The hypothetical risk of carcinogenesis to the breast from low dose mammography and the known benefits to the patient from detection of breast cancer at an early curable stage through mammographic screening will be assessed. Benefit/risk ratios will be derived based on the best available clinical and biological evidence. This analysis will support mammographic screening of all women above 40 years of age.

Committee SC-72 of the National Council on Radiation Protection and Measurements (NCRP) was formed to prepare a report on Radiation Protection in Clinical Mammography which will both supplement and expand upon the material presented in NCRP Report No. 66: "Mammography". Highlights of the current draft of the report are presented.

In order to provide quality assurance for the National Breast Screening Study currently underway in Canada, a mailed survey package was designed which allows for the monitoring of i) the exposure for the "average" craniocaudad view, ii) beam quality, iii) speed, contrast, and fog parameters of the film processing, and iv) the overall image quality. As well, measurements were made of depth-dose curves to a breast phantom for a number of techniques so that mean glandular dose to the breast could be estimated from entrance exposures. To measure the duration of the x-ray exposure, a device which uses a photodiode to detect scattered radiation from the breast was designed and has been used to check consistency of phototiming and x-ray output. At one centre, the average skin exposure for the craniocaudad view was 295 mR. The range was 150 - 800 mR, due to differences in tissue density and distribution, breast thickness and compression. The mean glandular dose was estimated to be 66 mrad. For the intercomparison of imaging at the centres, identical lucite phantoms have been constructed and distributed which require an exposure time similar to that of the "average" breast. An aluminum step wedge built up of 15 sheets of 0.4mm Al is used to obtain "sensitometric" curves for the mailed survey. Some of the factors responsible for the variation in patient dose and image quality in the study are discussed.

Since failure to biopsy a malignant mass during a relatively early stage of its development will result in a severely shortened life span for the patient, breast imaging techniques must be capable not only of detecting a mass, but of differentiating between the image characteristics of a benign and a malignant tumor. Ideally, the applied technique should be equally capable when used to examine the young breast and the older breast. The specific advantages and disadvantages of x-ray mammography and of ultrasound visualization, as well as the rationale of balancing the disadvantages of one technique with the advantages of the other, are discussed in this presentation.

The relationship between absorbed dose and image quality in xeromammography is examined as a function of various x-ray beam and breast parameters. Image quality is assessed in terms of measured density profiles across low contrast edges. The radiographic risk is quantified by means of the average absorbed dose to the whole breast. It is concluded the diagnostically useful images can be obtained at exposure levels substantially lower than currently quoted in literature.

Radiographs of a phantom containing excised breast tissue and idealized test objects embedded in tissue equivalent plastic have been obtained using a variety of mammographic film-screen detectors. A CGR Senographe (molybdenum target, beryllium window) was modified to accept a series of x-ray filters: 0.03 mm molybdenum, 0.5 mm aluminum or 1.2 mm aluminum. Radiographs were obtained at kVp's ranging from 26 to 32 kVp with the mAs adjusted to achieve the same optical density on the films. Entrance skin exposure was measured for the range of techniques from which dependence of mean gland dose on kVp was derived. Indices of image quality were also determined from the radiographs. A discussion of the effects of kVp variation on mammographic image quality, especially on low contrast discrimination is included.

Sonographic detectability of mammographically demonstrable breast calcifications depends primarily on the size of the calcific particles. Calcifications smaller than 2 mm in size are detected in fewer than 10% of cases, with equally low detection rates for benign and malignant particles. The inability of sonography to routinely identify malignant breast calcifications will severely restrict its use as a screening tool for cancer detection.

Signal-to-noise ratio, dynamic range and contrast sensitivity are three parameters used to describe performance or quality of electronic medical imaging systems or components. The meaning and significance of these parameters are discussed and their values for various systems, including conventional film-screen systems, are derived. An additional parameter, "useful dynamic range", is introduced and related to existing systems. Under optimum conditions the signal-to-noise ratio of a radiographic imaging system will be determined only by quantum statistics. For a quantum limited system contrast sensitivity is related to x-ray dose and independent of the particular system used, analog or digital. The "useful dynamic range" of most electronic imaging systems is limited by system noise to a value less than that typically exhibited by conventional film-screen systems, 20:1 to 50:1.

In most previous studies involving the ideal observer, the task considered has been that of simple detection where it is assumed that there is complete a priori knowledge of the background and of the possible object's shape, amplitude, and position. It is shown that redefining the detection task to include the possibility of an unknown, slowly varying background reduces the importance of the low-frequency components in the image for the ideal observer. More complicated tasks than object detection are also considered, such as determination of an object's position and width and the resolution of two objects. These higher-order tasks further enhance the importance of the high-frequency information content of the image.

The contrast-detail (C-D) diagram has been used widely yin the evaluation of imaging systems including screen-film combinations, computed tomography, and digital radiography. Despite its apparent simplicity and efficiency in generating observer performance data, the relationship between this method and the more rigorous ROC or forced-choice procedures is not well understood. With a high-quality image processing/simulation system, we made images by superimposing digitally simulated square patterns on radiographic noise. We then studied the statistical variability in the C-D diagram due to within- and between-reader differences, and the effect of different numbers of image samples. Results indicate that the percent standard deviations due to each of these variations are about 10-20%. We also compared the visibility thresholds measured by the C-D diagram method and by an 18-alternative forced-choice (18AFC) procedure. Although the thresholds defined in these two methods are fundamentally different, we found that the threshold contrasts measured by the two procedures have similar values.

A set of universal curves can be used to estimate the scattered x-ray spectrum from an arbitrary thickness of water and a monoenergetic primary exit spectrum. By approximating the primary exit spectrum as a discrete spectrum of narrow energy bins (photons per keV vs keV), these curves allow for a rapid estimation (±20%) of the resultant scatter spectrum. Experimental confirmation is shown.

The detection of a cardiovascular stenosis involves the complex interaction of technique selection, system performance and anatomy. Technique selection for instance involves choosing the correct focal spot, pulse width, KVp, and frame rate for a given patient and projection angle. In order to provide guidance in this selection process, these variables were investigated in terms of their impact on image quality. Conclusions which were confirmed clinically are summarized in terms of recommended techniques for a range of patient sizes. The confirmation includes clinical films which demonstrates the effects of parameter changes.

Radiological fluoroscopy, using x-ray image intensifiers (XRII) and TV systems, is limited in visual noise or mottle by the requirement of minimizing radiation dose to the patient. Various methods have been used for making this noise more acceptable, such as using laggy vidicons (rather than plumbicons) or digital stores to integrate and thus reduce noise. Integration implies some motional blurring and this problem has to be balanced against noise reduction. In obtaining static images from such XRII-TV system, further choices come into play, (a) optical exposure time in the multiformat camera and (b) whether or not the radiation is pulsed. We have performed mottle and motion blurring evaluations on radiographic images produced during continuous fluoroscopy via TV systems (i.e. videofluorographs) or directly from x-ray image intensifiers for various exposure times. These results are compared with each other and direct measurements of the TV system's temporal frequency response. It was found for the particular clinical system evaluated (Siemens 1023 line TV and Matrix Instruments FV1-1 multiformat camera) that (a) TV generated noise was sufficiently high to require several TV frames to be integrated onto the film image to reduce it to acceptable proportion. (b) That the vidicon had a consider-able lag--effectively integrating for ~ 0.2 s with consequent motional blurring problems. (c) Finally that clinically useful images were obtained using (allowing for system integration) 10μR to the imaging system. This is a factor of 10 smaller than 100mm photofluorographs and 100 times smaller than PAR speed film-screen combinations. The compromises to image quality which are intrinsic and those which are particular to the system used are distinguished. Based on these results an ideal, yet feasible imaging system is outlined for digital radiography which is optimized for low dose-low information content general radiography.

Occurrence of ghost images on radiographs is a generally feared phenomenon. A method is presented how to evaluate in a quantitative way the photographic effect of long term persistency of radiographic screens. It is shown that the effect not only depends on the radiation intensities present in the preceding exposure(s) but also on the radiation intensities present in the actual exposure. By the method an objective forecast of possible ghost imaging can be made. Results obtained with Curix MR400-screens are presented.

For the case of radiographic contrast media flowing through an otherwise stationary object, a theoretical analysis shows that image enhancement with optimal dose utilization can be achieved by varying the x-ray intensity during the sequence of exposures. The need for variable intensity derives from the conflicting requirements of continuous visualization of the time course of the contrast media and the fact that dose utilization in a difference image is maximized when all the x-ray photons are divided between minimum and maximum opacification. A set of optimal weights for combining multiple images are derived as a function of the x-ray intensity in each image and the effects of photon noise, camera (video system) noise, and digital truncation noise. It is shown theoretically that controlled variation of x-ray intensity may improve the signal to noise to dose ratio by approximately a factor of 2 when compared to conventional matched filtering. If the x-ray intensity remains constant for all images, the optimal filter weights reduce to those of the expected matched filter. A simple implementation of optimal filtering using only two intensities is described mathematically and demonstrated experimentally.

A circuit has been constructed to perform spatial frequency filtration on DSA images at real-time video rates. The 10-bit device performs low-pass or high-pass filtering, and with external memory can perform bandpass and more sophisticated filtering. Pixels in the convolving kernel are weighted independendently in the x- and y-directions to provide a Gaussian-like convolving function. The kernel width ranges from 3-30 pixels and appropriate weighting yields a FWHM of the Gaussian kernel function as small as 0.85 pixel width for horizontal image widths of ≤256 pixels and as small as 1.25 pixel width for a horizontal image width of 512 pixels. Applications to be investigated include scatter and glare correction for videodensitometry, enhancement of arteries behind large opacified structures such as the ventricle and aorta, noise suppression in low-spatial frequency DSA exams, edge-enhancement of images, and partial-pixel shifting. Peli and Lim have suggested a more sophisticated algorithm which enhances high-pass filtration only in dark regions of an image. This and other techniques may be implemented with the current circuit to enhance small detail in highly opacified regions such as the ventricle, while leaving the rest of the image unaltered.

While the quantitative nature of digital fluroscopic and digital radiographic imaging would seem to make possible many radiotherapeutic applications, significant problems must first be solved. The purpose of this study is to examine the potential benefits of digital imaging in radiation therapy, identify what problems must be solved to attain these benefits, and attempt to make some assessment as to the relative merits of developing such technologies. Among the potential benefits are ability to do contrast-subtraction studies for tumor and/or normal structure localization and portal placement, digital simulation of portal placement and treatment delivery, spatial localization and digitization of inhomogeneity boundaries, spatial localization and digitization of brachytherapy sources, and quantitative inhomogeneity acquisition for dosimetry calculations. Significant problems which must be solved include image receptor size limitations, image acquisition geometry related spatial distortion, three dimensional region calculation from limited views, and physical interpretation of digital image grey levels. Maximal benefits of digital imaging in radiation oncology is likely to be obtained by the development of large area image receptors for use in therapy simulators with direct data link to a combined image-analysis/treatment planning computer.

The resolution characteristics of a digital subtraction angiographic (DSA) system have been evaluated under two different magnification techniques, namely; the geometrical magnification, and the electro-optical magnification. It has been found that the DSA system resolution is less dependent on the focal spot size (or, modulation transfer function, MTF) than that of routine radiographic angiographic studies. The evaluation results have been utilized to re-establish the DSA examination procedure.

We have developed a simple technique to evaluate and monitor the performance of digital fluoroscopy imaging systems in a clinical environment. We use a specially designed phantom based on a commercially available fluoroscopic stepwedge, placed on top of a water bath which is interlaced with rubber tubing simulating arteries. Noise characteristics are found by taking a difference image of the stationary phantom. Dynamic range and fidelity of logarithmic amplification in the presence of scattered radiation and veiling glare are measured by analyzing the image of the phantom and the difference image of the stepwedge in two different positions. Vascular contrast resolution is measured by taking the difference image of diluted contrast agent injected into the tubing. Further analysis of the obtained images can reveal possible instability in the video scanning pattern or the image intensifier electronic lens system and line-to-line sweep circuit clamping instability of a TV camera. We have used this technique on a routine basis for quality assurance of a digital system for some time.

Coronary interventional procedures are placing new demands on fluoroscopic imaging performance. These procedural requirements are defined and related to the performance of a high definition video system. System factors considered include line rate, intensifier dose, spatial resolution and temporal response.

The benefits of a comprehensive quality assurance (QA) program in conventional diagnostic radiology are well understood. In computed tomography (CT), however, the situation is less clear. Since the advent of CT a number of manufacturers and individual medical physicists have recommended a variety of parameters for periodic monitoring or system performance evaluation. These recommendations are reviewed. There are no data, however, indicating that CT facilities, with the exception of some mobile units, have implemented a comprehensive QA program, including the daily collection of data and its interpretation and application to the identification and correction of problems with the CT system. Data from a nationwide study will be presented which describe the present status of QA CT facilities. QA programs on mobile CT systems, where there is more concern about consistent system performance due to the effects of moving the unit, have suggested that trends in performance can be identified. Preliminary data indicates that a similar situation may hold in the case of fixed facilities. This suggests that CT QA programs may permit detection of problems before they significantly affect the diagnostic value of the CT image. The recent development of automated CT QA software by a number of investigators will also facilitate the implementation of appropriate QA protocols.

A user based quality assurance program for assessing image quality changes in Computed Tomography is described. Measured image quality parameters include: CT number and standard deviation (noise) in a water phantom, contrast scale based on the CT number difference between plexiglass and water, and resolution as indicated by the modulation of a plastic bar pattern. For any quality assurance program to be effective, realistic values for allowable variation in image quality parameters must be established. This requires an understanding of the effect on image parameters caused by variations in the CT system parameters. In order to understand this interaction, a General Electric CT/T 8800 scanner was systematically "detuned" and image quality parameters monitored. CT system parameters studied include kVp, x-ray tube focal spot size, x-ray output, slice thickness and system alignment. The effect on the image for each system parameter is established and an attempt is made to identify system parameters whose status can be assessed by image parameter measurements.

Due to the widespread use of high-speed rare-earth screen/film combinations and pulsed cine and digital radiography systems in radiology departments, the short exposure time performance of the x-ray generator is of increased importance to the production of high qual-ity images. For short exposures slow rise time of the high voltage applied to the x-ray tube may result in the peak tube potential differing significantly from that indicated. Similarly, a pulse overshoot on the leading edge of the high voltage waveform may produce a peak tube potential greater than that indicated. The measured high voltage and current waveforms of three-phase generators representative of the major x-ray equipment manufacturers are presented as a function of tube potential, tube current, and exposure time. The implications of the results are discussed, and a comprehensive definition of generator minimum exposure time is proposed.

Digital radiographic flow patterns were analyzed during series of pulsed contrast bolus injections in both phantoms and canine arteries. The effects of stenoses in the phantoms and canine vessels on the flow patterns were also studied. Our digital radiographic results indicate that it is possible to determine the velocity of flow by the pulsatile injection mode and that the large variations in flow patterns observed is a strong function of the elastic properties of the flexible phantoms or arteries as well as of the degree of stenosis. Both cardiac gating and choice of injection site proved important in the canine studies. Work is underway to extend these results to NMR.

Digital tomosynthesis (DTS) is the implementation of classical geometric tomography on a digital radiography system. A set of original images is collected during a single tomographic sweep. These images are spatially shifted, corrected for distortion if necessary, and added together to synthesize a single cut in any desired plane. Any number of additional cuts can be synthesized from the same set of original images by altering the amounts of shift and correction. A practical system for performing DTS is described. Engineering limits and the results of human and animal studies are presented. Our work in DTS was prompted by earlier efforts of Bally and co-workers on an analog system, and evolved through an early digital prototype reported elsewhere. The present system, based on commercially available digital fluoroscopy equipment, is more flexible. Data collection time can be short (down to 1/2 second) to permit angiotomography, which may be the most important application for the method. Mask subtraction and spatial filtering to remove out of focus information are options. The synthesis programs produce a new cut plane every few seconds. Contiguous stored planes can be rapidly replayed. Contrast sensitivity in the DTS mask subtraction option is close to simple digital fluoroscopy subtraction, but spatial resolution is somewhat worse. A linear tomographic motion and 512 by 512 pixel image are presently used. Extensions to other motions and larger matrix sizes are under development.

The use of electronic imaging detectors in radiography and fluoroscopy allows the spatial variation of transmitted radiation intensity to be measured directly and stored in electronic devices for subsequent analysis and display. By eliminating the conventional use of photographic film in conjunction with scintillation screens or image intensifiers, potential cost and performance benefits can be accrued. Most electronic imaging detectors considered to date have poor resolution characteristics. However, this is usually due to problems in fabrication and charge readout and not to fundamental limits in performance. Fundamental limitations in the performance of electronic detectors result from charge deposited away from the position where an X-ray photon is initially absorbed and are dictated by the material composition of the detector. The charge transport arises from either secondary photons or secondary electrons. In this paper, we present preliminary results on computer calculations of the effects of secondary radiation transport on the intrinsic imaging characteristics of electronic imaging detectors as a function of their elemental composition.

Measurements and theoretical calculations of the limitations of film-screen mammography have been made. The results show that the use of mammographic film-screen systems to image the breast has three major drawbacks: 1) insufficient contrast enhancement for the visualization of subtle tumours and microcalcificatons; 2) limited latitude, i.e. in some patients, structures in thin and thick or dense regions of the breast are not clearly visible on the same film; and 3) dose inefficiency, since most mammograms are not quantum-limited. Digital techniques may overcome these problems by decoupling the image recording and display processes, allowing for contrast enhancement without loss of dynamic range. Using digitized film mammograms and phantom images, we have made a preliminary study of whether image processing can improve visualization of breast lesions. Contrast enhancement by histogram equalization and linear scaling with clipping, and edge enhancement by unsharp masking allowed the extraction of more information than was originally perceived on the film. In these images, however, film grain noise accounts for at least one-third of the total image noise, and thus limits the degree to which the images can be enhanced. To investigate the feasibility of recording digital mammographic images directly, an x-ray image intensifier (XRII) was evaluated. With the use of geometric magnification and small focal spot x-ray tubes, the XRII can produce images with sufficient resolution for mammography. Noise power spectra of both XRII and film-screen images were measured. The XRII was found to have better sensitivity and, if structural mottle can be removed, a higher signal-to-noise ratio than mammographic film-screen systems; however, it is limited by the high attenuation of the front input window at mammographic energies.

We are investigating, both theoretically and experimentally, the application of dual-energy techniques to the detection of calcifications in mammography. A prototype xenon ionization detector, developed for scanned projection digital radiography, is used to obtain images at two energies. These image pairs are processed to produce pairs of "material" images which give the equivalent thicknesses of lucite and aluminum of the imaged objects. The material images are then combined such that the final image has minimal soft-tissue contrast. When this has been done, calcifications may be more easily visualized due to the removal of soft-tissue background "clutter," despite increased image noise. The dual-energy procedure is described and experimental images of a phantom and breast tissue are presented.

"DIGITAL RADIOGRAPHY" is a generic term which has evolved recently to describe a collection of novel radiographic modalities where the primary image recording is accomplished through the use of x-ray-sensitive electronic detectors, rather than by the conventional use of phosphor screens and silver-halide-based films. The electronic detectors provide occult electronic signals which must be processed to render them ultimately visible and thus accessible to the diagnostician -- in the short term for diagnosis, in the long term for reference. The type of imaging involved is metric both in the spatial and the luminous dimensions, with strong requirements for imaging repeatability and image element signal integrity.

An analysis is presented of the application of a scanning linear X-ray detector to dual kVp chest radiography, and the single kVp imaging of anatomical areas other than chest. Clinical evaluation of an experimental digital chest unit utilizing a solid state linear detector array has provided information regarding the post-patient photon flux required for clinically acceptable images. Results from computer simulation show that successful energy subtraction imaging of the chest by the method of kVp switching will be unsuccessful unless it can be demonstrated clinically that a much lower number of photons per pixel is accept-able in energy subtracted images. In addition, it is shown that X-ray tube loading limitations preclude imaging of the abdominal and pelvic regions.

The first few years of any new imaging technology are characterized by rapid developments in equipment. Digital subtraction angiography (DSA) is no exception. These improvements must be accompanied by an increase in understanding of the interaction between equipment and the patient. Several articles have discussed the impact of various components on the signal to noise ratio (SNR). While most users have recognized the importance of certain biological factors, e.g. heart rate, cardiac output, etc., only recently has the equipment evolved to the state permitting systematic study of these variables.

In this paper we consider the choice of the magnetic field for an imaging system based on the nuclear magnetic resonance of hydrogen. We show by analysis that the quality, or contrast-to-noise ratio, of images based on T1 discrimination increases with field or frequency up to 2 T or 85 MHz. After a brief discussion of potential engineering limitations we present results showing that images of the human head with excellent anatomic detail can be produced at 1.5 T or 64 MHz.

In October 1982, medical magnetic resonance (MMR) clinical trials were initiated using a 0.15 Tesla resistive magnet. Measurement of standard image quality parameters were attempted in the first months following the installation. MMR signal strength is a complex interaction of tissue Tl, T2, spin density, and flow (including diffusion). Further, the operator selectable controls of TE (time to echo), TR (repetition time), and TI (inversion time) all tend to complicate any unique specification of the MMR signal. However, MMR can be approached from a conventional tomographic imaging problem. Image quality in MMR1suffers from partial volume affects and thus spatial resolution including slice thickness must be specified. System noise with MMR differs from all other conventional radiographic imaging modalities in that the sgurce of noise does not include a component that is derived from the statistical nature of the signal. Uniformity of signal response over an imaging volume in MMR is a prerequisite for quality examinations. Failure of the RF coil to provide a uniform secondary magnetic field results in a tip angle of other than the desired value. The spatial dependency of the secondary magnetic field is a function of the coil design.

Initial discussions were carried out in May of 1981 regarding a suitable site for magnetic resonance imaging facility. At that time minimum information was available regrading the special environmental demands required for this new diagnostic modHlity. What was known for certain was that the whole body superconducting magnets were physically large (8 x 8 feet) and could not be delivered in a conventional elevator or through standard passage ways. The weight of a large superconducting magnet (fifteen thousand pounds) also precluded the siting of this unit above grade except in cases of special load bearing floors. For this reason it was decided to initiate discussions with an architectural firm to arrive at a suitable design for a standalone medical magnetic imaging facility. Initially, the siting criteria were based upon magnetic field strengths from two units of between 0.15 and 3.5 Tesla. In time however, discussions centered upon placing two magnets of 0.5 and 1.5 Tesla with a design basis for the facility of two units of 3.0 Tesla each.

In nuclear magnetic resonance (NMR) the image pixel value is governed by four intrinsic parameters: the spin density ρ, the spin-lattice relaxation time T1, the spin-spin relaxation time T2 and, for non-stationary spins, the flow velocity v. The extent to which the signal is weighted toward one or several parameters is related to the history of the spin system preceding the detection pulse. In the present work T1 and T2 were determined in vivo for several regions in the CNS from inversion-recovery (T1) And multiple-echo (T2) images, using least-squares fitting procedures. From averaged values of T1 and T2 in grey matter, white matter and CSF, the signal intensity was calculated on the basis of the Bloch equations and plotted as a function of the intrinsic parameters for the three most common imaging pulse sequences. These data are in excellent agreement with images, recorded from normal volunteers on an experimental whole-body imaging system operating at 12.8 MHz (0.3T). The graphical presentation of contrast further will provide the radiologist with a straightforward tool for image interpretation.

Until recently the idea of using permanent magnets for magnetic resonance imaging has been ignored in the literature. This paper discusses the technical aspects of the design of permanent magnets for whole body imaging and describes the differences between permanent magnet systems and those using resistive or superconducting electromagnets. Material properties, magnetic circuit characteristics, and the problems of scaling to whole body size are discussed. Also discussed are the problems of site preparation, maintenance and repair.

A range of multiple exponential transverse relaxation values (T2) was obtained for a variety of benign and malignant conditions in human breast tissue. In breast tissue containing predominantly fat or fibro-fatty tissue, there was good discrimination between benign and malignant samples based upon T2 values. In the fibrous tissue samples there was significant overlap in relaxation values of benign and malignant tissues. In the fibrous breast the number of carcinoma cells had to be greater than 20% before discrimination could be achieved.

Since the early 1970's, Edholm and others have shown that compensation attenuators can provide fundamental improvements in radiographic image formation by modifying patient dose-distribution, reducing scatter fractions in dark regions of the image, and easing dynamic range requirements for film and video systems. Compensating attenuators have not yet been used widely in diagnostic radiology because of the difficulties of forming masks to compensate for the anatomical variations of individual patients. We have designed a software-based system which forms a heavy-metal attenuator from a digital image of the patient. At present, the attenuators are constructed manually from a pattern generated by the computer, but several techniques are being investigated which may permit fabrication and positioning during suspension of respiration. Phantom studies demonstrate that, in nonsubtractive applications, unsharp masking by the x-ray beam attenuator enhances local contrast, while in digitally subtracted images, attenuators eliminate dark regions where iodine signals otherwise are degraded by video and quantum noise. The technique can be used to reduce patient exposure in highly transmissive areas or, at the expense of increased tube loading, to increase exposure to highly attenuating areas in order to reduce image noise. Anticipated applications of this technique include chest radiography as well as conventional and digital subtraction angiography.

In this paper, we survey existing various impedance techniques and we propose to develop a new imaging technology, computerized impedance tomography (CIT),for imaging the thorax. By applying high-frequency current and measuring its distribution, it should be possible to reconstruct the impedance images of the body noninvasively and nondestructively for the frontal plane, the transverse plane, and eventually three dimensions. We overcome difficulties in the reconstruction procedure due to nonlinear and nonplanar current paths by solving Laplace's equation numerically in our model for every iteration and by using a nonlinear back-projection reconstruction algorithm to obtain the impedance image iteratively. Results of our simulation study with computer-generated phantom objects in the computer model indicate that reasonable impedance images in one transverse plane of the thorax can be obtained with 8 current projections. We discuss advantages and disadvantages associated with impedance imaging, the computer model, important variables affecting the quality of reconstructed impedance images, and back-projection algorithms used in reconstructing impedance images. We present reconstructed impedance images with 8 projection angles and different projection methods. We summarize numerical aspects, computer requirements, and limitations of impedance imaging. We also discuss the future of impedance imaging.

Radioluminescent imaging, as used in this discussion, is defined as the mechanism wherein a phosphor coating absorbs an x-ray photon and re-emits light photons in the visible spectrum. This visible light, in turn, can expose a photosensitive film, excite a photodetector, or be viewed directly. Mathematical models are developed for three different juxtapositions of radioluminescent coating and photoreceptor planes. These models predict the total light output reaching the photoreceptor plane as a function of x-ray and optical attenuation coefficients, coating weights and intrinsic phosphor properties. Although the analysis is one dimensional, the model is shown to be extremely useful in actual designs of radioluminescent imaging systems.

To meet the complex needs of a nuclear medicine division serving a 1100-bed hospital, a computer information system has been developed in sequential phases. This database management system is based on a time-shared minicomputer linked to a broadband communications network. The database contains information on patient histories, billing, types of procedures, doses of radiopharmaceuticals, times of study, scanning equipment used, and technician performing the procedure. These patient records are cycled through three levels of storage: (a) an active file of 100 studies for those patients currently scheduled, (b) a temporary storage level of 1000 studies, and (c) an archival level of 10,000 studies containing selected information. Merging of this information with reports and various statistical analyses are possible. This first phase has been in operation for well over a year. The second phase is an upgrade of the size of the various storage levels by a factor of ten.