A coded aperture consisting of a planar array of three off-axis zone plates oriented at 120° with respect to each other has been used to image gamma-ray sources up to ten inches in diameter. The zone plates do not have a common axis, and the high frequency zones of each zone plate are located in the central region. The maximum radius and fringe spacing are chosen so that the spatial frequency bandpasses for the individual zone plates do not overlap, and a hexagonal array of holes is used as a sampling screen (half-tone) to shift the object spectrum into the bandpass of the three zone plates. The aperture configuration permits more uniform sampling of a large object as compared with a system consisting of a single large off-axis zone plate. Design parameters are discussed, particularly those related to the imaging of bar phantoms and the generation of coherent artifacts. Reconstructed images of gamma-ray sources using the three zone plate system are presented.

We have previously described the con-struction and operation of a xenon-filled multiwire proportional chamber (MWPC)(Ref.1, 2). This type of detector offers high resolution gamma-ray imaging over large areas, with good uniformity of response and at costs that are low when compared to that of scintillation cameras. The main drawback of MWPC's is their low detection efficiency. This parameter can be increased by constructing thicker detectors, but at a sacrifice in spatial resolution, which is degraded by the increased transverse path presented to photons accepted within the solid angle subtended by the collimators used for imaging. On the other hand, increasing the density of the gas raises detector efficiency by increasing its effective mass. Because the detection process in a MWPC involves transfer of energy from a gamma-ray to an electron, the path length of this conversion electron determines the intrinsic spatial resolution of the detector. Thus, an increase in gas density also has the favorable effect of reducing this path. We have found that operation in a pressurized mode (Ref. 3) offers the advantage of simultaneously increasing efficiency and spatial resolution.

Tracer doses of y-emitting, radio-labeled pharmaceuticals have been used for a number of years for tumor localization and to provide the clinician with regional information concerning organ function and blood flow. An image of the radio-pharmaceutical distribution is formed by means of the emitted y-rays. Elevated or depressed concentration of the tracer agent as reflected by hot or cold spots in the image is indicative of abnormalities. The y-rays, which must be of sufficient energy to penetrate tissue with minimum scattering and absorption, cannot be refracted or reflected so the image must be formed by aperture limitation; (i.e. a pinhole in lead or multi-channel collimators). Such an aperture is extremely inefficient with solid angle efficiencies on the order of 10^-4; efficiency may be increased but only at the sacrifice of resolution or field of view. This, coupled with the need to minimize the radiation dose to the patient and to maintain reasonably short imaging times, results in low resolution images corrupted by noise arising from the statistical fluctuation of the y-photon emission. A typical 25 cm diameter image field seldom contains more than 1000 resolved image elements and the information is often carried by fewer than 100,000 photons.

A method for producing quantitatively accurate, three-dimensional images of the structures of the body, or in the case of nuclear medicine, the distribution of radio-activity within these structures has been a tantalizing but elusive goal of medical research for many years. Among the many approaches that have been tried, those coming closest to success have been the various forms of tomographic or section imaging. The form of tomography known as transverse axial or transverse section tomography is particularly attractive in terms of quantitative accuracy because information from parts of the object outside the plane of interest does not appear in the final image as it does in other forms of tomography.

Scintillation camera data storage, processing and playback capabilities are becoming an important if not essential function of the modern nuclear medicine laboratory, both in routine service studies and in research. However, the high cost of current dedicated computers makes them relatively inaccessible to many laboratories. Polaroid film has long been the primary medium of image acquisition and storage, offering the advantages of an immediately available image. Mostly used with a 3-lens camera, where each tens is set on a different f-stop to compensate for the poor latitude of the film, the final copy consists of three minified images with a crude form of background subtraction at the high f-stops. The usefulness of Polaroid film used this way was greatest when spatial resolution was relatively poor, and data density was limited by the radiation dose delivered by the available radioisotopes.

A mini-computer system has been implemented to perform interactive image processing. The computer system presents an in-tegrated modular approach to hardware development as well as software development. Digitally controlled scanners are interfaced with the system and are treated as extensions of core memory. This technique alleviates a mmajor constraint on the total amount of core memory that may be required for image processing. Scanning and processing may be accomplished simultaneously on a line-by-line basis. An image dissector scanner is the primary digitizer at this time. Several displays are interfaced to the mini-computer for on-line image output. Program, or algorithm, analysis is provided by image displays at various stages of the processing. A high contrast interactive Digital Disc Display serves as the primary image out-put at this time. Applications of the system include on-line analysis of chest radiographs for the detection of heart disease, feasible detection of increased or decreased vascularity in the lung fields, and automatic area measurements on cells from pathological slides.

Xeroradiography has been felt by many radiologists to be diagnostically superior to no-screen x-ray films for mammography. Recently we have shown that Xeroradiography has a modulation transfer function similar to that of "soft focus masking" or "dodging" wherein low spatial frequencies are reduced almost to zero while frequencies near one line pair per millimeter are greatly accentuated. Since lower spatial frequencies contain information mostly about the size and shape of the breast while the higher ones presumably contain diagnostic information about small calcifications and blood vessel structure, it would appear that Xeroradio-graphy is able to present the spatial frequencies of interest with great contrast without exceeding the dynamic range of the powdered image in large sections.

The paper gives some preliminary results of a continuing experimental study of adjacency effect. One of the important variables to consider in evaluating photo-interpretation by human observers is the adjacency effect, such as that resulting from the sharp contrast of black and white areas on a photographic surface. When human observers view a photographic surface on which black and white are adjoining, the adjacency or "edge" effect occurs. They notice a band of greater brightness just before the light area apparently shades into gray and a band of greater darkness just before the apparently gray area shades into greater darkness. Examination by a photometer or other optical instrument shows that these bands of light and dark do not exist and are only apparently present in the pattern. The apparent brighter and darker bands occur at the point of what was first known as the Mach Band effect (1916). The paper is concerned with a quantitative evaluation of the Mach Band effect when brief exposure time, less than .5 seconds is used.

Degradation of the radiologic image is caused by the finite size of X-ray sources. This effect is often the principal limitation on the resolution of a radiologic imaging system. If it were possible to reduce the source size, the contrast and resolution of an X-ray image would be improved. Unfortunately, X-ray tube construction considerations and the need for sufficient radiation output necessitate relatively large X-ray sources.

The mathematical function describing the transmission of x-rays through an object depends on many geometrical and physical vari-ables. A Taylor series expansion of the transmission function about an arbitrary point in the multivariable space displays the variety of terms which may be emphasized or isolated in order to accentuate certain information concerning the object under study. Conventional radiographic or fluoroscopic images are represented by the zero order term. This term can be used to derive higher order terms which depend explicitly on the spatial coordinates x and y, but contains only averaged or incomplete information about variables such as z (depth)E (energy), or t (time). One class of image enhancement procedures involves accentuation of whatever information is present on conventional radiographs by selective filtration of spatial frequency components or isolation of the x and y derivative terms. A second class of images with significantly different information content can be obtained by a priori use of knowledge about other variables in order to isolate other derivative terms. These images can be formed from linear combinations of zero order images associated with different values of the variable Ofinterest. Examples of this class of images include tomography, time dependent subtraction, and absorption edge imaging. Because evaluations of "enhanced" images containing one subset of the possible image terms cannot be assumed to pertain to images containing other terms, it is suggested that the term "image enhancement" is too general and must be used with care.

The title for this luncheon address was thought up rather hastily during a telephone conversation between our Executive Director (Joe Yaver) and myself. Joe made it quite clear to me that as Executive Vice-President of the Society of Photo-Optical Instrumentation Engineers it was my unavoidable duty to be here today and present this luncheon address. I am honored to be fulfilling that duty. The title is in many ways superfluous, but it allows me to make a number of remarks that relate to the general subject matter of this program - "Application of Optical Instrumentation in Medicine".

The purpose of the panel is to consider whether and, if so, how image enhancement can be effective in various situations in medical diagnosis. Further, I hope we can consider how one determines if an enhancement technique is effective or effective enough.

For purposes of this discussion, assume that a conventional radiographic image such as film already exists and an optional enhancement step is available as a potential diagnostic aid. To decide if the enhancement is worth the trouble, several factors have to be considered: 1. Diagnostic value of enhancement. 2. Ease and speed of processing. 3. Cost of enhancement.

For the past eight years, our Department of Radiology has used a multiplicity of both image enhancement techniques and display formats for visual presentations and analysis of two dimensional distributions of information. While such methodology and equipment has been available primarily for enhancing scintiscans in nuclear medicine, such has also been used for various research prob lems in diagnostic radiology.

Nuclear Medicine is a new medical specialty dealing with the diagnostic and therapeutic uses of radioactive isotopes. It is a discipline similar to Radiology but this latter specialty deals with the use of X-Rays. X-Rays originate from the electron shells of the atom, while the gamma and beta radiations used in Nuclear Medicine originate in the nucleus. At present, the major emphasis in Nuclear Medicine is imaging the distributions of nuclides within patients - a process often called nuclide scanning. A large number of relatively new radiopharmaceuticals have been developed which localize in specific organs or lesions. For example, one of the latest radiopharmaceuticals to come into use is Glucoheptonate labeled with Technetium-99m. A tiny amount of Glucoheptonate tagged with a relatively large amount (about 10 mCi) of radioisotope is injected intravenously. One hour later a large percentage of the isotope will be deposited in the renal tubular cells. There is much higher concentration in the kidneys than in other organs in the body. An image then of the distribution of the isotope provides a very clear picture of the normal kidney tissue. Tumor tissue or a cyst within the kidney will not concentrate the isotope and would show up as an area of decreased activity in the image. Similar imaging procedures have been developed for the brain, lungs, liver, bone and several other organs and specific lesions.

The first general question upon which I would like to focus the panel's attention is "In which medical applications are enhancement techniques already being used effectively, in which applications might we expect enhancement to contribute useful information, and in which types of applications is enhancement an unlikely candidate for improvement of the use of diagnostic information in radiologic images?"

Introduced about a decade ago, the fiber-optic endoscope has become an important and versatile tool in medicine. It originally was intended to replace the semi-rigid all-lens endoscopes to provide, (a) greater flexibility than was possible with the classical instrument which consists of a train of optical lenses, and (b) to provide "cold-light" illumination at the distal end of the probe. It thus became possible to develop a whole family of endoscopes for remote viewing and photography in bronchoscopy, gastroscopy, rectoscopy, proctosigmoidoscopy, cystoscopy, etc. The fiber-optic endoscope proved to be of smaller diameter, more flexible and provided more light for improved image clarity in visualizing color gradations in the tissue, an important factor in medical diagnosis. Also, the greater flexibility of the fiber-optic endoscope enables the visualization around corners and the elimination of "blind areas" obtained with the classical semi-rigid instrument. Although small in comparison with the classical endoscope, the fiber-optic instrument is still large enough to accomodate an optical lens system at the distal end. An imaging bundle of coherent fibers used in a typical endoscope has a diameter of from 2 to 5 mm. It usually consists of many fibers (50 thousand to 250 thousand in-dividual fibers) aligned to provide for the coherent transmission of the image. When constructed of individual fibers, lOpim in diameter each, the entire bundle remains flexible allowing for very small bending radii. Commercially available fiber-optic structures of coherently aligned individual fibers utilize fiber bundles with diameters down to 2mm where the ends are tied together while the rest of the fibers along the length of the bundle remain loose and therefore flexible.

In the fall of 1972, the Central Diagnostic Laboratories of Barnes Hospital, St. Louis, Missouri, evaluated several commercial instruments for automation of disc sensitivity test reading and colony counting in the clinical microbiology laboratory. None of the available devices were capable of reading sensitivity test results without assistance from an operator, and the costs of the more nearly automatic ones were so high as to render their cost justification questionable. It was decided that the construction of a fully automatic high speed petri dish reading instru-ment at a cost appreciably lower than existing devices should be possible if a high speed minicomputer were used as a control element and if the price of the optoelectronic components could be kept down by designing to limited performance criteria. Additionally, for reliability and ease of construction, readily available components were to be used throughout.

The Raman spectrum of long chain biological polymers displays a broad band which is distinct from conventional molecular Raman lines. A hypothesis based on a new Electromagnetic Molecular Electronic Resonance has been advanced to explain it. The resolved broad band data of Poly-L-Glutamic acid and Bovine Serum Albumin coincide with the predictions of this theory. The technique has been extended to the analysis of double stranded DNA's. . Spectra obtained from T7 and A viruses show that there is an Electromagnetic Molecular Electronic Resonance taking place along 1 orbital chains between coupled bases. This provides the first experimental proof that there is a IT orbital coupling from base to base in double stranded DNA's. A base dislocation along the DNA is detected in the spectrum. An important feature of this new technique is that it does identify a virus DNA, and measures its length, when the latter is still inside its protein coat.

In recent years the need for large flat screen fluoroscopic imaging systems has been pointed up by the development and subsequent use of new technologies and new procedures in diagnostic radiology. Videodensitometry, if it is to be extended to clinical applications involving studies of the thorax, kidneys, etc., requires an input screen having an area greater than that of the usual x-ray image intensifier (Ref. 1). A similar requirement exists for optimal use to be made of fluoroscopic tomography (Refs. 2, 3, 4). Further, if these very promising developments are to be made truly quantitative rather than relative, the distortion of the resultant optical intensity due to screen curvature needs to be eliminated. The use of digitized systems with on-line processing and with their capabilities for making precision measurements at the levels of minimum detectability, makes this development even more desirable (Refs. 5, 6). Examples of this need can also be found in the qualitative or imaging aspects of diagnostic radiology.

The technique presented is an approach to absorption edge transmission imaging which eliminates the need for monoenergetic x-ray sources, detector scanning, and energy analysis of transmitted x-rays. Subtraction of images formed with heavily filtered x-ray beams peaked at energies above and below the K absorption edge is accomplished with image-intensified fluoroscopy equipment coupled to a pair of storage tubes. Images of iodine and xenon concentrations of 1 mg/cm² have been imaged in 15 cm phantoms in a few seconds. When automated, the system will provide similar sensitivity in 20 cm phantoms, with imaging times of 1 second or less with resolution of the order of 1 mm. Several preliminary images are presented which relate to possible future clinical studies.

Disc recorders and image storage tubes are now being used with fluoroscopic television systems as a means for reducing patient ex-posure to x-rays. Electronic image storage systems also permit a number of methods for image processing and new techniques have been developed for absorption edge imaging, catheter placement, electronic spot-filming and pulsed fluoroscopy. In many cases, the image storage device may be credited with a reduction of patient exposure because the original system was not optimized. While storage devices do permit reductions of patient exposure, first consideration must be given to proper design and operation of the original system.

The most generally used method to determine the MTF of x-ray image intensifiers is through the use of a test object which impresses on the image intensifier a square wave input signal. These devices are usually fabricated from lead sheet by cutting a series of slits of width w separated by an equal width of lead. Different spatial frequencies are obtained by making a series of objects of different slit and separation widths. The output signal resulting from these test objects is the intensifier's response to a square wave. To obtain the MTF, this response is converted mathematically to that which would have resulted from a sinusoidal input signal (1).

We're on the threshold of a genuine departure in radiological practice: The direct spot-film and the full-size radiograph are being rapidly abandoned in a number of diagnostic procedures. Yet an overview of a sophisticated medical imaging system at the end of 1973 looks pretty much like a schematic we might have examined a year ago, or five years ago. Only the details have changed. But just these details - and the performance advances underlying them - are making the intensified spot film the x-ray recording medium of choice. This talk will catalog and describe these advances, several of which have really become commercially available in only the last year or two. Perhaps the most significant unifying "event", though, is the fact that these advanced systems started to move into hospitals in significant quantities during 1973. And this is the ultimate reason we should expect a strong trend away from the direct radiograph.

The experiment now to be described was carried out more than two years ago and an account of it has already been published in a radiological journal(1). It is presented here, then, not because of its technical novelty (though it was, so far as we know, a "first" at the time it was performed) but rather because it complements other methods, reported at this meeting, of transmitting radiological images over long distances.

The results of Nuclear Medicine organ imaging examinations are commonly displayed in pictorial form, either as transparencies, on x-ray film or 35 or 70 mm negatives, or as photographs usually utilizing Polaroid film. The ability to transmit this pictorial information to remote locations for consultation, interpretation and study is extremely useful in clinical diagnostic facilities, educational programs and research. Conventional television installations are prohibitively expensive for most of these applications. Furthermore, most organ images are stationary and do not require "real time" television. Slow scan or compressed video television systems are suitable for the transmission of stationary images over standard telephone circuits. This report relates our experience with two commercially available systems over a period of approximately three years in Nuclear Medicine diagnostic units in a University Health Science Center and a two hundred thirty bed private community hospital.

The use of television to visualize images at a distance from the source of the images offers diagnostic radiology the promise of revolutionizing the manner in which the medical speciality of radiology is practiced.

The Health Care Technology Division, Department of Health, Education & Welfare, is sponsoring two trials of the Bell System's PICTUREPHON service in hospitals serving Chicago's west side ghetto.

Radiographic information, like any other kind of information, has no value until it is presented to a human observer, in this case, the radiologist or referring physician, who will extract from, and apply, this information to benefit the patient. Information, in other words, is like the sound made by a tree falling in the middle of a forest--it doesn't really exist until there is a human receiver. The human observer is, therefore, the focus around which any radiographic information system must be developed.

The need for performance standards in image intensification is related to the concern of the radiologist for obtaining adequate diagnostic information at minimal patient dose. While we assume that patient dose can be adequately quantized for any imaging system under discussion, we do not presume to be able to define what is meant by adequate diagnostic information and it is this area which we would like to discuss.

I have been asked, this afternoon, to preface our discussion by describing the bench or, as I would rather phrase it, production testing of x-ray image intensified by a tube manufacturer and also what field test information tube manufacturers supply to their customers.

The various occupational responsibilities of the persons selected for this panel is an acknowledgement of the fact that when one considers field tests he must also consider the condition of the field in which he finds himself. I find myself enclosed in two fields. One is that in which the intensifier can be set up in a laboratory for sophisticated measurements. The other is that in which the intensifier is a component of an assembled fluoroscopy system and can be removed only at the expense of great irritation on the part of the radiologist. In this latter field, absolute measurements can be carried out only with a low degree of sophistication and a high degree of uncertainty. These measurements should be expected to yield numbers for comparison purposes only. In the first field, some of my remarks will undoubtedly overlap those of the two panel members from industry. I shall try to keep this overlap to a minimum. In the second field, overlap can be expected with the remarks of Dr. Shea on the users point of view. I will also try to keep this overlap to a minimum. Possibly, some diametrically opposed remarks will be made and a real discussion develop within and without the panel.

The theme of our panel discussion today is Instrumentation as applied to x-ray intensifiers.

This PDF contains the transcript of the Discussion Following Panelists' Opening Remarks from Volume 0043.

The preparation of a radiograph with adequate contrast and resolution does not in itself assure the transfer of the infor-mation necessary for a medical diagnosis. The last physically describable event in the complex chain of events, which constitute the radiological examination, is that of the viewing of the finished radiographs by the radiologist. If the viewing condi tions are substandard, there is a possibility of losing the diagnostic information at this late stage in the examination. Although radiographic viewing conditions are critical to the entire examination process, they have not been subject to extensive human factors engineering studies. In this paper we will describe radiographic film viewing at our institution, and we will discuss some of the problems inherent in such a task,

KODAK RETNAR minification products have been developed to minify radiographs onto a film chip carried by an aperture card with mini-mum loss of diagnostic quality. Converting radiographs to aperture cards reduces storage requirements and adds the convenience of a miniature unit record. The products center around a new film developed for minifying radiographs. The film combines extended exposure scale and near unity contrast with a direct positive photographic response. A bright, efficient reader displays the film's broad density range. A camera-processor exposes a radiograph at a 12.5 reduction on the camera card film, processes and dries the film while mounted in the card, and prints eye-readable information on the card. This unit can be operated by clerical personnel in a typical file room or office environment.

As a part of the broad program in biological image processing being pursued at the Jet Propulsion Laboratory, the development of an integrated electron microscope/ Computer system has been undertaken. The ultimate goal of this project, is the recovery of useful information concerning the structure of biological systems.

A precise correlation, of physiological parameters with cineradiogram is of major importance in the evaluation of cardiac function, in patients undergoing cardiac catheterization. The correlation between a separately produced, paper or photographic recording and a cine requires a timing device, and as a procedure is not only time-consuming but at higher filming speeds is burdened with an error.