, greatly enhancing the capacity for detailed diagnosis from the visible thermographic image. In addition, the detector signal can be computer analyzed for the acquisition of additional diagnostic information, as well as for enhancement of the visible display.

Of the many factors which affect the quality of a mammogram, one of the most important is the radioluscency of the patient. It is well known that there is a significant variation in the patient population, yet the analysis and optimization of mammography imaging systems (and others) generally is based on the concept of an average patient. With the concern over dose in mammography and the resulting quest for the lowest doses consistent with acceptable diagnostic quality, it is of interest to investigate the implications of patient variability for the optimization of reduced dose mammography. In this paper we report on the use of the model of a mammography examination developed by Muntz to investigate the questions associated with a patient population having dispersed physical characteristics. Data on compressed breast thickness for a series of 120 patients receiving mammograms at LAC-USC Medical Center was used for information about the dispersion in patient characteristics. It is concluded that significant changes in mammography risk factor can be associated with patient variability. The risk factor variations are so large that the concept of the average patient needs to be re-thought for the case of reduced dose mammography.

Sensitometric response and modulation transfer function (MTF) measurements are used to scale noise power spectral (NPS) measurements to the exposure axis. The resulting value of noise equivalent quanta (NEQ) can be compared with the actual number of exposure quanta required to form the image (q ) ; the ratio yields the system detective quantum efficiency (DQE). Two systems used for mammography are studied.

A data management system, implemented on a minicomputer, has been developed to provide rapid and convenient access to parts, service, performance, and radiation safety records for diagnostic x-ray equipment. The system is designed to provide on-line information retrieval and reporting for the regional medical physics/x-ray service organization for our seven medical centers and their associated clinics. The on-line data base contains: complete information describing each x-ray system; parts inventories at each medical center and at regional service headquarters; and detailed summaries of all service, preventive maintenance, performance, and radiation safety reports for the previous two years and the current year-to-date. Data includes, in part, labor times, system downtime, problem areas and types, parts used, tube potential and current calibration, half-value layers, and radiation safety noncompliances to be corrected. In addition to retrieval of information for display on a CRT terminal, printed reports can be obtained on demand which summarize the financial transactions on a monthly and year-to-date basis. Additional reports describing parts inventories, x-ray system histories, and statistical summaries of service problems may also be printed on demand. The specific contents of the data base, correlative capabilities of the system, and examples of available reports will be discussed.

A microcomputer system, based on the Intel 8080, is implemented to collect data from various sources in a cardio-vascular examination room. The use of the equipment before, during and after the examination, is registrated on magnetic tape. A prototype of this registration apparatus (data-logger) is installed on the cardiovascular department of Dr. Buis in the academical hospital of Leiden (The Netherlands).

An evaluation of information flow within the radiology department revealed efficient and accurate techniques for information aquisition but little ability to analyze and disseminate the information. A computer was aquired to replace an obsolete card access system for filing patient x-ray numbers. The computer, while costing about the same as a replacement card system, enabled the acquisition of patient information and its distribution throughout the department. The system was adapted to existing procedures and personnel, causing a minimum of disruption in department operations. The system is capable of billing, reporting and generating statistical reports through the use of simple filing and sorting programs. These programs were developed in-house at a lower cost and have more flexibility than limited packaged programs. Adding an 80 megabyte disk to the system's original three 2.5 megabyte disks and expanding the computer memory to 96K words from 32K, has enabled other departments in the hospital to utilize the system. Other clinical applications in radiation therapy, nuclear medicine and information imaging have been considered. With the increasing utilization of computers in radiology, an understanding of computer capabilities as well as their applications to multiple tasks are important in reducing costs and improving patient care. The development of an in-house system has been a valuable learning experience.

Routine radiographic chest examinations have been performed under various radiological conditions that have utilized differences of x-ray tube potential, intensifying screen, source-to-image receptor distance (SID), grid, etc. In the process of acquiring chest radiographic equipment for our new radiology facility, purchase specifications were formulated. The specifications were designed so that the x-ray unit would accommodate multiple methods of routine radiographic chest examination. It was to be equipped with an automatic film changer and an automatic film processor. The operation modality of this unit encompasses most, if not all, presently utilized routine radiographic examinations for chest. The specifications of this idealized unit are based on clinical requirements and technological requirements, but at the same time conform with all applicable regulatory restrictions.

One of the first American Science and Engineering Co. (AS&E) CT scanners installed in a clinical environment is operational at the University of Colorado Medical Center. The performance characteristics of several aspects of this stationary detector array scanner have been measured. Included in these characteristics are the resolution capabilities of the scanner in terms of MTF for high contrast objects, and the perceptability capabilities for objects of lower contrast. Other characteristics include patient dose, beam profiles, CT number linearity, noise, algorithm dependences, and the performance of selected electromechanical features such as patient bed incrementing and light beam/radiation field coincidence. Equipment and techniques for measuring these quantities are described. Although data are presented solely for the CT scanner, the procedures for evaluation of performance and periodic quality control may be generalized to other CT scanners. A summary of quality assurance tests being performed and a brief description of these tests in tabular form are presented.

The performance and dosimetric characteristics of the first U.S-installed hybrid head scanner developed commercially by CGR have been evaluated. Using a translate-rotate motion with a xenon gas detector array, the scanner produces excellent low contrast resolution at low dose levels (single slice surface dose of 1 rad). Console selection of scanning parameters, viz.: scan time, kV, mA, beam filtration, slice width, and pixel size permits optimization of image quality for various exams. For optimizing either spatial or low contrast resolution, a push-button choice of reconstruction algorithms is available. Using the AAPM CT phantom, spatial resolution and line spread functions for both algo-rithms were determined from the graphical display of CT numbers. Low contrast resolution was measured at various contrast levels as a function of kV, dose, beam width, pixel size, and algorithm. Values for noise (σ)were also recorded for these scanning parameters. Patient dose distributions for various operator-selectable beam widths (3,6,9 mm) were obtained using film and TLD's. CT number linearity and contrast scale were determined using different plastic rods.

A method is presented to systematize the selection of a CT Scanner. The method consist of four parts: initial contact with vendor, paper analysis of the machine, physical survey of the machine, and final analysis. The paper discusses the details of the method and presents the results of the physical surveys of about twenty-five scanners.

The effect of CT scanner geometry on spatial resolution, contrast sensitivity, and patient dose are considered and used to determine the limiting factor or factors for a particular scanner. To this end analytical expressions for the components of the system modulation transfer function (MTF) are identified, and a Geometrical Figure of Merit is introduced based on the information capacity per rad. This formalism is used to compare four commercial CT units representing the main types presently in use. Improvements in scanner geometry are suggested which will improve this Figure of Merit for a given design. A new geometry, which would be relatively inexpensive to manufacture, is introduced and considered in the context of the formalism presented. Such a geometry utilizing state of the art components is shown to be superior to present and improved versions of CT scanners currently available.

The effect on dose of varying the imaging slice thickness in computed tomography was evaluated. A cylindrical Plexiglas phantom (20 cm diameter, 15 cm length) was used. The dose at each depth was estimated from measured dose profiles. The maximum surface dose was independent of slice thickness. The minimum surface dose in the scan plane increased with slice thickness. This indicated that scatter contributes to the dose in the scan plane. A computer program has been developed. It predicted the maximum surface dose for a scan procedure with an accuracy of ±1.2 mrad/mAs. The model takes the effect of scatter into account, assuming a zero-th order scattering cross section, which is a constant, independent of photon energy. A least squares fit with the experimental data permitted to estimate the scattering cross section (.095 cm ²/g, for the 10 mm slice thickness).

The Stand Alone Viewing Console is an autonomous console, offering a wide range of viewing functions. The input is on flexible disks, these being the medium used for recording Tomoscan 300 images. In this paper, the mini-computer-based organisation is explained and the use of the special-purpose imaging hardware unit (VDU) during viewing is shown in detail.

In conventional computed tomography images, only the average CT number, which is a first-order statistical parameter, is used to characterize the tissues by giving an estimate of tissue density. Second-order statistical parameters such as the signal variance and cross-correlation function have also been used to obtain additional information to discriminate between certain tissues and lesions. However, the contribution of quantum noise to the signal variance and cross-correlation function creates, for the conventional CT patient dose, a background signal often larger than the signal containing the information about tissue structure. The misleading information, called "artifacts," in second-order image statistics caused by quantum noise is studied. The distribution of the contribution of quantum noise to second-order statistical parameters is object-dependent, space dependent, and "nonlocal." The term "nonlocal" refers to the wide-ranging effects of nonlinear artifacts on the distribution of second-order statistical parameters. We present an analysis and examples of some of the artifacts encountered and also comment on how these artifacts affect the potential of these parameters for clinical use, such as tissue characterization.

Since the X-ray tubes used in CT scanners produce a polychromatic spectrum, the beam quality changes as the X-ray beam penetrates the scanned object, giving rise to differential beam hardening. Because of this effect, one cannot interpret CT numbers directly in terms of material densities or X-ray attenuation coefficients. Various schemes to correct for the beam hardening effect have been published and are used in commercial scanners. Although computer simulations have demonstrated the potential effectiveness of these schemes, uncertainties in the correction data limit the precision of the results, totally aside from limitations of photon counting statistics. In this work, the most basic beam hardening correction scheme, the water-equivalent correction, is analyzed. A formula is derived which relates uncertainties in the correction data to uncertainties in the reconstructed coefficients. Numerical estimates are given, showing that the uncertainties in the CT numbers can be on the order of 2½% to 10% even with the best correction data available. More sophisticated correction schemes are discussed from the point of reference developed for the water-equivalent correction. The analysis allows one to set requirements on the precision and stability of machine parameters and correction data.

A previously proposed method for beam hardening correction is illustrated using a mathematically described phantom based on an actual reconstructed cross-section of a human head. The method consists of a combination of polynomial correction of the polyenergetic data and an iterative estimation of the monoenergetic data based on previous reconstructions The quality of reconstructions is found to be similar to those which can be obtained from monoenergetic data.

The convolution/back-projection algorithm has certain characteristics which enable it to reconstruct useful images when the projection data or data-processing operations are incomplete or when they are approximated rather crudely. In this paper we show how the shape of the convolution kernel can be exploited for reconstruction from truncated projections and for ultra-fast approximate projection filtering. The results of simulations indicate that these techniques would be of value in practical computed tomography.

In dynamic radiological modalities the frame duration is often constrained to be comparable to characteristic times of the imaged dynamic process. The results can be interpreted as temporally averaged 'time exposures' only if the regional sensitivity distribution of the instrument is time independent. This condition fails for all transmission and most emission computed tomographic (CT) devices, in which only a fraction of the total number of rays is being measured at any instant. The CT image discrepancies from the ideal temporal average over the scan duration can influence the extraction of quantitative results from the data. A general theory of these discrepancies is difficult because a time dependent data collection geometry imaging a time dependent object is no longer spatially invariant in general. We have developed a conceptual scheme for the qualitative assessment of discrepancies due to time dependence. The scheme is evaluated by comparison of its predictions with computer simulations of two physiologically relevant time dependences imaged in six CT geometries.

We describe a program that allows a user to generate 3-D phantoms and the corresponding radiographic data that might be expected for an arrangement of cone-beam x-ray sources and 2-D detector arrays. The software described is the firgt part of a project (GOLEM) to include in a single framework: the generation of 3-D phantoms and test data, fully 3-D reconstruction algorithms for cone-beam x-ray sources, and (fully-3D) algorithm evaluation.

The loss of detection sensitivity incurred by any stage of image processing may normally be characterized by the frequency dependence of the detective quantum efficiency (DQE) of that stage of processing, provided the image is represented in continuous coordinates. However, limitations to the DQE concept arise when discretely sampled projection data are used to obtain discretely sampled computed tomographic (CT) reconstructions. The source of these limitations is the aliasing produced by the discrete sampling which mixes contributions from various frequencies. An associated problem is that the SNR for the detection of an object can depend upon the position of the object relative to the discrete reconstruction pixels. The effective SNR for discrete images must take into account this variation. While there may be no loss in the detection SNR for reconstructions in continuous coordinates (DQE = 100%!), a reduction in the SNR will result from aliasing for discrete reconstructions. A simple one-dimensional model elucidates the characteristics of discrete CT reconstruction.

X-ray exposures to patients were measured in 50 rooms in hospitals in Ontario. For a given procedure, the total exposure for a satisfactory examination differed from one room to another by as much as a factor of 60. Skin exposures reached 90 R and more for barium enemas and barium meals, very high exposures for patients who may not be seriously ill. The factors primarily responsible for these large differences in exposure were fluoroscopic exposure rate (0.65-25 R/min) and time (0.75-12 min), kVp and filtration (8:1), choice of screen-film combination (6:1) and attenuation of table tops and phototimers (4:1). The very high exposures can readily be reduced 50% to 75% at minimal cost and without any decrease in diagnostic acceptability of the images. The use in some rooms of minimal fluoroscopic exposure rates and of 70 mm fluorography, giving total skin exposures of less than 5 R, shows that the risks from diagnostic X-rays can be made negligible with modern, properly adjusted equipment.

The speeds of screen-film combinations available in Canada have been measured at 75 kVp with appropriately filtered beams. The influence on speeds of kVp, attenuation of the front of the cassette, age and use of screens, reciprocity failure and film variations from batch to batch is surveyed. The significance of the observed variations in speed (in selecting screen-film combinations) is considered. A set of radiographs of head and body phantoms has been used to examine the criteria which radiologists use in defining a 'good' radiograph and from this a qualitative assessment has been made of the priorities in selection of screen-film combinations.

Measurements of the time variations of kVp and exposure rate during diagnostic x-ray exposures give valuable information about faults in generators and about the radiation exposures to the patient. Simple solid state devices which display these parameters as a trace on a storage oscilloscope have been developed. Measurements have been made of their linearity, sensitivity and energy dependence. These have been used to optimize the design of devices suitable for measuring the exposure and kVp at the collimator, as well as the transmission of x-rays through grids and the patient.

The ultimate test of an X-ray imaging system is to establish whether radiographs contain all the image structure necessary for an accurate interpretation of the image for the particular anatomical region in the radiograph. There is therefore a need to define, as closely as possible, standards for the image quality required in each radiographic view and a need to provide radiographic phantoms which will test the X-ray imaging system for its ability to meet these standards. From the practical point of view, simple, inexpensive phantoms are required which may be used daily to check overall imaging performance. Simple phantoms for testing the performance of X-ray systems and specifically for angiography have been developed. These test system performance, and in particular response of phototiming devices and visibility of contrast material. Priorities in establishing standards are suggested.

An experimental rotating fan-beam (3rd generation) CT scanner has been designed and constructed at The Ontario Cancer Institute. The scanner incorporates an array of 201 xenon ionization detectors. These operate at a moderate pressure of 5 atm. with a collecting potential of 1 kV. A digitizing charge integration circuit is coupled to the collecting electrode of each detector so that detector data.can be digitally multiplexed onto a computer bus. The scanner is capable of acquiring scan data in 10 seconds per tomographic slice. The experimental nature of this scanner allows flexibility of x-ray technique, data accumulation, reconstruction algorithms, and image manipulation. The design of the detector and associated electronics is described and the method of scanner calibration is outlined with emphasis on some of the problems associated with fan-beam scanners.

The usefulness of CT scanners has been demonstrated in imaging many areas of the body, particularly the brain. However, one serious problem common to all scanners still remains and is related to the x-ray source. Because the source of x-rays produces a spectrum of radiation rather than monoenergetic x-rays, the CT numbers will depend upon the size of the patient so that they are in fact functions of the source, the size and shape of the patient and the point of reconstruction in the patient. For this reason brain tissues near the bone are imaged with elevated CT values so that it appears that the bone extends into the brain and therefore lesions near the skull would be missed. This artefact is known as the "spectral artefact". Another problem due to the polychromatic nature of the x-ray beam is the fact that it is very difficult to obtain meaningful attenuation coefficients and hence the electron density (σ), and average atomic number (Ζ) for each pixel directly from the CT image. The p is required for realistic dose calculations for radiotherapy and may be useful in diagnosis of tumours in the rest of the body. Two methods for correcting the "spectral artefact" and obtaining ρ and Ζ from CT scans will be discussed: 1) Dual detector; 2) Split filter. Results obtained from an experimental CT scanner will be shown.

One of the major uses of CT information is in dose calculations when the patient is treated with high energy radiation. Variations in electron density in tissue will affect both the primary and scattered radiation. The primary can easily be dealt with but the scattered component is very difficult to take into account. One requires detailed information on the relative importance of scatter from different volume elements. We have studied the scatter phenomenon in a water phantom irradiated by ⁶⁰Co radiation. We measured the effect on the dose to a point in the phantom when small volumes of water in the phantom are replaced with styrofoam which is nearly air equivalent. In this way we have obtained information on the amount of scattered dose from different parts of the phantom. However, the experimental procedure of replacing water with air has a perturbing effect on scatter from other regions of the phantom. In extreme cases, removal of scattering medium can actually increase the overall scattered dose to the point of measurement. In this paper we will present one set of experimental data and show an example of its use in a dose calculation for a nonhomogeneous phantom.

X-ray transmission computed tomography (CT) has been used primarily in medical diagnostic radiology since 1972. More recently, the application of CT to radiation therapy planning has been developed. In comparison to conventional radiography, CT images provide vastly improved anatomical localization of the tumour and of surrounding healthy tissue. Qualitatively, this information assists in the selection of appropriate radiation fields. Quantitatively, CT data can be processed to yield the electron densities of the tissues for calculation of the corresponding radiation distribution in the patient. In such calculations, the variations in tissue density and the three dimensional nature of structures are considered by employing a series of contiguous transverse section CT images which cover the entire patient volume to be irradiated. In this paper, the utilization of computed tomography for radiotherapy planning and for monitoring tissue response following radiotherapy will be demonstrated with a clinical example (tumour of the chest wall). The results presented were obtained with a computerized treatment planning system (TP-11, Atomic Energy of Canada) modified to interact with a commercial CT scanner (TR-40 Synerview, Picker X-ray Corporation). This system has been in clinical use for the past six months at the Princess Margaret Hospital.

Nuclear magnetic resonance (NMR) zeugmatographic imaging has some features in common with transmission and emission computed tomography, but there are a number of significant differences. These result, for the most part, from the highly non-linear character of the NMR phenomenon, and, from the slowness of changes in the nuclear magnetization, which take place in times ranging from microseconds to seconds. As a consequence, there is an intrinsic limitation on the sensitivity with which NMR signals can be observed, and the speed and, resolution of imaging applications are almost always limited by the signal-to-noise ratio available. Some biologically and medically interesting images already produced will be shown and future prospects, as well as the advantages of this new technique, outlined.

Probability theory provides a simple method for physicians to use their "intellectual linkages" to their past clinical experience in making current diagnoses. Only a pencil and paper are required for making a few likelihood calculations. To illustrate this method, evaluation was done of a new diagnostic sign (presence of knee ossification centers) for differentiating rubella from cytomegalic inclusion infection in young infants. Two practical questions can be answered by use of this likelihood method: 1) How certain can one be about either diagnosis when centers are present or absent? 2) How can other radiologists apply these results to their individual cases?

The Medical Micro-Dose X-Ray System was developed for human use from airport baggage inspection equipment and is the subject of this report. A heavily collimated lmmxlmm beam of x-ray scans the object. The x-ray beam moves linearly in relation to the object in both x and y axes. The x-ray beam is detected by a sodium iodide crystal. The resultant information is digitized and a direct video display is produced. 1024 levels of gray are stored and the displayed image is computer variable (much like a computerized tomography (CT) image). This variability allows the stored image to be displayed as many useful images. Linear resolution of 0.67 line pairs per millimeter has been obtained. A routine scan of fifteen by twenty inches is produced in sixteen seconds with a dosage of less than five hundred microrads to the patient. The medical uses of the Micro-Dose System are being investigated and preliminary results are favorable.

The image quality available on photofluorographic spot films (70,100 or 105 mm) has been gradually improving as high resolution image intensification has evolved. This paper describes that a resolution of 2.5 to 3 line pairs is now available on the Siemen's Videomed H television monitor. Such resolution now approximates that available on 100 mm film. The authors have been using a 70 mm camera to record the television image and have also modified a multiformat camera to record images in fluoroscopic examinations. This development is described and illustrated using clinical examples. The reduction in radiation and film cost is emphasized.

Recent improvements in the ultrasonic computed tomography system at the University of Colorado Medical Center include use of separate transducer apertures for attenuation and velocity imaging. In addition, scanning time to acquire attenuation, speed of ultrasound and pulse echo data of each plane has been reduced to four minutes. Artifacts due to reflection of ultrasound from waves on the water surface and problems of aligning the scanning tank with the breast have been reduced by covering the water surface with an acetate membrane and polyethylene liner including a circular hole for the breast. A detailed description of the scanning system is presented in this paper, and examples are given which demonstrate many characteristics of the attenuation and speed of ultrasound in imaging of the breast. Five of the 42 patients scanned to date with the ultrasonic CT system also were among those imaged with a newly available pulse echo breast scanner. An example is given which illustrates the correlation between pulse echo and ultrasonic CT images.

A. multiple x-ray source high-speed transaxial scanner system (DSR) is about to undergo evaluation. studies. The capH.bility for programmable scanning modes and operator interactive retrospective reconfiguration of scan data makes the DSR a very powerful research tool. The physics and technological basis for system design and selection of several major components of the DSR scanner are discussed.