For more than five years now the Bureau of Radiological Health has been making quantitative measurements on the physical imaging parameters of radiologic imaging systems' components. This work has taken place in the laboratories of the Bureau's Divison of Electronic Products. Our earliest efforts addressed questions concerning the performance of films, intensifying screens, and the x-ray tube focal spot. In order to seek a consensus on methods for measuring such quantities as sensitometric response, modulation transfer function (MTF), and noise power spectra for film/screen systems, MTF and summary measures of performance of focal spot, image tubes, and n9w image detection schemes, we presented a symposium entitled "Medical X-Ray Photo-Optical Systems Evaluation" in Columbia, MD in October of 1974. Many of the papers presented by experts in the field at the Columbia symposium might well serve as draft documents for standardization of several measurement techniques, most notably those on film/screen systems.

Presently available descriptors of image quality can be sorted into three categories: physical descriptors of the imaging system, measures derived from these, and empirical descriptors of human observer decision performance. A basic problem today is our inability to relate the first type of description to the third. Signal detection theory is proposed as a useful tool for attacking this problem.

Recent advances in technology have made radiographic magnification of the skeleton clinically feasible. A new electron gun micro-focal spot tube combined with new high-resolution recording systems were used to perform magnification radiography, which was then compared with conventional contact radiography. Quantitative evaluation included measurements of speed, contrast, resolution, and noise. Qualitative evaluation included an analysis of 215 clinical cases in which both techniques were used. The superior image quality of direct radiographic magnification is confirmed and the clinical areas in which it proved most helpful are defined.

With the radiographs shown i Fig.1, the late Dr. Kurt ossmann used to demonstrate some of the fundamentals of radiologic imaging by means of simple test objects and screen-film systems. The plastic beads simulate low-contrast, round objects such as gall stones. The needle corresponds to a high-contrast, long object such as an °pacified blood vessel. The image on the left is made with a sharp, but noisy system; that on the right is made with an unsharp, less noisy system. The speeds of the two systems are identical. Detail visibility of the needle is better in the system on the left; therefore, detail visibility is sensitive to the resolution or sharpness of the system. The detection of plastic beads, however, is easier on the right and therefore is ilfluenced more by the noise or r.diographic mottle. Although these test objects are very simple, Dr. Rossmann was successful in illustrating not only the effect of these b.sic parameters on the radiographic images, but also the complexities involved in diagnostic radiologic imaging.

A method using radiologists' subjective judgments was developed to compare the quality of the diagnostic image information from two different screen/film combinations (Hi Plus/RPL versus Lanex/Ortho G). A sample of 148 comparison film pairs was obtained in the course of performing 74 urograms using the two film/screen combinations. Each film pair was evaluated by three radiologists, using a blind film reading format, in regard to: 1) anatomic diagnostic criteria visualization, 2) radiologic diagnostic certainty, 3) presence or absence of quantum mottle, and 4) prediction of which film of a pair was the rare earth screen/film combination. There was a significant difference favoring Hi Plus/RPL in perceived quality of visualization of anatomic criteria. However, the difference was more statistical than of practical importance. Diagnostic certainty differed only marginally and slightly favored the Hi Plus/RPL combination. No significant differences in perception of quantum mottle were attributable to either screen/film combination. Radiologists were able to correctly identify the screen/film combination a significant proportion of the time.

The widely used DuPont Lo-dose mammography screen-film system was compared with four combinations of higher efficiency screens and single-emulsion films. Laboratory comparisons of the basic image properties (resolution, noise, contrast) of all these screen-film combinations were made both qualitatively (images of microwire meshes, tiny plastic beads, step wedges) and quantitatively (modulation transfer functions, Wiener spectra, step wedge optical density differences). Subjective single-blind estimations of resolution, noise, and contrast were also made directly from the mammograms of 100 consecutive patients radio-graphed with each of the five screen-film combinations. Both laboratory and clinical evaluations of image quality showed strikingly similar results: improved contrast, minimally reduced resolution, and varying degrees of increased noise for each of the high efficiency recording systems. Similarities between the results of laboratory and clinical studies should help to establish the clinical relevance of the laboratory analyses, and recognition of the relative precision of the clinical evaluations should allow for their use in place of laboratory studies when sophisticated physics support is not available.

There is a mystique associated with diagnostic radiology that discourages the use of analysis as a means of understanding and improving radiological examinations. While it is true that intuitive manipulation of the generators and detectors that are available can lead to satisfactory optimizations, it is also true that the equipment that is available may not necessarily have the correct characteristics. Far better optimums may be possible with equipment of different, but state of the art, capabilities or characteristics. This paper presents two cases of modelings for mammography that are examples of this situation. It remains to be established whether or not the modeling has diagnostic significance. However, since the conclusions appear to be significant it seems important to pursue such projects. To accomplish this, it is required to generate information and data that can be useful to the modelings. The extensive imaging research that has recently occupied many diagnostic radiological physicists is only part of the necessary information. More emphasis needs to be given to research that is useful in helping to describe the details of the x-ray photon image. One of the examples of modeling that is discussed is related to signal to noise ratios for the detection of calcifications in mammography. The other is a study of contrast in magnification mammography.

The objective of this work is to relate the performance of radiologists as measured by the probability of detecting a nodular abnormality on a chest film to the properties of the nodule and its surrounding structures measured by a microdensitometer from the film. A parameter called conspicuity has been defined. It consists of measurements of the size, shape, contrast, and the edge gradient of the nodule divided by a measurement of the complexity of the surrounding structures that tend to camouflage the nodule. Conspicuity has been measured for simulated nodules and for a series of real nodules that were originally missed in clinical practice but later detected. Correlation between film reader performance and conspicuity was found.

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 results of two types of visual detection experiments will be presented and discussed. One set of experiments was done using single bar targets optically superimposed on a variety of backgrounds. The experiments were designed to determine functional relationships between the target contrast at the detection threshold and a number of variables. The variables included bar angular width, bar angular length, back-ground noise level, and viewing distance. The second set of experiments was done using radiographs of aluminum wires. The radiographs were produced using a conventional neuro-angiography system with a 0.38 mm focus. The aluminum wire diameter at the detection threshold was determined as a function of geometric magnification for several screen-film combinations. The aim of the experiments is to determine how to calculate visual signal-to-noise ratios. Some progress has been made toward this goal.

The clinical value of radionuclide images depends on many factors, some controllable and some not. Noncontrollable factors include signal-to-background ratio, contrast gradient, and just-perceptible density difference. Several controllable factors affect the level and proportion of false negative and false positive results: the number of counts in the image, the signal-to-noise ratio, film gamma, contrast, viewing distance, computer and other image manipulation techniques, and selection of the criterion for calling a study abnormal. It is not always true that the more counts an image has, the better it is. For typical clinical situations the number of counts required in an image ranges from 50,000 to 2,000,000, depending on signal-to-noise ratio and film gamma. The clinical value of the images does not necessarily increase as the number of counts increases, the improved resolution and SNR being balanced off by patient motion, field nonuniformity, film nonuniformity, and loss of apparent contrast at high count density. Signal-to-noise ratio varies between two and five in typical clinical situations, and can be improved by better counting statistics. The signal-to-noise ratio is to be distinguished from the signal-to-background ratio, which is not a controllable parameter. Correct viewing distance is an often overlooked parameter. The eye acts as a bandpass filter, perceiving objects with greatest sensitivity when the object occupies between 5 and 10 minutes of visual arc. The correct viewing distance depends on the size of the image being examined, which in the case of Polaroid images varies between 0.1 and 1 cm, corresponding to optimum viewing distances of 30 cm to 3 m. Some organs, such as the liver, may contain filling defects that are smaller than the lower limit of resolution of the imaging system. A technique has been developed in which the observed fluctuation of count rate over the liver is compared with the expected fluctuation (as derived from the number of counts ner unit area). When the observed fluctuation is significantly greater than the expected fluctuation, liver uptake is said to be "nonuniform" despite its appearance to the naked eye. Work is still in progress in this area but initial results have shown a reduction in false negatives in widespread fine metastases and degenerative liver disease in the absence of focal defects as seen on liver scans. Finally, the incidence of false negatives and false positives is affected by the choice of selection criterion by which one decides when to call an image "abnormal." If one knows the approximate distribution of values (such as relative counts per unit area) in normal and abnormal studies, one can maximize the utility of the diagnostic study by selecting an appropriate criterion that is based on the utility of false negative and false positive outcomes. In other words, if the observer is able to formulate a trade-off between false negatives and false positives which would, in his mind, equalize the penalty for being wrong in either case, he can develop a diagnostic strategy that will be reproducible and that will maximize the utility of a test as he sees it. Another observer having a different trade-off can maximize his own utility using a different criterion, and the two observers can estimate what fraction of the normal and abnormal population they would disagree on, based on their separate criteria.

The introduction of new diagnostic tests and the excessive use of existing tests have come under increasing scrutiny today for two reasons. First, the proportion of the Gross National Product devoted to health care is increasing and, second, expensive technological devices are being promulgated rapidly for both diagnostic and therapeutic purposes. Since 1965 the total health care allotment has trebled to $122 billion and its percentage of the Gross National Product has increased 41 percent (Table 1).(1)

Sensitometric methods employed for determining the characteristic curve of interest in medical imaging are described. These methods are based on the way in which various recording systems are exposed. In diagnostic radiology, an x-ray intensity scale sensitometer is used for measurements of the characteristic curves of conventional film and screen-film systems. The effect of reciprocity law failure on the sensitometric evaluation of screen-film systems is illustrated. The shape of the characteristic curves as a function of the light spectrum emitted from three types of intensifying screens is presented. Sensitometric techniques for quality control of automatic processors are also discussed. In radiation therapy, a sensitometric technique is described for determination of the proper characteristic curve for a film which can be placed beneath the patient before radiation treatment and removed afterwards so that an image of the anatomy actually irradiated is obtained. A characteristic curve can also be obtained by means of which photographic densities measured on the therapy verification film is related to exit absorbed dose.

This paper will present a definition of Sensitometry in Computerized Tomography and discuss practical techniques used to determine relative CT sensitometry. Several factors that influence CT sensitometry are: (1) the x-ray beam spectral distribution, (2) the composition of the sample materials, (3) volume averaging, and (4) spatial nonuniformity of the scan field including the x-ray intensity distribution within the sample volume. Examples of comparative sensitometry among several CT units is presented as well as discussion of a technique for obtaining a continuous rather than discrete output.

Perception of a lesion in a Radiograph or Nuclear Medicine image requires that lesion contrast exceed both a threshold contrast value for the human visual system, and the contrast of any random fluctuations due to quantum mottle in the image. In principle, any lesion contrast value can be raised above visual threshold by choosing a film with sufficiently high gamma, and the quantum mottle reduced to any desired minimum by using a high enough photon count density in the image. In practice, one must compromise these parameters within the constraints of limited dynamic range for the film and the need to minimize discomfort and radiation dose to the patient. This talk will show how quantitative relationships can be developed to calculate the minimum detectable lesion contrast value and maximum useable photon count density for any film. In the development of diagnostic imaging procedures, these relationships offer considerable assistance in optimizing the choice of film with regard to the characteristics of the human visual system and the limitations of present-day instrumentation.

A new standard method for the sensitometry of X-ray screen-film-processing systems is under development by ANSI Subcommittee PI12-31. Devised not only for systems manufacturers but also for the ultimate system user, the method employs radiation qualities simulating those for X-rays emergent from human phantoms. Standard test objects are specified that approximate the primary to scatter ratios found for three commonly used techniques (chest, skull and pelvis, and extremity radiography). The method permits the determination of system speeds for each technique under scatter and under non-scatter conditions. By specifying the use of aluminum modulators with the test objects the method also measures contrast, or the system's capability of visualizing changes in object thickness.

Intensifying screens are a significant source of blur in most medical x-ray imaging systems. In the analysis of x-ray system performance it is desirable to specify intensifying screen blur in terms of a single parameter which can be easily combined with focal-spot blur. This is achieved by describing screen performance in terms of equivalent blur, which is defined as the dimension of a square blur pattern which has a uniform intensity distribution. This paper describes a simple method for obtaining equivalent blur values from the MTF of intensifying screens and gives the values for representative screens.

A proposed ANSI standard for classifying radiographic intensifier screens has been under test at the National Bureau of Standards. This standard establishes procedures for characterizing, on an absolute basis, the optical spectral output of fluorescent screens per unit of incident x-ray exposure. The testing procedure has undergone revision since an earlier status report was given. Calcium tungstate screens, however, still form the basis of comparison in this procedure because of the long acceptability and stable output qualities of this screen. The nature of the revisions and current output data on the screens will be described.

Reciprocity law failure was measured for four brands of medical X-ray films exposed with intensifying screens. Three of the films are green light sensitized for use in combination with green light emitting rare-earth screens. These films showed larger reciprocity failure effects than one conventional blue sensitive film, Dupont Cronex-2. The development conditions had a small effect on reciprocity failure. The effect of reciprocity failure on the measurements of physical image quality parameters is discussed.

The observed higher contrast of radiographs made with rare-earth screens as compared with calcium tungstate screens is shown to be associated with the scattered component of the radiation absorbed by the screens. This is confirmed by screen absorption measurements of the primary and scattered radiation transmitted through two regions of a chest phantom. The primary/scatter absorption ratio (or contrast) is higher for the rare-earth screen, indicating lower relative response to the scattered radiation. It is also shown that, although the contrast becomes lower for the rare-earth phosphor as compared with the calcium tungstate when the scatter is reduced, the information-content in the radiation absorbed remains higher for the rare-earth screens.

Radiographic signal detection is considered to be a point process. Realism of radio-graphic signal detection experiments is extended by allowing a random number of signals and by requiring accurate location of the signals. A substantial improvement in the simplicity of the theoretical framework of multiple signal detection experiments results from using this free response approach. The generation of false positive events is adequately described as a Poisson process, represented by a single measurable parameter. A scheme for reinterpretation of earlier conventional Receiver Operating Characteristic (ROC) curves in terms of the free response approach is presented, and a description of a much simpler and more efficient experimental procedure is provided.

The wide variation in patient exposures for the same examination at different facilities, as shown by the Nationwide Evaluation of X-Ray Trends (NEXT) data, has been previously reported. In order to investigate this variation, an image receptor module (IRM) was designed to be used in conjunction with the NEXT system. It has been used to collect information on the types of films, screens, and grids used for over 900 radiographic exams in 12 States. Numerical relative speed values were experimentally determined for many of the more frequently used film/screen combinations, enabling a speed index to be assigned to these combinations when reported in the IRM. Principal findings to be discussed are the variation in film/screen speeds used for a given examination; the variation in exposures at skin entrance for a given examination, even when film/screen combinations of the same speed range are used; and relationships between the speed of the image receptor system and the exposure at skin entrance.

In 1975 the concept for our hospital based x-ray service group was formulated. Several factors including increased vendor supplied service cost for parts and labor, maintenance requirements on new equipment, heavier equipment workloads, and increased x-ray room down time, contributed to our decision to develop a proposal for an "in-house" service group. Using a systematic approach to ascertain service requirements on x-ray equipment, a decision based on cost effectiveness and other advantages can be made with respect to formation of a hospital based service group. Service requirements were determined by obtaining data relating to vendor supplied costs (labor, travel, and parts); the size of the department; complexity of equipment; and departmental demand. A comparison of cost for vendor supplied service and budget requirements for a hospital based organization was prepared. At present we have eight x-ray service engineers to meet our requirements. The purpose of this presentation will be to discuss our approach to the decision to pursue a hospital based service organization and to compare our results for six hospitals.

The improvement of radiographic quality has been a subject of concern to physicians, physicists, and design engineers ever since Wilhelm Conrad Roentgen produced the first visual image of human skeletal structure. However, despite exhaustive attempts to define and quantify the multitude of parameters which effect the quality of the image, the final result must succumb to the subjective opinion of the radiologist. It is generally accepted, however, that patient motion, which may be voluntary, due to disease, or due to normal autonomic movement of viscera, serves to degrade radiographic quality by blurring the image. Conversely, any method which serves to reduce the effects of motion will also serve to improve image quality. The variety of methods which have been applied to reduce motion unsharpness may be grouped into three categories: 1. The use of drugs 2. Physiological Gating 3. The reduction of x-ray exposure time

For a given patient (attenuator) thickness, substantial differences in the x-ray tube potential can be observed on the image intensified fluoroscopic x-ray installations equipped with kVp variable automatic brightness stabilization (ABS) circuitry. The ftinctional behavior of such fluoroscopic x-ray equipment is investigated through measurement of (1) the fluoroscopic x-ray tube potential, (2) the fluoroscopic x-ray tube current, (3) pie entrance exposure rate, and (4) the exit exposure rate as functions of the attenuator thickness.

Certification of major components of x-ray equipment by the Bureau of Radiological Health of the Food and Drug Administration has created a renewed interest in the performance of diagnostic x-ray systems. Although the certification performance criteria, major component specification data, and assembly data is submitted by the manufacturers to the Bureau of Radiological Health; assembly and calibration of the diagnostic x-ray system depends on the assembler. Even though x-ray system components exhibit certification labels, there is no guarantee that the systems meet the compliance standards. In addition to the manufacturers' specifications and HEW requirements, we have developed supplemental performance specifications. Acceptance testing of diagnostic x-ray equipment prior to final payment has been used to check that the manufacturers meet the criteria established by BRH; to check the calibration of the various x-ray systems (generator/tube, collimator, automatic exposure control, etc.); and to develop supplementary testing criteria. The performance criteria, specifications, and equipment acceptance tests, including before and after measurements for several units, will be discussed. The results will demonstrate that our desired specifications can be achieved.

When the decision to replace equipment in a Radiology room is made, the process which follows is:

In the past two decades, new radiographic devices enhanced by advanced space age electro-mechanical engineering, i.e., automated x-ray generators, examination tables, automatic collimators, high speed patient to radiologist film transports, highly sophisticated closed-circuit television systems, disc and tape storage of x-ray produced video images, and others continue to come off the manufacturer's drawing board and into radiological departments daily. This new technology intensifies a longstanding and rapidly growing problem the severe shortage of trained and qualified professional radiological engineers. Equipment manufacturers and their representatives have been hard pressed to provide competent service to this ever increasing demand. Recognizing the enlarging dimension of the problem, the Department of Radiological Sciences, University of California, Los Angeles, in March of 1970, took two giant steps first, the establishment of an in-house radiological engineering staff and second, the development of a two year training program for Radiologic Electronic Specialists. Class number eight has convened with their studies commencing on August 1, 1977.

Since only electrons which have relatively high energies can escape from a metal, a thermonic emission current, when initiated, reduces the average energy of the remaining electrons and molecules and hence, if not compensated for, the temperature of the emitter. Investigations of several diagnostic x-ray tubes and generators indicate that there is a noticeable, exponential fall-off of x-ray tube current during an exposure (time constant = 130 msec) attributable to this phenomenon. With three-phase generators the equilibrium values observed were approximately 15% lower than the corresponding values at the beginning of the exposure. Smaller fall-offs were observed with single-phase equipment. Theoretical expressions derived from basic thermodynamic considerations, which are in good agreement with observations for the magnitude of the effect and its time constant, are presented.

In this paper, we will report on a detailed study of fluoroscopic spot film systems and the methodology we have developed for achieving optimal clinical image quality at minimum patient exposure. In addition to those quantities that affect the image quality in routine radiography, the spot film image is dependent upon the gain and resolution ability of the image intensifier, the functioning of the automatic exposure timing system, and the imaging optics of the spot film camera. Hence, a detailed examination of these systems in new equipment, and an on-going program to monitor both the quality of the images produced and the condition of the imaging system components is required. We will report on the procedure we have developed for the complete examination of spot film imaging equipment and the means for assuring their optimal performance. Clinical films made on a series image intensifiers will be compared to the physical imaging properties and patient dose rates of these systems. A simple method for the daily check of spot film image quality will be presented. The wide variation in the imaging ability of new spot film systems will also be discussed.

A novel proximity-type x-ray image intensifier tube is described. This device has the main advantage of fewer sources of unsharpness so that this advantage can be traded for higher MTF performance or dose reduction performance. In addition, the size and shape of this device are not restricted as in the case of current minifying-type x-ray image intensifier tubes. Various systems applications employing this device, such as in simple direct-view intensified panel-like fluoroscopy, low dose TV fluoroscopy and direct radio-graphic miniaturization, are discussed.

A pilot study for potential applicability of a coherent optical. Fourier transform technique to differentiate between healthy and pathological human blood cell smears has been conducted. Types of abnormalities tested include acute myelogenous leukemia, sickle-cell anemia, target-cell anemia, acute lymphatic leukemia, infectious mononucleosis,and eosincphila. Although limited numbers of slides and restricted types of abnormalities were examined, normality and various abnormalities are distinctly identifiable through coherent optical transform analysis. The morphological and constitutional variations have been qualitatively examined as the sources of both the observed deviations in the characteristic transform patterns and the photographic or densitometric profiles. Limitations in the presently employed optical Fourier transform method and the prerequisite conditions for standardized sample preparation are briefly discussed.

The characteristic lines of molybdenum-anode x-ray tubes yield high contrast images on film of thin breasts. For thicker breasts, the higher spectral energies are transmitted and the lower energies are mostly absorbed, resulting in higher patient exposure and reduced contrast. However, filters from rhodium (Z=45) to tin (Z=50) can be used to shape the energy distribution of tungsten-anode tubes to improve contrast and reduce patient exposure. Measured spectra and phantom test data are described.

A system has been developed to allow single frame acquisition of superior quality digitized video images. The system consists of a high-quality x-ray image intensifier coupled to a high-resolution, low-noise video system. The output video signal is fed to a digital memory having 512 x 512 pixels and 8-bit accuracy. The memory output is directly interfaced to a PDP 11/70 image processing facility to allow computer controlled acquisition of video imagery as well as real-time visualization of computer processed images. Computer pro-cessed images are returned to the memory for display. Techniques and examples of processed radiological images will be discussed, and examples shown. Additional digital memories are under construction for 12-bit dynamic range.

This paper describes considerations that are important for the successful application of x-ray image intensifier video systems to diagnostic radiology. Particular attention is given to spatial resolution amd maximum signal-to-noise ratio.

Psychophysical contrast and resolution responses obtained from r.diologists in two experimental studies are compared with similar data from previous work by the authors. Five observers participated in two experiments that were designed to compare the effectiveness of two video systems (a standard system set at 525 scan lines and a high-resolution system set at 1023 scan lines) for displaying nodules of different size and contrast. We co elude, based on the results of all studies, that video systems exist that have a sufficient contrast transfer to allow detection accuracy similar to that achieved with film.

The history of radiographic duplicating technology has at times been concerned with improving the extraction of information from radiographs. Recent work in diazotype radiographic duplicating film has resulted in a continuous tone direct positive film which is capable of visual enhancement of edge-sharpness and contrast. The convenience it affords of room light handling contributes to its value as an interpretive and diagnostic aid in conjunction with the original radiograph. Specific examples are discussed.

The physician's need for standardization of ultrasonic instrumentation might appear to be too obvious a requirement to be worthy of further comment. Nevertheless, many engineers do not appreciate the extent to which the diagnostic process in ultrasound as currently practiced is dependent upon the subjective impressions of the interpreter and the variations that occur in similar models of the same machine or the drift in any single machine with time.

The technological advance of gray scale imaging has been a major factor in the rapid growth of the clinical application of diagnostic ultrasound. Gray scale imaging implies that the displayed brightness of a target in the ultrasound image is a function of the amplitude of its echo. Several techniques have been investigated for the measurement of the dynamic range of gray scale ultrasound scanners. The measurement methods include (1) an electronic pulse burst generator which emits a sequence of pulses decaying at a selectable rate of 0, 1.25 or 2.5 dB/cm, (2) a polypropanol wedge designed to attenuate acoustic echoes over a range of 0 to 60 dB, (3) a series of stainless steel reflectors varying in angle from 0 to 10° so that returning echoes vary over a 40 dB range. All three instruments are easy to use. However, the wedge and variable angle devices must be calibrated for use with individual transducers. The electronic test device excludes the transducer when testing system dynamic range. Therefore, the pulse burst generator seems to be the most suitable for clinical or survey applications at this time.

The main contributions of the AAPM and AIUM in ultrasound instrument performance evaluation are made through special sessions, individual members' presentations at the national meetings, and in publications of these two societies. Committee work is proceeding within the AAPM on a draft of recommended acceptance tests. In addition, the AAPM provides coordination for the NCI-sponsored Centers for Radiological Physics. The Centers are presenting regional workshops on ultrasound instrument performance evaluation and are developing techniques such as meaningful axial resolution measurements for gray scale scanners. These developments are summarized. Two existing ALUM standards are designed for performance evaluation of ultrasound equipment. These standards are "Standard Specification of Echoscope Sensitivity and Signal to Noise Ratio" and the "Standard 100 mm Test Object and Recommended Procedures for Its Use." The ALUM now offers a commendation to manufacturers who specify appropriate safety related instrument performance information, and recently has begun work to cooperate with the National Electrical Manufacturers Association, the AAPM and the Acoustical Society of America in helping to define standards for acoustic safety aspects of diagnostic ultrasound equipment. AIUM subcommittees also are active in ophthalmological equipment evaluation and transducer characterization standards.

The Standards Committee of the American Institute of Ultrasound in Medicine uum and the National Bureau of Standards (NBS) cosponsored a meeting on Pulse-Echo Ultrasonic Transducer Characterization at the NBS in Gaithersberg,Maryland, June 16, 1977. More than 100 people from both the medical ultrasound and non-destructive testing communities attended. Initiation of work on a standard for these transducers was the purpose of the meeting. A morning session of invited and contributed papers was followed by the organizational meeting of the ALUM Subcommittee on Pulse-Echo Transducer Characterization. The purpose of this paper is to discuss standardization problems of the transducer, a vitally important and often neglected portion of a diagnostic ultrasound system. A brief summary of the ideas discussed that day and the direction chosen by the committee is included.

This paper discusses needs for tissue equivalent test objects for use in ultrasound quality control. Currently available test objects are useful for assessing specific aspects of an ultrasound imaging system such as depth calibration and B-scan registration accuracy. A phantom which yields scattered echo signals of a similar magnitude as detected in tissues and producing the same frequency dependent attenuation of the beam would be useful for scan uniformity checks, evaluation of swept gain settings, and possibly for gray scale checks. The phantom material would also be useful for measuring pulse-echo response profiles of ultrasonic transducers.

The Food and Drug Administration is currently planning to publish for comment a Proposed Recommendation to encourage diagnostic radiology facilities to establish their own quality assurance programs. These programs will improve the quality of health care by reducing unnecessary patient radiation dose, improving the quality of the diagnostic image, and reducing the cost of radiological services. The general concepts of the Proposed Recommendation are discussed.

Our experience with Quality Control Programs in many x-ray departments shows that the reject rate can frequently be reduced about 45% through proper controls. A Reject Analysis establishes the current reject rate and its associated film cost. Based on the reject analysis, appropriate corrective programs are undertaken. Typical corrective programs are processor monitoring and maintenance, screen/cassette program, inventory control and technique consultation. Six months to one year after the Reject Analysis, a follow-up Reject Analysis measures the effect of the Quality Control Program. Data compiled from many x-ray departments indicate an average reject rate of 13%. The chest exam usually accounts for the most repeats. Incorrect density and poor positioning are the major reasons for repeated exams. The Quality Control Program usually reduces the reject rate to about 7%, a 45% reduction. If the annual film cost is $150,000, the savings from this reduction can be as much as $9,000 in film alone. Other costs such as chemicals and equipment are also reduced.

Although large hospitals have access to sensitometric testing tools, the majority of hospitals do not have sufficient capital equipment budget to afford this sophisticated special purpose equipment. By using refrigerated pre-exposed strips and visual analysis, it is possible to monitor the photographic parameters of the speed, contrast and gross fog in a reasonable manner. This paper considers the methods employed in this processor monitoring procedure and offers evidence regarding the reliability of using refrigerated pre-exposed processor control strips.

Two types of fast fan-beam CT body scanners are in widespread clinical use. A bi-modal type employs a series of linear translations followed by rotational steps. Rotary scanners use pure rotary motion of the source and detector or source alone (stationary detector type). We have conducted comparison studies using production models of an EMI-5005 and a GE CT/T installed at the Department of Radiology, U.C.S.F. These include quantitative determinations of noise and resolution, semiquantitative comparisons using phantoms, and qualitative evaluation of CT images obtained from double-scanned patients. Image quality of the two machines is roughly comparable under normal use. The most important differences involve the frequency and character of artifacts. The smoothness and speed of the rotary system's scanning motion greatly reduces that systems sensitivity to patient-motion-related streak artifacts. On the other hand the translate-rotate system currently offers higher spatial resolution and optional slow, high-dose, high-sensitivity scans.

To date, in Computed Tomography (CT) various tech-niques have been utilized for measuring and reporting exposure. This report outlines some techniques for measuring exposures in CT. Due to the x-ray field geometry exposure measurements must be made carefully to assure accurate results. Films and TLD have generally been used and can be very time consuming. These two methods will be discussed. The utilization of an ionization chamber specifically designed to measure CT exposures will be introduced.

The geometrical limitation of the spatial resolution inherent in a CT scanner has been analyzed. In order to assess this limitation, an equation for calculating the theoretical attenuation profile optical transfer function (OTF) from the geometrical dimensions of the system was derived. Four commercially available CT units were studied which are representative of the three basic design types in use. For each unit the OTF was calculated and the results are presented. Using this information in conjunction with the pixel size and data sampling frequency of the attenuation profiles, an analysis of the factors limiting the spatial resolution of each unit was made. In addition, subtle effects on the image quality of a scan due to the geometry of the data-taking mechanics are discussed and correlated with the observations of experimental comparisons. A suggestion is made for the optimal design of a CT unit that will encompass the conclusions of this analysis.

If the standard filtered backprojection algorithm with a filter of the form g(f) = |f|h(f) is applied to noisy projections, all of which have a noise power spectral density (NPSD), Sproj(f), then the resulting computed tomographic (CT) reconstruction has a two dimensional NPSD of the form, S(f) ~ |f||h(f)|2 Sproj(f). For proper reconstruction, h(f) must approach a non-zero constant as f 0. Provided Sproj(f) is constant, i.e. white projection noise, the CT noise at low frequencies is supprbssbd by the |f| factor. This low frequency suppression results in a long range negative spatial correlation of the CT noise. If white noise is spatially averaged over a circle of diameter d, then the variance in the averaged values will behave as a2 ~ d-2. For CT noise the variance drops faster than d-2. Simple signal-to-noise ratio considerations suggest that the dependence of minimum detectable contrast upon the diameter of the circle to be detected could be significantly differ-ent in the presence of CT noise than in that of white noise. Simulated reconstructions of a suitable detectability pattern demonstrate these differences may not exist unless the image is spatially smoothed before observation. It is pointed out that the pixel width used in the image display should be from 1/3 to 1/2 the width of the point spread function in order to avoid discrete binning problems.

In the field of x-ray reconstruction tomography, many sets of attenuation data are gathered at various angular orientations to the object. Each set of data is divided into many individual elements where each element represents a single pencil beam measurement of attenuated x-rays at some perpendicular distance from the origin. Given this scheme of data collection, it is reasonable to ask the following question: For a given reconstruction diameter, is it more important to collect more sets of projection data - views, or more elements per projection - rays? The answer to this question is basic to the design of computer assisted tomographic (CAT) scanners. Using simulated projection data, we show that for the case of Rotate/Translate geometry, the reconstructed picture is much more sensitive to rays than to views. That is, given the design choice between increasing the number of views or increasing the number of rays per view, one would clearly prefer to sample rays more frequently. This concept may be generalized to any geometry CAT scanner being particularly relevant in the design of pure rotational fan beam CAT scanners.

A method for analyzing dual-energy computed tomograms is presented. The method requires three calibration values at each energy: air, water, and a KI solution. The first two values are used to convert the CT numbers to the ideal Hounsfield scale. The KI value is used to calculate quality factor Q that describes the effective x-ray spectrum. By using the Q-values, the Compton and photoelectric coefficients of any tissue or material may be easily obtained. One can further calculate the electron density and effective atomic number, if desired, and perform chemical identification.

In our institution performance evaluation checks of diagnostic ultrasound imaging equipment have in large part been carried out by medical physics staff and students. Routine tests primarily include checks for changes in machine sensitivity, depth calibration, horizontal and vertical display agreement and B-scan registration accuracy. More recently a triggered pulse burst generator has been used for evaluating gray-scale characteristics. The routine tests are now being handed over to the chief ultrasound technologist who will work under supervision of medical physics personnel.

The intent of this session is to present an open, frank and informal discussion of the advantages, disadvantages and impact of the several state and federal regulatory agencies regulations on medical x-ray equipment on the users of such equipment. Speakers were invited from both camps as well as from the equipment manufacturing industry. The topic is to be considered broadly so that "users" includes owners (e.g. hospitals), clinicians and even the public itself.

It's been some 80 years since Professor Roentgen admitted his ignorance by referring to his discovery of the "X" ray. Although countless electrons have struck their anodes in the ensuing decades, we still honor that admission. Societies, associations, clubs, journals, and individuals may try their best to refer to Roentgen rays, but daily jargon, as well as legal codes, refer to diagnostic and therapeutic x-rays.

There is no question that the government at all levels has the responsibility of protecting the public from hazards, both proven and implied. The impression which is often given is that government regulation is required in the field of radiology because the radiological community is not capable of, or inclined to assure, proper radiation protection of the patient and other individuals. This mistaken impression which was rampant in 1967, and which led to the passage of the Radiation for Health and Safety Law of 1968 (PL90-602) was in part due to the fact that most radiologists and physicists have been unwilling to lower patient dose at any cost - rather they have insisted that benefits not be reduced and that the risk-benefit ratio not be increased as a result of a decrease in dose. This, of course, should not be construed as a resistance to a decrease in dose, but rather as an effort to be certain that the proven values of x-ray usage not be compromised. It should be noted that since 1967, various government regulatory agencies have more realistically moved in the direction of evaluating both the risk and the benefit sides of the risk-benefit equation rather than the somewhat unilateral evaluation of only the risk which was prevalent in the '60s.

During the past decade there has been intense pressure to provide increasing regulations involving the practice of Medicine. As part of this process we have encountered increased regulations with implications that many more are being prepared with regards to the users of x-ray equipment. Some of the pressure for this type of interface in the regulation of x-ray equipment comes from consumer groups who have now become active participants in the planning process for health care. Other pressures come from federal, state, and local regulatory agencies who feel that this opportunity now permits them through the regulatory process to ultimately limit the costs of health care. Certainly this has been the thrust of local and regional planning agencies and the Certificate of Need legislation which has proliferated throughout our country. Although in most of these instances the concern is expressed that this will lead to improved quality of patient care it is my impression at this date that this increased use of a variety of regulatory agencies has not obtained this objective but instead has increased the bureacracy of health care planners and at the same time has increased the cost of health care equipment particularly in the area of diagnostic radiology.

High kilovoltage radiography had its beginning with supervoltage technique and is presently accomplished with the 350 kV Chest X-Ray System. The physical reasons to explain the improved visibility with this technique are: (a) absorption coefficients of bone and soft tissue, (b) more uniform bone visibility, (c) x-ray spectrum, (d) visual responses, and (e) depth resolution. In our study, 350 kV and 120 kV chest x-ray techniques are compared by evaluating some 15 parameters. Based on our data and the results of previous large clinical studies, a list of disadvantages and advantages if offered. The entrance dose for an average 350 kV PA chest radiograph is 8 mR, which is about 1/3 that with 90 kV technique.

Obstructive lesions in the coronary arterial tree may impair capillary or myocardial contrast accumulation, commonly known as myocardial blush. Densitometric assessment of the blush in cine-fluorography presupposes that the part of the cine-frame that represents the myocardium (or well defined part of it) can be retrieved on sequential frames. In the present operator-interactive approach, the blush is determined by digital subtraction. The procedure is the following: (1) A mask is selected by the operator before injection of the contrast medium at a selected instant within the heart cycle. (2) On one frame from the post-injection film series taken at the same phase of the heart cycle as the mask the operator may identify as many as 3 analyzing windows (polygons) which cover various areas on the myocardium. The computer then finds the same areas represented by the windows on each frame, computes the integral of the optical density distribution over them, and subtracts these values from the respective integrals on the mask. A plot of the subtracted data as a function of time gives the myocardial contrast turnover at selected sites.

New comprehensive surgical procedures for the correction of severe craniofacial anomalies have resulted in a need for precise anatomical information for presurgical planning. The extreme variability in expressivity of these anomalies has served to make this need particularly urgent. This, in turn, has led to requests for increasing numbers of torrographic scans for assessment of these patients. Both standard and computed tomograms are used for this purpose. Unfortunately, there is an extreme paucity of information on comparative radiation dose levels and comparative image quality among these methods and their use may, in some instances, place patients at undue risk. For purposes of the current study, a phantom head was constructed which allowed the measurement of the Threshold Quality Factor (TQP) as a function of radiation dose. Quantitative results obtained from this phantom and from cadaver heads on three CT units (CT/T GE, Delta 50, and EMI 5005) and from one standard tomographic unit (CM Stratomatic) were compared in terms of dosimetry, quantified image quality, and subjective clarity of anatomical landmarks. Based on findings in this study, it is suggested that a diagnostic hierarchy could be formulated which would significantly reduce the total radiation levels to which specific types of patients are exposed.