Scanning equalization radiography (SER) is an improved method of chest imaging. The reason for this improvement is due to the fact that SER optimized film exposure producing more subject contrast and better visualization of anatomic detail. This new system has produced improved diagnostic accuracy on a wide variety of cases by all the tested radiologists. Also, there were fewer false positive errors by each of the radiologists. We found that this improvement in image quality and decrease in diagnostic errors was not related to the image complexity (# of abnormalities/radiograph). Also 11 of the radiologists, regardless of their diagnostic skills had a marked preference for the SER images over the conventional.

The field of medical imaging has experienced many significant advances in recent years with the evolution of a host of computer assisted imaging methods. This growth has also been evident in the areas of more conventional radiography through improved resolution and sensitivity in screen/film technologies. However, in spite of these improvements the fundamental principles of radiographic projection imaging have not significantly changed since its earliest demonstration. A case in point is the nature of the irradiation technique itself which routinely uses a field. of radiation of spatially uniform intensity. These uniform fields can result in large variations in transmitted exposure when used in radio graphy of the chest, head and neck. These wide exposure variations often exceed the useful exposure range of conventional radiographic film/screen combinations and result in large portions of the image being rendered with suboptimal contrast. In chest radiography this is particularly evident, resulting in images where the thick mediastinal, diaphragmatic and heart regions are rendered with negligible contrast when the thinner lung zones are properly. exposed.

The use of a trough compensation filter to reduce the large difference in x-ray transmission between the lung parenchyma and mediastinum results in beam hardening and increased scattered radiation which causes a reduction of subject contrast in the lung fields. When the Rotating Aperture Wheel (RAW) multiple-scanning-beam device is used with the filter, this image degradation in the lung fields can be prevented while obtaining improved visualization of the mediastinum. Quantitative comparisons of image contrast and scatter distributions over chest radiographs obtained using various com-binations of anti-scatter device and filter are examined.

Recently, anatomically shaped lead acrylic filters have been introduced for chest radiography. Two of these filters were compared to several (Al, Cu, Y, and Pb acrylic) uniform filters. Phototimed exposures at 80, 100, 120 and 140 kVp were made on a realistic chest phantom. The optical density in the lung field was kept constant for all filters and kVp's. Exposure time and entrance exposures to the mediastinum and lungs were measured. When compared to standard (3.2 mm Al HVL) aluminum filtration a reduction of (44% - 60%) of lung exposure and better visualization of the mediastinum and retrocardiac areas were noted. However, a significant (80 - 350%) increase in exposure time was required and mediastinum exposure increased by 40 - 100 percent. When using uniform filters, in addition to the standard aluminum filter, entrance exposures to the lung and mediastinum were reduced by 30 -50% for Yttrium and Copper and by 27 to 35% for lead acrylic. Exposure times increased by up to 36%, 64%, and 52% respectively. When using spatially shaped filters, improved image quality and reduced lung exposure results, however, one must be aware of the significant increase in exposure time especially at low (80) kVp's. There is also an increase of medias-tinum exposure and possibility of positioning artifacts.

The performance of a new high line rate (1023) video imaging system (VHR) installed in the cardiac catherization laboratory at the University of Colorado Health Sciences Center is compared to the previously installed standard line rate (525) video imaging system (pre-VHR). Comparative performance was assessed both quantitatively using a standardized evaluation protocol and qualitatively based on analysis of data collected during the observation of clinical procedures for which the cardiologists were asked to rank the quality of the fluoroscopic image. The results of this comparative study are presented and suggest that the performance of the high line rate system is significantly improved over the standard line rate system.

Nuclear magnetic resonance imaging (MRI) imaging is a new medical imaging technology with a great potential. A growing number of systems are available for clinical uses. While the quality of images from MRI scanners have been improving, methods by which the NMR imaging performance can be measured have been scarce in the literature. This paper describes a compact MRI phantom and the method that can be used to monitor the performance of MRI systems. The phantom was evaluated as we were completing the task of building the first high field 1.5 T whole body MRI system at Columbia Presbyterian Hospital during 1983 and 1984. During the period we felt a need for a phantom that is compact and all inclusive for quick evaluation of pertinent imaging parameters without the use of any time consuming image analysis routines. Particular attention had to be paid to study the multislice 2 dimensional and 3 dimensional imaging capability of MRI that are used in our lab.

While proton imaging has become a routine diagnostic modality, sodium imaging has been difficult to obtain because of its inherent weakness in signal strength due to significantly lower concentration in tissue. -This difficulty of poor signal to noise ratio can be over-come in several ways, by using 3 dimensional data acquisition rather than 2 dimensional, a better low-noise electronics and fast pulse repetition rate for more signal averaging and multi-echo acquisition. The biochemical properties of sodium, yielding a dramatic contrast changes associated with pathology, also lead to facilitate favorable conditions for clinical imaging compared with proton imaging. The combined use of the advanced imaging techniques and the biochemical properties of sodium can produce clinically useful images. We have demonstrated an additional use of sodium imaging by obtaining simultaneous images of proton and sodium for direct comparison of the two distinctly different but related modalities.

Using analytic and experimental methods, we have investigated the design, fabrication, and performance of axially symmetric shields which minimize the field perturbation within the central bore of an MR magnet. Magnetic resonance shields have been built for 1.9 tesla/60 cm bore and 1.5 tesla/100 cm bore superconducting magnets. The 1.9 T system is enclosed by a simple cylindrical shield having a weight of 49,000 pounds. The 1.5 T system shield is a 100,000 pound tapered thickness cylinder with conic ends. The shields are constructed from cold rolled steel sections which are welded in place at the magnet site. A low carbon, specially annealed steel with particularly desireable properties of induced magnetism has been employed. The 5 gauss line is constrained to a point 4.0 meters to the side of the 1.9 T magnet, 3.0 meters to the side of the 1.5 T magnet and 7.0 meters off the end of both magnets.

A new incoherent method for 3-D object repositioning is proposed, which was verified by experiments. This technique allows overall repositioning instead of individual points of the object, as usually. When a good repositioning is achieved, or if the object remains in the same place, deformations or movements will be observed at real-time. The method consists of two steps: Recording of a grating projected onto a 3-D object, and re-projection of the developed record onto the repositioned object. A photographic enlarger is used in both steps. The mismatching with the original position is observed as deviations of the grid fringes with respect to the original pattern. The sensibility may be governed by varing the grid frequency. This method does not require more stability conditions than the usual moire techniques, and its operating range is from a few millimeters to several meters. Several applications are suggested.

We have developed the DICOM.-8 digital imaging computer for video image acquisition, processing and display. It is a low-cost mobile systems based on a Z80 microcomputer which controls access to two 512 x 512 x 8-bit image planes through a real-time video arithmetic unit. Image presentation capabilities include orthographic images, isometric plots with hidden-line suppression, real-time mask subtraction, binocular red/green stereo, and volumetric imaging with both geometrical and density windows under operator interactive control. Examples are shown for multiplane series of CT images.

Measurements of the performance of two high-resolution radiographic screen-film systems at 50 and 30 kVp are given in terms of the noise equivalent number of recorded x-ray quanta (NEQ) and the detective quantum efficiency (DQE). Additionally, the MTF, contrast transfer function (CTF), and noise power spectra (NPS) of these imaging systems are presented. The NEQ, DQE, CTF, and NPS are shown to be strong functions of both spatial frequency and exposure. The shapes of these surfaces are significantly different from those published for general-purpose radiographic systems; these data provide a useful context within which screen-film system performance can be interpreted.

Fundamental properties of the family of rare earth tantalate phosphors are reviewed, including crystal structure, morphology, density and luminescence. Ultraviolet emitting YTa0A becomes a good choice for commercial applications, based on low print-through with double emulsion films, high DQE and high sharpness to speed ratio. Speed, DQE and MTF data for Quanta Detail and Quanta Fast Detail screens made from YTaO4 are presented.

For many years medical x-ray processing has been dominated by roller transport machines with automatic developer replenishment. These processors, with high film volumes, are centrally located in radiology departments. The high film volume is an important aspect in the formulation of the traditional developers that have been used in these processors. Developer lifetime is long, 2-3 months, and most often the developer is recharged, not because of sensitometric performance but rather for processor cleaning and maintenance.

In order to determine the relative effectiveness of improperly stored, preexposed sensitometric film in the evaluation of processing during a several month period of time, a comparison was made of medical x-ray film which was exposed with a light sensitometer and immediately processed (freshly exposed) and film that had been exposed with a light sensitometer at an earlier time (preexposed) and processed along with the freshly exposed film. The freshly exposed and preexposed films were further divided into a properly stored group, in our darkroom, and an environmentally stored group. The environmentally stored films were kept in a light-tight box in the trunk of a car during the months of August through October, 1983, and August through September, 1984. Each sample set of data included (1) properly stored and freshly exposed film, (2) properly stored and preexposed film, (3) environmentally stored and freshly exposed film; and (4) environmentally stored and preexposed film. A complete set of films was processed at specific times during the study. Initially the time interval for processing between two consecutive sample sets was measured in minutes, with this time interval increasing to hours, days, and weeks. The last sample set was processed 70 days after the study began. Density differences as large as 0.70 between the freshly exposed and preexposed films were observed. A second trial was conducted in order to repeat some of our earlier observations. Two additional film types were incorporated into the second trial. Observations will be discussed along with implications for quality assurance test procedures.

Port films are routinely exposed during radiotherapy treatment to verify correct patient positioning. Contrast is always poor, due to the high energy of the radiation, and spatial resolution is seriously degraded by scatter and large penumbra. We place port films on an illuminated panel facing a vidicon camera, and obtain digitized images for processing in our DICOM-8 digital imaging computer. Averaging up to 256 frames reduces camera noise. Convolution filtering, windowing, histogram equalization and blurred mask subtraction lead to significant improvement in the perception of anatomical structures. Digitization of the x-ray simulator film, using real-time subtraction for accurate alignment, permits the direct comparison of the actual field size with the prescribed treatment area.

Interest in diaphanography for the early diagnosis of breast cancer is increasing, and a number of clinical trials are being conducted. However, little is known about the transmission of light through tissue, or the mechanism by which shadows of lesions are formed. In this work we are studying the scattering and absorption of infra-red light by homogeneous suspensions of red blood, sarcoma and leukemia cells, as well as both normal and pathological samples of human breast tissue.

The performance characteristics of a new dedicated mammography unit (a Mammo Diagnost U-M, Philips Medical Systems, Inc., Shelton, Conn.) are presented. The unit has capabilities for contact imaging with or without a grid and magnification imaging without a grid, using a molybdenum target and either of two different beam filtrations (molybdenum or aluminum). Characteristics evaluated include the performance of the beam restriction system, accuracy of technique parameter selector indication of kilovoltage, tube current and exposure time, focal spot size, exposure output and beam quality characteristics as a function of kilovoltage and filtration, compression device attenuation and automatic exposure control system performance. The resultant data have been used to calculate breast entrance exposure and average glandular dose for each imaging modality and to compare the different imaging modalities in terms of equivalent bandwidth, fe, using established models for the modulation transfer functions of the focal spot and image receptor. Modifications to the system to enhance clinical utility and to allow for routine calculations of patient exposure are also described.

Laboratory studies and results of our work to reduce the x-ray exposure of xeromammographic examinations are described. Significant exposure reduction was achieved and demonstrated with breast phantom images through improved photoreceptor sensitivity and a higher sensitivity electrostatic image development process. The contributions of x-ray absorption and discharge efficiency of an experimental selenium photoreceptor are discussed. Development of the latent charge image on the photoreceptor with an electrophoretic process is explained, and the impact of key variables such as covering power and other toner properties on this process are analyzed. Phantom images of the present commercial xeroradiographic system are compared to those made with the experimental higher sensitivity process and are discussed in light of image quality criteria.

A performance comparison between two digital imaging systems was done using various phantoms to both simulate a clinical situation and provide insight into each system's imaging characteristics. One system analyzed was a complete Philips package including a DVI-II digital imaging system with the most recent software updates and a fourteen inch "C" arm mounted multi-mode image intensifier. As the room contains only one manufacturer's equipment, it represents the current approach to digital imaging. The second system was an ADAC DPS-4100 digital imaging system added on to CGR Septar components, including a nine inch dual mode image intensifier. The add-on approach has recently fallen on disfavor. The comparison was undertaken to determine the relative advantages and disadvantages for each approach. Image characteristic comparison was performed using a commercially available digital subtraction angiography phantom. Measurements of each system's high contrast spatial resolution, low contrast detectability and linearity, dynamic range and subtraction effectiveness were made. Clinical demands were simulated with a standard skull phantom. This enabled evaluation of each system's image acquisition requirements, post-processing capabilities and patient dose.

Purchase of a digital subtraction angiography (DSA) system involves medical, financial, and technical considerations. The Lahey Clinic DSA phantoms were developed to assist in technical evaluation of DSA equipment by providing a noninvasive means of assessment of radiographic and digital equipment components before purchase. The same set of phantoms can also be used to assess performance of a. system at the time of installation and for routine evaluation. The critical components of a DSA system are detailed with special emphasis on the image intensifier and television camera. Individual phantom patterns are described, and images are presented.

Based on an acceptance testing procedure by means of the American Association of Physicists ini Medicine (AAPM) digital subtraction angiography (DSA) phantom described previously , experiences were gained with respect to the expected level of system performance and the image quality reproducible for typical DSA systems currently installed in the clinical environment. There were more than 20 DSA systems involved in this study and encompassed several DSA manufacturers' products. A DSA system is employed as an example to highlight the importance of the initial calibration of the DSA imaging chain, and to demonstrate the differences in the image quality obtainable for the same system.

Using standard angiographic equipment involving an image intensifier-video camera combination, subpixel precision can be achieved in the estimation of blood vessel diameter using 512x512x8 digital image resolution. Analyses were carried out with blood vessel phantoms ranging from 1.27 to 5.46 mm in diameter. The relationship between actual and measured diameters was found linear between 3 and 5.46 mm; below 3mm it became non-linear. The studies were performed under ideal conditions with a water block as scattering medium, without motion and without subtraction on uniform background.

The rapid scanning beam Rotating Aperture Wheel (RAW) device is capable of achieving about 1,000 beam passes per second over each point in the field of view, allowing millisecond exposure times. Screen-film images of a chest phantom were obtained with exposure times of tens of milliseconds at scan speeds of a few hundred beam passes per second. These are the shortest exposure times ever achieved by a nongrid, purely scanning beam anti-scatter device. Because the scanning beam speed and exposure times were not synchronized for this study, well understood overlap artifacts appeared on the original radiograph. However, this artifactual density variation was subtracted, and the noise variation caused by the different quantum fluence in the alternating regions was imperceptible. A number of artifact correction methods are discussed. Finally, the application of various RAW device configurations to the millisecond exposure times used for cardiac cineradiography are outlined.

The personnel radiation exposure which is possible in interventional radiology is much greater than that in other areas of diagnostic radiology. Following a review of recent film badge records for personnel involved in interventional radiology at our institution, we have examined various methods to reduce head and neck exposures during interventional procedures. Calculations and measurements demonstrate that with even a modest workload ancillary shielding materials are necessary in order to maintain head exposures within regulatory limits. This report presents a newly developed shielding method (surface shield) for use during interventional radiology procedures. The surface shield is inexpensive and reduces head and neck exposure by up to 75% without compromising patient access or radiologist convenience. Due to the special demands of interventional radiology, periodic review of procedures and radiation protection principles with the personnel involved is important in keeping personnel exposure as low as reasonably achievable.

Single kVp and dual energy CT imaging were utilized for the assessment of bone mineral content. In this approach, the CT values obtained from the trabecular portion of the lumbar vertebrae were related to known mineral concentrations in compartments of a calibration phantom that is scanned simultaneously with the patients. Both pre-reconstruction and post reconstruction dual energy CT methods were utilized in this study. The results obtained from single kVp scans were compared with the data from dual kVp scans. The influence of marrow fat content, variations in the calibration phantom, beam hardening effects and scanner drifts upon the accuracy and precision of the bone mineral measurements were investigated.

The measurement of spinal bone mineral content has been achieved by using automated contour detection on CT images. The method reproducibly provides a definition of the cortical area and calibrates a region of interest over essentially the whole spongeous bone. The algorithm can be used in single and dual energy CT studies. In single energy examinations the analyses employ a correction procedure based upon the method introduced by Cann and Genant. The procedures were tested on 40 randomly selected vertebral images.

The linear dependence of Computed Tomography (CT) numbers on linear attenuation coefficients was investigated. CT numbers were fit to linear attenuation coefficients using the least squares method for energies from 55 to 80 keV. The variance was recorded. The energy representing the lowest variance was assumed to be the effective energy of the scanner. Six points were insufficient to unambiguously define the effective energy. Nine materials were used in all subsequent studies; unambiguous values were obtained. Linearity was determined as a function of field of view, phantom size from 10 cm to 30 cm diameter, and energy. Two plastics (polysulfone and acetal) are described which extend the linearity curve to 215 and 360 Houndsfield numbers respectively (scale is 1000). Practical applications were investigated.

This study was undertaken to assess the effect of reconstruction time upon computed tomography (CT) patient throughput. Time studies were performed to analyze patient throughput and the distribution of examination time on a Pfizer 0450 and a Siemens DR3 CT scanners. Analyses were performed for both head and body examinations. An overall reduction in examination time of about 32% was observed for the DR3, owing primarily to the nearly instant image reconstruction available on this scanner. Our initial time study on the 0450 scanner proved to be useful in predicting the increase in patient throughput on the DR3. Hence the time study approach was felt to be valuable in both its predictive use as well as a tool to analyze current operating characteristics.

Diagnostic-quality X-ray computed to ography (CT) images, with spatial resolution of 0.05 to 0.1 mm, have now been made of small laboratory animals with a "micro-CT" scanner incorpo rating a special, high-resolution X-ray detector array. The images show clearly defined internal organ structure and tumors in intact rodents.

During the preparation and planning phase of the PACS project at Georgetown University Hospital it was realized that PACS requires truly the state of the art technology in data communication, image processing and man machine interfacing. It was also realized that un-like many other technology intensive devices used in radiology, PACS cannot be seen as an independent system that will provide well defined services. PACS will be the backbone of the department operation in clinical, educational and managerial functions. It will indeed be the nerve center of the radiologic services affecting every aspect of the department. PACS will have to be designed to perform in a cost-effective manner to widely varying needs within the radiology departments. The integration of ever changing complex technology that will impact every aspect of a radiology service is not a trivial matter. This transition period going from current manual film based PACS to Digital PACS can be long, expansive and disruptive unless careful planning preceeds the implementation. PACS is still an emerging technology at its infancy. Performance monitoring and evaluation of diversified functions have to be also established so that improvement to the system can be efficiently implemented. Thus the evaluation criteria should be also established as early as possible.

A survey was conducted in which the medical community at a university hospital was asked to respond to questions concerning Picture Archiving and Communications Systems (PACS). 20% of the physicians (medical doctors and senior medical students) responded. Results show that physicians perceive a need for PACS. Although all physicians saw advantages in PACS, radiologists, in particular, saw the system as beneficial to their practice. In contrast, non-radiologists saw more drawbacks to the system most notably decreased portability and excessive cost of the system. Most physicians thought they would utilize radiologic information from remote sites in such a system. They also suggested several ways in which the system might be uniquely modified for their subspecialty needs. The implications of PACS for the practice of radiology will be discussed.

The radiation dose, image signal-to-noise ratio, image resolution and cost of film/screen and digital subtraction angiography have been investigated. Theoretical analysis indicates that the dose normalized signal-to-noise ratio for DSA is 50% better than film/screen methods. Iodinated vessel phantoms have confirmed this result. An evaluation of 55 patients indicates improved small vessel detection, better delineation of diffuse contrast stains, and fewer artifacts with DSA when compared with film/screen results on the same patients. Substantial reduction in cost for DSA methods is attributed to film utilization reduction as well as reduced examining times.

A potential advantage of digital radiographic systems is their ability to enhance images by various types of processing. Digital unsharp masking is one of the simplest and potentially most useful forms of enhancement. The efficacy of unsharp masking in clinical radiologic diagnosis has not been investigated systematically, however. The effect of digital unsharp masking on the detectability of two types of subtle abnormalities, pneumothorax and interstitial infiltrate, was studied in an observer performance test. An ROC analysis of this preliminary data suggests that unsharp masking may improve diagnostic accuracy for pneumothorax. Radiologists' performance in identifying interstitial infiltrates was degraded by the image processing, however, and false positive diagnoses tended to be more frequent.

A computer program Automated Nodule Detection System Version-1 (ANDS-V1) that locates candidate nodules (CNs) in a chest radiograph has been developed. The CN locations are displayed to a radiologist for further inspection. This program was tested on 37 chest radiographs that contain at least one nodule. At least one nodule in 92% (34/37) of the radiographs was detected by the circle detection program (untrained ANDS). Among the CNs detected in a chest radiograph are numerous false positives. The radiologist must inspect 45 sites on the film to be certain that all detected nodules had been inspected. The number of false positive sites presented for inspection is reduced by the Nodule Expert program. Only 10 CNs must be inspected for the radiologist to be confident of having inspected all nodules detected by the program. The true positive rate is 86% (32/37) when the Nodule Expert program is used. Features for discriminating among various classes of CNs are described.

We have used a prototype digital beam attenuator (DBA) system to generate patient-specific digitally-prepared compensating filters for chest radiography of a human subject. The compensated radiographs demonstrate substantially more information in areas such as the mediastinum and upper abdomen which normally are underpenetrated in conventional chest radiographs. The compensated image was acquired with high contrast, high speed film-screen receptors improving the visibility of pulmonary parenchymal detail while minimizing patient radiation exposure. Currently we are limited by a two-hour preparation time and position the attenuator manually. We are developing a second generation DBA system featuring fast (15 second) fabrication times and automatic positioning of the attenuator. We expect that these features will relieve some of the misregistration errors present in our initial examination.

Film-screen mammography is the most sensitive means available for the detection of early breast cancer. Its performance, however, is often limited by the restricted latitude over which the gradient of the radiographic film Provides appreciable contrast and by film granularity. We are currently investigating means for acquiring the mammographic image directly in digital form, using a scanned fan-beam x-ray system and high resolution image intensifier coupled to a self-scanned photo-diode array. Such a system will potentially provide efficient scatter-rejection and images which are limited only by quantum fluctuation. At the same time it would make practical such image enhancement techniques as contrast amplification and unsharp masking as well as dual-energy image processing for the improved detection of calcifications. In this paper a prototype digital mammographic system is described and preliminary images of surgical specimens are presented.