There has been a tremendous amount of effort put into the design of diagnostic radiology workstations; however, few workstations have been clinically accepted. Among the requirements for a clinically acceptable workstation are good image quality, a well designed user-interface, and access to all relevant diagnostic information. The user-interface design should reflect radiologist's film reading habits and encourage new reading methods that take advantage of the electronic environment. As part of our effort to improve diagnostic workstation design, we surveyed radiologists in the UCLA Department of Radiological Sciences. Sixteen radiologists from the fields of pediatric, genitourinary, thoracic, and neuroradiology participated in the initial survey. We asked their opinions regarding our PACS infrastructure performance and our existing diagnostic workstations. We also asked them to identify certain pathologies that they found to be less evident on workstations as compared to film. We are using this information to determine the current limitations of diagnostic workstations and to develop a user interface design that addresses the clinical requirements of a busy teritiary care medical center the radiologists who use it.

A Global PACS is a medical imaging system which enables the doctors to capture, archive, and retrieve medical images over wide area networks. In previous work, we have developed a distributed software for remote consultation and diagnosis in a Global PACS environment over the Internet or NSFNET. The doctors are able to interactively perform a remote consultation with basic image annotation commands. In this paper, we present a new mechanism for adding voice to this scenario and performing synchronization of voice and image annotation in a remote consultation and diagnosis session.

Today's radiology practice in a large medical center requires information from multimedia databases utilizing text, images, and voice. Establishing access to this information provides convenience, efficiency, and a robust decision support environment for the radiologist. This paper describes these databases and a method to access this information through a local area network at a radiologist's desktop computer.

Recent advances in personal computer technology have made these low-cost platforms quite attractive for the implementation of radiological workstations, particularly in those cases where high throughput is not critical (such as clinical review workstations). In light of this observation, the objectives of this on-going project are threefold: (1) to identify the characteristics and performance specifications that are desirable in a radiological review workstation (within the UCSF PACS architecture), (2) to review current personal computer technology in terms of its ability to support such a workstation, and (3) to design and implement a prototype hardware and software architecture. This paper outlines our progress to data and discusses some of our projections for the near future.

A prototype IS&C (Image Save and Carry) system is under development at the National Cancer Center Hospital in Japan as a project of the Ministry of International Trade and Industry. The purpose of the project is to develop a medical information filing system for personal health care using the IS&C system. This paper describes the outline and present status of the project.

An radiologic reporting system by means of oral presentation in an IS&C (Image Save and Carry) environment has been developed and tested. The rewritable and compact magneto- optical disk (MOD) according to the standard specification of IS&C filing format is employed to store oral diagnostic reports together with read radiologic images. The MOD of 5.25 inches in diameter has 600 MB memory capacity. Advantages of the system are: simultaneous retrieval of oral reports and the images which a radiologist interpreted, and capability of media circulation in addition to the function of filing. Thus the MOD in our IS&C environment has a multimedia function of both off-line communication and filing. When medical images are interpreted and oral presentations are registered, the digitized oral reports and the digitized images are filed to the MOD automatically. Referring physicians can get the diagnostic reports by oral speech and can see the images at the same time in front of personal computer. Furthermore, integration with a voice recognition machine is being tried in our experiment.

Continuous advancement in the field of high speed computer networks and real-time distributed database technology have made new applications in the field of medical imaging possible. The Global PACS prototype is one such application which has been implemented on a national scale via wide area networks. It integrates different system components, such as imaging equipment, viewing workstations, database management and the archive system, fiber optic networks and a high-speed backbone network. Global PACS includes a distributed database archive system spread throughout wide geographical areas and interconnected via the Internet and NSFNET. This paper presents a new approach using a file management system for a Global PACS distributed database management and archive system.

As part of the Picture Archiving and Communication System effort at the University of Florida, two General Electric high speed advantage CT scanners, a General Electric CT independent review console, a Siemens Magnatom MR imager and a General Electric Signa Advantage MR imager have been connected to a common image archive. Clinical images have been archived from these scanners for about 2 years resulting in about .75 Terabytes of image data. The original archive connections for these scanners were proprietary vendor connections which have recently been replaced by a DICOM implementation utilizing the TCP/IP protocol. A description of the old and new systems will be presented as well as a basic description of the DICOM server which was developed. Timing measurements of the performance of both the old and new interfaces were performed and are presented as well.

This paper will present a technical description of the NOVUS Image Archive System and discuss evaluations underway at two of our Beta test sites. Testing has been underway in a university hospital setting and at a clinical off-site imaging center. Results obtained to date regarding efficiencies gained and cost savings realized, and an outline of future expansion plans for the technology will be addressed.

Since radiation oncologists make the treatment plan by prior experience, information about previous cases is helpful in planning the radiation treatment. We have developed an supporting system for the radiation therapy. The case-based reasoning method was implemented in order to search old treatments and images of past cases. This system evaluates similarities between the current case and all stored cases (case base). The portal images of the similar cases can be retrieved for reference images, as well as treatment records which show examples of the radiation treatment. By this system radiotherapists can easily make suitable plans of the radiation therapy. This system is useful to prevent inaccurate plannings due to preconceptions and/or lack of knowledge. Images were stored into magneto-optical disks and the demographic data is recorded to the hard disk which is equipped in the personal computer. Images can be displayed quickly on the radiotherapist's demands. The radiation oncologist can refer past cases which are recorded in the case base and decide the radiation treatment of the current case. The file and data format of magneto-optical disk is the IS&C format. This format provides the interchangeability and reproducibility of the medical information which includes images and other demographic data.

A PACS system has been developed through a multi-phase collaboration between the Mayo Clinic and IBM/Rochester. The current system has been fully integrated into the clinical practice of the Radiology Department for the primary purpose of digital image archival, retrieval, and networked workstation review. Work currently in progress includes the design and implementation of a gateway device for providing digital image data to third-party workstations, laser printers, and other devices, for users both within and outside of the Radiology Department.

A clinical picture archiving and communication system (PACS) requires immediate access to both patients' previous and current images as well as the corresponding radiology reports. We have developed a prefetch mechanism for the PACS system in the UCSF Department of Radiology that automatically retrieves historical images and reports for reviewing at remote display stations.

One of the primary advantages of a Picture Archive and Communications System is the availability of historical images. The historical images may be prefetched from the archive the night before a patient visit based on a pull list or may be prefetched after the patient has registered for a radiological exam while the exam is being performed. A set of rules has been devised for prefetch of selected images. Further, a Madigan Army Medical Center statistics of the number of images that are prefetched and used with a new exam are examined. Madigan has more than 300,000 historical images on line for an aid in the diagnosis of new exams. The percent of patients with historical exams is examined as a function of the age of the historical exam. The result of the study is a sample of the number of images that will be prefetched and the distribution of the images over image types and time for a large set of exams.

The organization of modern medical care environments into disease-related clusters, such as a cancer center, a diabetes clinic, etc., has the side-effect of introducing multiple heterogeneous databases, often containing similar information, within the same organization. This heterogeneity fosters incompatibility and prevents the effective sharing of data amongst applications at different sites. Although integration of heterogeneous databases is now feasible, in the medical arena this is often an ad hoc process, not founded on proven database technology or formal methods. In this paper we illustrate the use of a high-level object- oriented semantic association method to model information found in different databases into an integrated conceptual global model that integrates the databases. We provide examples from the medical domain to illustrate an integration approach resulting in a consistent global view, without attacking the autonomy of the underlying databases.

A knowledge-based approach to predicting and pre-fetching prior images has been proposed to better meet radiologists' information and performance requirements during primary reading in a totally digitized radiology environment. The knowledge required to accurately predict the related images that a radiologist would wish to review for the diagnosis of a new examination includes classification/subclassification of reason-for-examination (based on examination requisition and patient information), and anatomical-portion/disease/abnormality-dependent cross-modality, cross-anatomical-portion and temporal examination relationships. Developing knowledge-based system to embody and process such knowledge involves a number of challenges. Specifically, because image retrieval knowledge infers many patient, examination and other pertinent data attributes, knowledge representation and processing need to be effective with knowledge and data relationships. Furthermore, because distributed and object- based architecture offer performance and easy-expansion advantages to database systems that house image-retrieval-related multimedia data, the design of a knowledge-based image retrieval system needs to handle distributed object-oriented data and knowledge interactions efficiently. A new generation Image Retrieval Expert System (IRES) has adopted a coupled knowledge-base/database architecture using an object-oriented knowledge and data representation. This paper will depict the architectural design as well as the detailed design/implementation results of IRES in this new generation.

The University of Pennsylvania Medical Center has been operating a prototype clinical PACS for a number of years. The details of this implementation have been described elsewhere. Early on in the design phase we realized the need to pre-fetch images on a PACS network and incorporated rule-based pre-fetch mechanisms into our Folder Manager software suite. As our PACS expanded and as the pool of users increased to include MR technologies and researchers in addition to radiologists, the case for an on-demand retrieval mechanism became compelling. The design of such a paradigm posed special problems in that the system had to be robust and user-friendly. This paper discusses the design and implementation of such a system. Usage statistics collected over a one-year period are also presented. The retrieval profile shows certain patterns that can be exploited to improve the design of the PACS.

Currently, most medical information is scattered across diverse independent information systems, such as Hospital Information System (HIS), Picture Archiving and Communication System (PACS), and other medical databases, with varying degree of interconnectivity and data integration. In this environment, our objective is to collate text data, mostly in HIS, with relevant image data, mostly in PACS, transparently from different local sources and have them as a logically integrated whole. Even though there has been much research on designing PACS DBs and discussing issues of data integration, very few have addressed the dynamic nature of data integration in the context of medical informatics. With various reasons including security, it is desirable for a user to be able to create an integrated view based on data semantics, given data range, and security classification. In addition, the integrity of the integrated view should be maintained even with the local schema changes. In order to accomplish these capabilities, generic tools which can perform semi-automated integration of information have to be developed. In this paper, we will discuss the technical background regarding data integration of different local PACS DBs and HIS. Also, based on the integrated-view (or schema), several aspects of query processing are discussed.

We have developed integrated RIS/PACS system which supports examination, interpretation and management in the diagnostic imaging department. The purposes of the system is offering a mean to support immediate and chronological comparative reading without film file transportation. The system was installed to the Toshiba hospital in Tokyo in May 1993, concurrent with a renewal of the hospital facilities.

The University of Florida Medical Center has developed an On-Line Medical Record (OLMR) that serves as a repository of patient information from a number of individual department databases, the Radiology Information System included, and builds a comprehensive electronic patient based chart. The OLMR, widely used by clinicians to view information on test results, will be expanded to add image and graphics display capabilities, and will require pointers to PACS images.

We describe the clinical requirements of the integrated federation of databases and present our client-mediator-server design. The main body of the paper describes five important aspects of integrating information systems: (1) global schema design, (2) establishing sessions with remote database servers, (3) development of schema translators, (4) integration of global system triggers, and (5) development of job workflow scripts.

To assess the type and frequency of data entry errors, we implemented an interface between a radiology information system and a Kodak image-management system. During a patient verification process that included 2,550 cases transmitted from 15 acquisition devices and archived on our system, 195 cases (7.6%) had mismatches that could not be corrected without manual intervention. These included multiple misspellings, errors in the patient identification numbers, and errors due to the incomplete entry of information from the emergency room (trauma cases). Our results clearly demonstrate the need for a comprehensive patient identification and verification function prior to permanent archiving of imaging information.

HIS/RIS-PACS interfaces are inevitably very complex due to the large differences between HIS/RIS and PACS. Furthermore, because of the high degree of mutual variation among HIS/RIS system and among PACS systems, the interface must be different for each combination of HIS/RIS with PACS. To solve this problem, a generic HIS/RIS-PACS interface has been designed and realized within the scope of the EurIPACS/HIPIN topic. The internal architecture of the HIPIN interface has been designed into a common part, which schedules and processes the messages, and specific adapters for each of the connected systems, which match the communication profile and the message syntax of the system to which it belongs. The applicability of this HIPIN interface both in technical and in clinical aspect is being demonstrated by testing and by using the HIPIN interface at two hospitals, namely the Free University Hospital in Brussels and the Philipps University Hospital in Marburg.

In this work we present our initial effort in developing a scalable and modular network architecture for cost-effective medical communication system providing real-time universal access in multimedia format. Target applications of such an integrated network design include telemedicine, teleradiology, electronic claims processing and supply ordering, teleconsultation, and home patient monitoring.

A joint project by Shands Hospital, the College of Medicine at the University of Florida, and the Gainesville Regional Utilities (GRU) has produced the design and partial completion of a comprehensive Metropolitan Area Network (MAN). The purpose of this MAN is to supply an integrated medical information network to a number of remote clinical sites both existing and planned. The joint cooperation with GRU will allow the replacement of the current power distribution intercommunications with reliable high speed optical paths.

The purpose of this project is to develop a graphical tool to analyze and optimize the image flow pattern in a hierarchical, large scale PACS local area network. The PACS network is characterized as heavy image traffic, long image flow path, heterogeneous computer nodes, multiple protocols, and physical media. A smooth operation of the PACS network depends on a centralized, timely report of image flow pattern, trouble spots, and an alarm mechanism for critical conditions. Crucial timing parameters such as queue delay from acquisition to display, processing time for archiving and reformatting, and retrieval speed, can be locally collected and reported to the network manager for statistical characterization. Moreover, network traffic information and system component failures anywhere in the image flow can be reported graphically. High urgency failure trigger an immediate alarm to the designed operator. Our network management software with user friendly graphical interface is easy to use even by non-computer personnel such as hospital administrators. It achieves total automation in trouble reporting and thus leads to improvement of the network layout and operation condition based on the analysis of image flow pattern.

During the past year, the Department of Radiology, School of Medicine, Indiana University designed, specified, and installed a campus wide network. The network supports three functions: a laser camera network to allow the transfer of hard copy images across the campus; a positron emission tomography (PET) network to allow the interconnection of the workstations comprising the PET system; and a future personal computer network to allow support of departmental administrative functions with an upgrade path to allow the display of soft copy images in physician offices and other locations in the department.

For a local area medical network, the performance of transfer protocols on top of the standard LAN protocols is explored. In this simulation the medical LAN is FDDI at 100 Mbps and the transfer protocols to be investigated are the commonly available TCP/IP and the relatively new, lightweight XTP. The performance measures used for their comparison are response time for an image and total transfer time for a series of images in one examination The image workloads ranged from current, potential digital image loads to larger image transmission loads in the future. Two different values for FDDI's target-token-rotation-time were also explored. Contrary to expectations XTP using less overhead bits than TCP/IP did not greatly outperform TCP/IP.

This paper describes deployment of a Medical Doctor's Workstation (MDWS) hosting access to the Institute's radiology image and information system testbed. On-line acquisition of CT, MR and CR images for the physician's patients was supported. JPEG-compressed images reduced image data communications requirements by a factor of ten. Image interpretation at the remote site was supported by access to on-line radiology reports. The communications link to the MDWS was provided by a pair of Combinet ethernet-to-ISDN bridges interfaced to a basic rate (two B channel) ISDN line. Results of two successive trials in a referring physician's office are reported, along with the steps taken to improve limitations identified by the system's user.

Intercontinental teleradiology is a newly implemented operational environment for the Medical Diagnostic Imaging Support (MDIS) system project, with phased teleradiology implementation at five sites coming on-line on the Korean peninsula, and four sites being phased in on the island of Oahu, Hawaii, between September 1993 and November 1994. Early implementation and testing efforts began between McConnell Air Force Base, Kansas, and the MDIS Project Office, Fort Detrick, Maryland, in the Summer of 1993. Emphasis is on explaining lessons learned and technical considerations for improving patient care on a global basis. Data on system speed, reliability, image quality, image interpretation and report turn-around is presented. The discussion will cover lessons learned on setting up an intercontinental teleradiology system and various configuration requirements for global teleradiological imaging, diagnosis, and reporting. The scope of the MDIS teleradiology implementation includes U.S. based Picture Archival Communications Systems at major medical treatment facilities which will do consultative and primary diagnostic reading of radiological images sent to them from smaller facilities all over the world.

This paper presents a performance evaluation of a digital telepathology system. In telepathology, several operations are involved in displaying source images from a remote microscope on the screen of a local workstation. At the remote site, images from a microscope are acquired by a camera and frame grabber, compressed, stored, and transferred to the local site. At the local site they are stored, decompressed, displayed, and diagnosed by collaborating pathologists. In this paper, we present the evaluation of a Roche Imaging Systems telepathology workstation.

Availability of initial radiographic images acquired in the Emergency Department (ED) for patients admitted to an Intensive Care Unit (ICU) has been a problem in our operations. It is not uncommon that images from the ED are delivered to the appropriate ICU several hours after admission, and this problem is typically magnified `off hours'. We installed a film digitizer in the ED and required technologists to digitize all chest radiographs as they came out of the film processor. These images are archived and transmitted to a workstation located near one of our busier medical ICUs. The system has been operational for eight months, and it provides reliable timely access to such images. Careful review of a large number of cases clearly demonstrated that such a system is not only feasible, but extremely effective in improving both perceptions and actual quality of radiology services in this difficult environment. Image quality was found to be acceptable for this purpose.

The U.S. military through the Medical Diagnostic Imaging Support (MDIS) system is installing teleradiology at multiple medical treatment facilities throughout the US and abroad. The goals are to improve patient care, maximize limited resources, and realize cost savings. This presentation reviews early experience with clinical use of the MDIS teleradiology configuration. Emphasis is on lessons learned in the areas of image quality, speed of image transmission, communication between sites, and the advantages of the MDIS two-way teleradiology configuration. The data is accumulated from the combined experience of the authors at multiple different sites within the continental US, Hawaii, and Korea.

The processor and network technologies are removing the old barriers to the visualization applications. For digital ultrasound imaging, the increased network speeds will replace the frame-grabbed, image imaging systems in favor of full-motion, uncompressed video. The latter system provides the sonographer with valuable feedback from a radiologist that can be incorporated into the study in real-time. This paper deals with the challenges and methods of implementing such a system over an ATM network, both in a local area network as well as a wide-area network.

Many technologies and products related to groupware, the purpose of which is to support cooperative activities in business fields, have been developed over the last decade. To explore the potential of groupware for medical applications, we have focused on face-to-face image reading teaching conferences in radiology and have developed a computer-aided conference support system. The purposes of the system are to facilitate the access to the various types of data to be used in conferences, to encourage synergy in interaction between participants, and to record discussions and thereby enable review of the conferences. The system consists of workstations, databases, and a local area network connecting these components. During a conference, each participant is assigned a workstation and uses it to view and share the data provided. The system displays patient image data on the monitor of each workstation, allows participants simultaneously to enter their comments about the images and links that connect comments to parts of images, and displays all comments and links. Comments can be sorted according to keywords and image parts selected by the lecturer to reveal problems participants have in common.

Digital imaging technology has introduced some new possibilities of forming teaching files without films. IS&C (Image Save & Carry) system, which is based on magneto-optic disc, is a good medium for this purpose, because of its large capacity, prompt access time, and unified format independent of operating systems. The author have constructed a teaching file library, on which user can add and edit images. CD-ROM and IS&C satisfy most of basic criteria for teaching file construction platform. CD-ROM is the best medium for circulating large numbers of identical copies, while IS&C is advantageous in personal addition and editing of library.

The Unified Medical Language System metathesaurus and medical semantic network (National Library of Medicine, Meta 1.3, 1993) currently contains 270,797 medical terms derived from multiple controlled vocabularies, including ICD-9, ACR, MeSH, CPT, and SNOMED. We briefly describe the structure and content of Meta 1.3, focusing on its potential value to the radiological community.

A project has been undertaken to prove the feasibility of a DICOM implementation that conforms to the modality interface user conformance profile (UCP). Five application entities have been defined. An implementation is currently underway based on the Radiological Society of North America '93 DICOM demonstration software developed by the Mallinckrodt Institute of Radiology. Two application entities implement the service class user functions specified by the modality interface UCP. The remaining three application entities implement service class provider functions. Image communication has been segregated from patient and study management functions. Based on preliminary findings we believe the DICOM modality interface User Conformance Profile to be feasible and implementable.

The Digital Imaging and Communications in Medicine (DICOM) 3.0 Standard is a multi-part document which defines point to point and network-based communication. The standard is modular in design to allow individual parts to change independently and to allow new parts to be added as needed. Part 8 of the standard specifies how DICOM applications can be layered on top of an Upper Layer Protocol defined for TCP/IP or on top of ISO network profiles.

This paper gives an overview about the DICOM trial implementation by CEN/TC251/WG4 and some background information about the development approach. It also highlights the integration of requirements into runtime software and related tools to follow future DICOM updates (additional parts) and to set-up a testbed of services for specific user orientations. Furthermore, this paper gives an status report about currently implemented DICOM features and explains the necessary future steps to install DICOM in operational environments existing in Europe.

The recent publication of the DICOM 3.0 standard for transferring images and other medical information between computers promises greatly enhanced communication and interoperability for medical applications. As part of a prototype implementation of the DICOM 3.0 standard we have also considered the problem of converting DICOM Information Objects into other standard image formats. This paper concentrates on the issues involved in converting DICOM image formats into IPI (Image Processing and Interchange) formats. The issue of conformance is restricted to consideration of information attributes which can be adequately represented in IPI formats and the tradeoffs involved when using different encoding schemes.

At the Baltimore VA Medical Center, the DHCP Integrated Imaging System and a commercial Picture Archiving and Communication System (PACS) work in concert to provide a wide- range of departmental and hospital-wide imaging capabilities. An interface between the DHCP and the Siemens-Loral PACS systems enables patient text and image data to be passed between the two systems. The interface uses ACR-NEMA 2.0 Standard messages extended with shadow groups based on draft ACR-NEMA 3.0 prototypes. A Novell file server, accessible to both systems via Ethernet, is used to communicate all the messages. Patient identification information, orders, ADT, procedure status, changes, patient reports, and images are sent between the two systems across the interface. The systems together provide an extensive set of imaging capabilities for both the specialist and the general practitioner.

Imagery and textual communications among healthcare information systems, medical imaging equipment, and picture archiving and communication systems (PACS) have always been difficult as each of these components varies with platforms, modalities, and manufacturers. With the emerging of industry standards, it become feasible to integrate all these heterogeneous, disparate medical images and textual data. This paper describes two such major industry standards: Health Level 7 (HL7) and ACR/NEMA. In conforming to the HL7 standard, we are able to share medical information between the hospital information systems, radiology information systems, and PACS. By adapting the ACR/NEMA 2.0 standard, we also can convert medical images generated from a variety of modalities and manufacturers to its standardized data format. The conversion is based on the data dictionary defined in the ACR/NEMA 2.0 document.

The first development phase of a multi-modality Picture Archiving and Communications System (PACS) has been completed at the Veterans Affairs Medical Center West Los Angeles. This initial phase involved the development of an ultrasound mini-PACS system using a commercially available PACS product line (IMPAX^TM) from AGFA Division of Miles Inc. The PACS provides acquisition, archival, display, and network printing capabilities.

We propose a scalable approach to ultrasound PACS. The general lack of any network interface capability on a large percentage of installed ultrasound scanners limits the solution available in the near term. A staged implementation beginning with a small number of ultrasound scanners interfaced to a single networked acquisition station is proposed. Initial mini-PACS may provide better utilization of the shared resources, such as archive and print servers and imagers, which would be cost prohibitive in a one-machine-per-scanner configuration. As the system requirements grow and ultrasound systems add direct network support, mini-PACS performances can overcome the initial single acquisition node bottleneck encountered with video-capture based systems, and ultrasound PACS can be integrated into a full hospital-wide PACS.

In an effort to improve patient care while considering cost-effectiveness, we developed a Picture Archiving and Communication System (PACS), which combines imaging cameras, computers and other peripheral equipment from multiple nuclear medicine vectors. The PACS provides fully-digital clinical operation which includes acquisition and automatic organization of patient data, distribution of the data to all networked units inside the department and other remote locations, digital analysis and quantitation of images, digital diagnostic reading of image studies and permanent data archival with the ability for fast retrieval. The PACS enabled us to significantly reduce the amount of film used, and we are currently proceeding with implementing a film-less laboratory. Hard copies are produced on paper or transparent sheets for non-digitally connected parts of the hospital. The PACS provides full-digital operation which is faster, more reliable, better organized and managed, and overall more efficient than a conventional film-based operation. In this paper, the integration of the various PACS components from multiple vendors is reviewed, and the impact of PACS, with its advantages and limitations on our clinical operation is analyzed.

This paper describes the evaluation of the clinical accuracy of a specific system for digitizing laser printed film (hard-copy) magnetic resonance images and transmitting them to a remote site for interpretation. The system tested is viable for digitization and transmission of hard- copy magnetic resonance images and subsequent interpretation on a remote soft-copy display station.

At the University of Florida, billing and results reporting are done through the Radiology Information System, and the radiologists needed software to meet the ACR requirements and to keep track of data on needle localizations and implants. Prior to this project all data were kept manually on paper and using various non-integrated software packages.

Almost three years ago we conducted a study to assess the impact of digital viewing stations in intensive care units. Two intensive care units, the MICU (Medical Intensive Care Unit) and the SICU (Surgical Intensive Care Unit) were studied. The results demonstrated that clinicians in the SICU (with digital viewing capability) were 91/2 times more likely to view exam results within an hour than those in the MICU (without digital display) were to receive any kind of notification of the exam results within an hour. Some of the differences observed between the two units were attributed to the differences in operation. In an effort to determine the portion of the differences attributable to differences in operation of the two units, another study was performed in the MICU after a digital viewing station was installed. These results were compared with the results obtained before the digital viewing stations were installed. Results have shown that there is not statistical difference in the times within which the exams are viewed for the two cases.

As part of a study of the use of a PACS workstation compared to film in a Medical Intensive Care Unit, logs of workstation activity were maintained. The software for the workstation kept track of the type of user (i.e., intern, resident, fellow, or attending physician) and also of the workstation image manipulation functions used. The functions logged were: no operation, brightness/contrast adjustment, invert video, zoom, and high resolution display (this last function resulted in the display of the full 2 K X 2 K image rather than the usual subsampled 1 K X 1 K image. Associated data collection allows us to obtain the diagnostic category of the examination being viewed (e.g., location of tubes and lines, rule out: pneumonia, congestive heart failure, pneumothorax, and pleural effusion). The diagnostic categories and user type were then correlated with the use of workstation functions during viewing of images. In general, there was an inverse relationship between the level of training and the number of workstation uses. About two-thirds of the time, there was no image manipulation operation performed. Adjustment of brightness/contrast had the highest percentage of use overall, followed by zoom, video invert, and high resolution display.

Preliminary data is presented from a prospective study of a Picture Archiving and Communication System (PACS) in a Medical Intensive Care Unit (MICU). These data compare the efficiency of image information communication and utilization when the MICU operates in a conventional manner using standard x-ray film as an image acquisition, storage and display medium and a digital manner when images are acquired by Computed Radiography (CR), transmitted and displayed digitally. The CR images were made available for viewing more quickly than conventional film images due to the increased automation provided by the management system of the PACS. Despite the improvement of the availability of images, the time required for the MICU physicians to utilize the image information did not change.

The digital applications in radiology are a controversial advanced which potentially will influence all areas of patient imaging. It is utilized and accepted in angiography, computed tomography, magnetic resonance, nuclear imaging and sonography. More recently Computed Radiography has gained credibility in mobile scenarios as well as specific applications from cervical spine radiography to digital fluoroscopy. Usually this acceptance is related to benefits of lesser radiation exposure or an improved presentation with an incorrect radiographic technique. One advantage of interpreting from digital information is the potential manipulation of the image presentation to the observer through windowing, leveling and edge enhancement pre and/or during image review. Additionally this digital data can be transmitted over distance and represented as hard and/or soft copy for primary or consultative review. The number and quality of the images to be viewed, the environment of the review station as well as the observer experience with conventional radiographic as well as digital image evaluation are important aspects of delivering the radiologist's product i.e. the final interpretation. This paper assesses that product, specifically addressing the question `Is the radiologist's report the same whether derived from the original analog image or from its digitized image.' The object of this study is to determine whether a digital system (3M PACS) designed for consultative viewing in a satellite department can also be used directly for primary diagnosis of conventional chest exams.

Digital communication systems have been proposed as a means of improving the flow of information between radiologists and other physicians. In the intensive care unit (ICU), physicians require more rapid access to images and interpretations than physicians in most other hospital settings. Thus these systems must be designed to ensure that rapid exchange of radiological information can be achieved. To better define system design for the electronic communication of radiological information to ICUs, this study examined bottlenecks in information flow through an integrated PACS-Radiology Information System.

Siemens Medical Engineering Group has been committed to Diagnostic Imaging from the very beginnings of x-ray imaging. The forerunner of the today's factory in Erlangen co-operated with Professor Conrad Röntgen and manufactured x-ray tubes for his research work and for the very first patient examinations in 1896. Encouraged by the positive results for the hospitals from the use of our equipment and the possibility of additional benefits expected by networking modern digital modalities, Siemens committed itself to digital image management systems (PACS) in the early 1980s [1]. Siemens now has more than ten years of experience with different application scenarios. Many of these scenarios are now running with SIENET which is the Siemens networked digital image management system. These SIENET systems are in multiple hospital or clinical settings in countries around the world. This past history has allowed us to accumulate experience from multiple configurations installed and used in these differing clinical routine environments

One of the aims of the EU subsidized project EurIPACS, is to provide evidence to assess effects and costs of PACS. To this aim, assessment studies are carried out in three hospitals. These hospitals all implement a partial and/or small-scale PACSystem in the period 1992 - 1994. The effects of these systems on communication, coordination, and efficiency are assessed by before-after measurements, according to one evaluation protocol. Most before- measurements have been completed, on the basis of which draft guidelines for implementation have been made. This paper describes the first result and the draft guidelines.

In this paper, we present our approach to developing Global Picture Archiving and Communication System (GPACS) applications using the Open Software Foundation (OSF) Distributed Computing Environment (DCE) services and toolkits. The OSF DCE services include remote procedure calls, naming service, threads service, time service, file management services, and security service. Several OSF DCE toolkits are currently available from computer and software vendors. Designing distributed Global PACS applications using the OSF DCE approach will feature an open architecture, heterogeneity, and technology independence for GPACS remote consultation and diagnosis applications, including synchronized image annotation, and system privacy and security. The applications can communicate through various transport services and communications networks in a Global PACS environment. The use of OSF DCE services for Global PACS will enable us to develop a robust distributed structure and new user services which feature reliability and scalability for Global PACS environments.

A multi-phase collaboration between Mayo Clinic and IBM-Rochester was undertaken, with the goal of developing a picture archiving and communication system for routine clinical use in the Radiology Department. The initial phase of this project (phase 0) was started in 1988. The current system has been fully integrated into the clinical practice and, to date, over 6.5 million images from 16 imaging modalities have been archived. Phase 3 of this project has recently concluded.

A modern radiology department utilizes multimedia technologies to facilitate its operation. Multimedia technologies include hardware platform, information systems, databases, communication protocols, display technology and system interface and integration. Current existing PACS can not handle this multimedia information because of its closed architectural design. This paper describes the concept of second generation PACS and its design and implementation at the University of California, San Francisco. The second generation PACS concept consists of four major components: folder manager, platter manager, image and data format standardization, and integration of heterogeneous information systems.

The Medical Diagnostic Imaging Support System at Madigan Army Medical Center has been operational in a phase approach since March 1992. Since then, nearly all image acquisition has been digital with progressively increasing primary soft copy diagnosis utilized. Nearly four terabytes of data will have been archived in compressed form by the two year anniversary including more than 300,000 Computed Radiography images.

The Department of Veterans Affairs (VA) DHCP Imaging System records clinically significant diagnostic images selected by medical specialists in a variety of departments, including radiology, cardiology, gastroenterology, pathology, dermatology, hematology, surgery, podiatry, dental clinic, and emergency room. These images are displayed on workstations located throughout a medical center. All images are managed by the VA's hospital information system, allowing integrated displays of text and image data across medical specialties. Clinicians can view screens of `thumbnail' images for all studies or procedures performed on a selected patient. Two VA medical centers currently have DHCP Imaging Systems installed, and others are planned. All VA medical centers and other VA facilities are connected by a wide area packet-switched network. The VA's electronic mail software has been modified to allow inclusion of binary data such as images in addition to the traditional text data. Testing of this multimedia electronic mail system is underway for medical teleconsultation.

In this paper we discuss the desired PACS system metrics and the large-scale PACS implementation strategy for data acquisition and distribution based on the specification of these PACS system metrics. We also describe a prototype PACS acquisition subsystem in terms of its architecture, implementation, reliability, efficiency and manageability.

The initial installation phase of Large Scale PACS has been completed at three miliary medical centers. As part of the MDIS Project, these PACS have common design characteristics, but differ in implementation. The acceptance test procedure and the results of the test are discussed. Acceptance testing included evaluation of system and component availability over a thirty day period, component and system-wide features and functions, as well as aspects of system integration. Tests were performed using the input and output functions of the PACS. Calibrated signals were injected at various points and system outputs were measured. Differences in component and system performance have been recorded in a database for tracking of status and corrective actions as well as comparison of one system to another. Variation in the performance of a PACS is both observable and measurable. Components such as diagnostic workstations, high performance film digitizers and modality interfaces were tested. Several minor deficiencies were found and are discussed. The variance in component and system performance is traced primarily to the quality of configuration and calibration practices.

Most economic studies of Picture Archiving and Communications Systems (PACS) to date, including our own, have focused on the perspective of the radiology department and its direct costs. However, many researchers have suggested additional cost savings that may accrue to the medical center as a whole through increased operational capacity, fewer lost images, rapid simultaneous access to images, and other decreases in resource utilization. We describe here a new economic analysis framework we have developed to estimate these potential additional savings. Our framework is comprised of two parallel measurement methods. The first method estimates the cost of care actually delivered through online capture of charge entries from the hospital's billing computer and from the clinical practices' billing database. Multiple regression analyses will be used to model cost of care, length of stay, and other estimates of resource utilization. The second method is the measurement of actual resource utilization, such as technologist time, frequency and duration of film searches, and equipment utilization rates. The costs associated with changes in resource use will be estimated using wage rates and other standard economic methods. Our working hypothesis is that, after controlling for the underlying clinical and demographic differences among patients, patients imaged using a PACS will have shorter lengths of stay, shorter exam performance times, and decreased costs of care. We expect our analysis framework to explain and resolve some of the conflicting views of the cost-effectiveness of PACS.

Many have argued that the introduction of a large-scale PACS system into a hospital will bring about reductions in the length of inpatient hospital stays. There is currently no convicting empirical evidence to support such claims. As part of the independent evaluation exercise being undertaken alongside the Hammersmith Hospital PACS implementation, an assessment is being made of the impact of PACS on length of stay for selected patient groups. This paper reports the general research methods being employed to undertake this assessment and provides some baseline results from the analysis of total hip replacement patients and total knee replacement patients treated prior to the introduction of PACS.

The PC-based software package PACER supports cost analysis of PACS. PACER is one of the products of the EU subsidized project EurIPACS. The installation of either partial or hospitalwide PACSystems typically involves a number of changes in requirements for resources like equipment, personnel, and use of film. These changes can to some extent be expressed in monetary units: costs or (quantifiable) benefits. The analysis thereof however, is not as straightforward as it seems. The system may grow over the years, and effects on for instance personnel are not always clear. The software package PACER assists in the process of establishing the cost components of the current film based situation and a possible PACS and allows cost analysis to be performed on the results. The calculation of results can not only be performed for a full scale PACS, but also for the various stages of transition. This paper describes the structure of the software package and possible application areas are illustrated with an example case composed of two European partial PACSystems.

Digital medical images (scanned x-ray photographs, computed tomography, magnetic resonance imagery, ...) will consume 2 - 3 orders of magnitude more space then similar numbers of other images managed by digital storage subsystems. We investigate what a hospital will have to pay for a Picture Acquisition and Communication System (PACS). Assuming that tangible displaced costs--the costs of creating, storing, and managing film-- define the upper bound of what hospitals will be willing to pay for workstations, network, and digital storage, we find the EDP solution to be affordable today and attractive in two years. The archival medium currently favored by PACS researchers in optical disk technology. Performance and cost characteristics are likely to make magnetic tape juke box technology a better choice. The published literature teaches much of what is needed for optimal data compression. Uncertainty about liability will persuade some radiologists to save data beyond the noise-determined limits. A practical system will charge for data stored and transmitted and allow each user to balance cost and risk.

There have been several excellent papers providing architectures for teleradiology. Effective teleradiology systems can be fielded today. However, cost issues arise which easily blur a decision to deploy a teleradiology system for a given hospital or regional hospital system. In this paper, a T1 infrastructure is assumed that is comprised of dedicated T1 links as well as fractional T1 links. The effects of teleconferencing are included in the analysis. Plots of the telecommunication costs provide visualization of the cost and performance issues as a function of varying degrees teleradiology and teleconference utilization. 1993 tariffs in North Carolina will be used as a baseline to arrive at some basic teleradiology cost plots and metrics. The graphs are produced by gnuplot that is freely available on many anonymous ftp sites and runs on Unix workstations as well as personal computers. The plotting commands used for the graphs are available at The Bowman Gray School of Medicine of Wake Forest University anonymous ftp site.

It is necessary to estimate system performance to ensure cost-effective system design. In addition, discrepancies between estimated performance and system performance must be established as a guidelines for the use of the simulator in advanced. We have developed and evaluated a PACS system simulator, `UNSUI/P'. UNSUI/P is based on discrete event simulator, `UNSUI', developed by Toshiba Research and Development Center. UNSUI/P was evaluated by simulating a number of PACS and comparing the results against actual system performance. Evaluations for a single component and for the entire system were carried out. Discrepancies between simulation results and actual system performance ranged from 30% to 50% for the response time in the first evaluation. In the second evaluation, simulation of the delay time for image transfer from a modality to a viewing workstation was performed and discrepancies in the delay time appear to be about 15% to 40%.

A simulator for a centralized PACS realization was developed and used with realistic input parameters, which provides reasonable functional outcome. The PACS designers may use it to build a partial or full-scale system and develop operational projections by simulating differing scenarios for PACS design and carefully evaluating whether there are any serious bottlenecks in the system and all the key requirements are met. To analyze a specific PACS configuration, the configuration simulator requires numerous parameters. These parameters are supplied by the designer and by simulation library objects that may be customized. To facilitate a comprehensive analysis of PACS, the simulation parameter sets are stored separately as a scenario, i.e., the designer initiates a session by specifying the simulation scenario number. The user-supplied assumptions, intermediate computed relationships, and program-selected library objects are all associated with this unique scenario number. The results of each simulation may then be compared with alternative scenarios in order to find the best configuration.

This paper describes a possible transformation of a PACS model into an implementation. The transformation is based on an intermediate model that specifies a system as a collection of concurrent objects that exchange messages. Characteristic for the intermediate model is that the processing of messages by objects is specified separately from communication. Due to this separation, objects do not need to known how communication takes place (synchronous or asynchronous), and do not necessarily have to know each other. The approach is illustrated using the Mimose model

A significant problem in building large-scale picture archiving and communications systems (PACS) is the production of reliable and accurate software, within a specified amount of time and cost, without impacting existing operations. PACS software management is particularly difficult because most PACS involve highly distributed processing over very heterogeneous components. We applied total quality management techniques to the problem of PACS software management. All potential users of PACS software were identified as `customers,' and we optimized the quality of service provided to them. Our methodology involves each of these customers at each stage of the software development cycle to help ensure that PACS functions meet the requirements and priorities of the majority of PACS users. We used this approach in rewriting major sections of key software modules in the UCLA PACS, resulting in code development within predictable periods of time at a predictable cost, thus enhancing our software development cycle.

A Picture Archiving and Communications System (PACS) must be able to support the image rate of the medical treatment facility. In addition the PACS must have adequate working storage and archive storage capacity required. The calculation of the number of images per minute and the capacity of working storage and of archiving storage is discussed. The calculation takes into account the distribution of images over the different size of radiological images, the distribution between inpatient and outpatient, and the distribution over plain film CR images and other modality images. The support of the indirect clinical image load is difficult to estimate and is considered in some detail. The result of the exercise for a particular hospital is an estimate of the average size of the images and exams on the system, of the number of gigabytes of working storage, of the number of images moved per minute, of the size of the archive in gigabytes, and of the number of images that are to be moved by the archive per minute. The types of storage required to support the image rates and the capacity required are discussed.

During the last decade, Picture Archiving and Communication Systems (PACS) have often been discussed and occasionally implemented. Many of the early installations were research- oriented, providing essential data and feedback, but yielding very few solutions to clinical problems. In the last five years, PACS has slowly gained recognition as a tool for addressing immediate problems and challenges associated with the film-based diagnostic imaging process. As the technology has migrated from laboratory to clinical environment, personnel responsible for implementing PACS have been challenged with an unenviable and often impossible task. A structured and thorough planning process should be executed in preparation for PACS. In general, this process is applicable for all PACS projects. On the average, this process will require 12 - 18 months for completion. Once the decision to move forward with a project has been made, it is imperative that the institution makes the commitment to allocate the necessary personnel, financial, and other resources to ensure the success of the project. Following this process can not guarantee complete success in all cases. However, doing so can minimize the problems encountered during and after implementation and reduce the risk of unfulfilled expectations.

The complexity of modern Picture Archiving and Communication Systems (PACS) confronts the potential user with a bewildering array of specifications, of which those effecting image quality are of primary importance. This paper reviews some of the basic concepts of PACS image acquisition and display and the relationship of the specifications for each to image quality. The key parameters of images quality for image digitization devices are spatial resolution, dynamic range, pixel accuracy and signal to noise ratio. On the output side critical factors are brightness, spatial and contrast resolution, stability, and uniformity. The concept of frame buffer depth vs. display depth for CRT monitors is reviewed, and the correspondence of CRT images to film based images is discussed. The principles of sampling theory and the Nyquist limit are also discussed. Through an understanding of the concepts presented in this paper, the PACS user (or potential user) will be in a better position to evaluate PACS for his/her clinical application.

Recent progress in medical imaging equipment besides evolution of each modality is the standardization of the image format, and communication method between them. When such an integrated diagnosis and therapy system is realized, operators must manipulate more than one equipment in order to receive the full benefit of the multiple-modality environment. This paper describes our basic concept, and some prototype design toward unified operation of networked medical equipment.

The ability to access a vast array of radiological and pathologic diagnoses through computer searches of local medical facility databases is a by-product of the continued development of filmless imaging systems. The Department of Defense (DoD) Medical Diagnostic Imaging Support initiative is expanding through the addition of on-line systems at several DoD health care facilities. Madigan Army Medical Center, as the initial site, will soon be 90% filmless, with over one million images archived. Multiple other DoD medical centers are under installation. The eventual goal is an interconnected network of PACS systems of DoD medical centers and their supported medical facilities throughout the United States. To access this potential pool of medical information requires a centralized database capable of acting as a diagnostic index system. The establishment of a multi-center film library index begins with an initial analysis of issues regarding data storage and access, indexing, cross-coding with pathological files, communication formats, cost sharing, and patient confidentiality. In initiating these first steps to developing this telecommunications library these issues and their implications are discussed. The final implementation of this system will facilitate markedly improved research and teaching capabilities in both radiological and pathological fields.

Automating image prefetching to meet reference requirements for primary reading an area having significant potential for improving the radiology service level and one to which attention has been long-due. IRES, a knowledge-based approach to image retrieval, has a natural appeal to radiologists over ad hoc rules of thumb because its embedded image retrieval knowledge is in concert with the image reference patterns and heuristics most of them use. This paper discusses the knowledge acquisition process and validation framework, presents some interesting validation results, and outlines future research directions.

By using a film digitizer it is possible to digitize X-ray film images with high quality. But from a practical point of view, conventional film digitizer systems are not satisfactory, because the identification of each film, that is indispensable for an image filing system, should be manually recorded. This procedure takes a long time and is cumbersome. In this paper we propose an automatic recognition method of film IDs using Synthetic Discriminant Function filters with an error check code, which greatly improves the recognition ratio. From experimental results it is confirmed that the proposed method is practically feasible. A full automatic film digitizer system is also described.

Powerful communication methods based upon multi-media and network technology are becoming available at reasonable cost. At the same time there is an increased awareness of the need to control costs and improve use of scarce resources by effective use of technology. Combining powerful communication methods with teleradiology will enhance its effectiveness and may aid in reduction of costs.

The discussion is no longer whether to use a Picture Archiving and Communications System (PACS). The fundamental question is whether to use a departmental or modality oriented system or to implement a Full Hospital PACS. There is a vast difference in the approaches that goes beyond simple cost justification. A Full Hospital PACS promises to fundamentally change the paradigm of radiological consultation as digital images with improved availability, accuracy, speed and economy become available to all of the physicians of the hospital any place and at any time. Certainly there will be savings in film and staffing for film rooms as the hospital goes primarily filmless. The benefits to the hospital and to the patient go far beyond these savings, resulting in hours saved per day for each radiologist, clinician, and support staff at the hospital. Measures of radiologist productivity are changing as the images become available to the clinician at the same time as they are available to the radiologist. Marked increases in patient exams performed have been measured. Radiology procedure backlogs have been reduced. The benefit to the patient is access to better medical care on short notice and with reduced costs. The ramifications of a Full Hospital PACS go beyond cost justification on savings using existing methodology to the consideration of whole new possibilities such as real time studies, time sequenced trend analysis, and an evolution to telemedicine.

The DICOM standard (Digital Imaging and Communications in Medicine) has been developed in order to obtain compatibility at the higher OSI levels. This paper describes the implementation of DICOM into our developed low cost PACS, which uses as much as possible standard software and standard protocols such as SQL, X Windows and TCP/IP. We adopted the requirement that all messages on the communication network have to be DICOM compatible. The translation between DICOM messages and SQL commands, which drive the relational data base, has been accommodated in our PACS supervisor. The translation takes only between 10 and 20 milliseconds. Images, that will be used the current day are stored in a distributed, parallel operating image base for reasons of performance. Extensive use has been made of X Windows to visualize images. A maximum of 12 images can simultaneously be displayed, of which one selected image can be manipulated (e.g., magnified, rotated, etc.), without affecting the other displayed images. The emphasis of the future work will be on performance measurements and modeling of our PACS and bringing the results of both methods in agreement with each other.

A new radiological exam spawns a collection of image movements in a PACS system. Historical images are fetched, images are moved to workstations for diagnosis, selected images are moved to other workstations for clinicians and for consultations. The distributions of the new images over the working day will establish the image load to which the PACS system must respond. At Madigan Army Medical Center 10.1% of the daily load of images occurs in the busiest hour of the day. Individual days may have an hour that has up to 15% of the average daily load. The PACS system must be able to handle the images that are spawned by the average 10.1% busy hour load with margin to be able to handle a 50% overload with acceptable degradation of the performance of the system.

At the Radiology Department at Skejby University Hospital a single modality PACS has been operational since September 1992. The system is being operated in combination with a CR reader (Digiscan). The PACS has undergone a clinical test during the same period. Ongoing upgrading of the system is being carried out. The image quality has constantly been very high. Even in test version it was possible to use the system beneficially and with few disruptions. The system has been highly accepted by all users, including distributed parts of PACS. A digital radiology system in combination with PACS and diagnostic reading of soft-copies helps advance the diagnostic quality and increases the efficiency at the Radiology Department and at the hospital in general.

This software was designed to satisfy the requirements of the 1993 DICOM demonstration and is not intended to provide a complete system. However, this software does provide an example implementation which can be used to gain understanding about the DICOM Standard and has successfully communicated with the 20 vendors who participated in the DICOM demonstration at the 1993 RSNA. This software is written in ANSI-C and has been compiled under the following operating systems: SunOS, Solaris, Ultrix, OSF/1, Irix, NextSTEP. This poster describes the implementation of the software, how the software can be used, and how the software can be obtained.

Osaka University hospital moved to the new hospital in September 1993, where HIS and RIS are installed and a part of PACS will be installed to provide safe and convenient circumstances for patients and hospital staffs. We investigated the flow of radiological examination orders, X-ray films, and diagnostic reports in the previous film-based system, and estimated the data volume in digital PACS. We designed (1) the total hospital information system and PACS, (2) RIS, FCR and controller connection and (3) RIS, handy terminal and FCR connection.

The new hospital is called `Intelligent Hospital,' where HIS and RIS are installed and a part of PACS will be installed in March 1994. The aim of this system is to optimize the flow and stock of materials and information in the hospital and to provide safe and convenient circumstances for patients and hospital staffs. The preliminary study of a small PACS suggested to us that the integration of HIS, RIS and PACS increase each value of them as a total information system. In this paper we will show the software and concept of our system. (1) Integration of HIS and RIS; concerning with radiological examinations, HIS has the radiological examination order entry system and its scheduling system. The radiologist can get the patient's information at examination and reporting on the RIS terminal and input the drug and film consumption. The data transferred to HIS as billing data. The examination report is also transferred to HIS. (2) Integration of RIS and modalities; RIS sends the patient's data and examination code to FCR (Fuji. Co.) system. (3) Integration of RIS, HIS and PACS; HIS shows reference images with report on the HIS terminal.

In the field of (medical) electronic data interchange a lot of `standards' or `protocols' are developed in order to make interchange work in a multi-vendor environment. These standards range from providing guidelines for a file format, to specifying possible dialogues between systems. Which standard to choose among the candidates, depends on the purpose it is used for. In the case of HIS/RIS and PACS integration a fullfledged communication protocol is needed.

The introduction of a new technology, such as PACS, on a large scale into a clinical environment requires a comprehensively designed training plan to achieve success in implementation of the system and minimal impact on the clinical operations of the facility. A tremendous range of experience and computer literacy will be found in the personnel who require training in some aspect of operation of the PACS. This presentation will describe the development and implementation of such a comprehensive plan.

The Mayo Clinic and IBM (at Rochester, Minnesota) have jointly developed a picture archived system for use with Mayo's MRI and CT imaging modalities. This PACS is made up of over 50 computers that work cooperatively to provide archival, retrieval and image distribution services for Mayo's Department of Radiology. This paper will examine the performance characteristics of the system.

This paper describes a research project which explored the reasons for variation in reporting time by senior radiologists. The objective was to identify factors which are important to control for in an assessment of the impact of the introduction of a PACS on reporting times. An observational study of reporting by senior radiologists at Hammersmith Hospital was initially undertaken for a period of 25 days. At all times during the working day on the days chosen for the study, an independent health service researcher observed the radiologist reporting on both urgent and non-urgent images. Data were collected on a variety of factors including the time taken to produce the report, the number and nature of the images viewed and the disturbances that occurred. Regression analysis techniques were used to analyze the data in order to identify the factors which explain variation in reporting times. The paper reports the research methods and results from the baseline data on working with conventional hard copy film at Hammersmith Hospital.

Quantitative description of the value of Picture Archiving and Communication System (PACS) in the health care environment is essential to assess the benefits with which PACS provides us. It is necessary to obtain data related to our film-based system before PACS installation in order to make quantitative comparison of before-and-after PACS operation, so that we can assess the clinical effectiveness brought by the PACS itself. We conducted flow study to acquire the whole chart of data flow related to radiological examination. And then defined the parameters and employed them to conduct time study for measuring several items concerning order entry, report making and image delivery at outpatient clinics and wards of eight departments using stopwatches, IC cards and time-recorders. This paper describes the results of measurement. For example, the location of each clinic/ward affected the time spent for images/reports delivery, and that hand writing time occupies more than half of the time required for making a single diagnostic report in 16 percent of cases. These results indicate that acceptability by medical staff is the key to successful operation of PACS. For instance, simple man-machine interaction at image display workstation must be properly realized. By using the same parameters mentioned above, we can compare those data with the results measured after PACS operation in a few years.

Image Save and Carry (IS&C) is the standardized format for recording medical images on 130-mm (5 1/4 inches) rewritable magneto-optical disks (MOD). One of the most important characteristics of an image display station is its display speed, and this paper describes the software techniques used to increase the speed with which data is transferred between an IS&C MOD drive and an image display workstation. The techniques make it possible to transfer more than 900 kBytes in a second while reading.

Medical information system as an infrastructure is required to be developed in Japan in order to establish, for example, a preventive medicine and better health care, that are indispensable in an aging society. For this purpose Japanese government is giving financial support to research projects, where compact media such as magneto-optical disk and IC-card are supposed to be used. This paper describes the current activities of these projects and discusses the requirements that medical information system for wide use should satisfy.

The initial transition of the Radiology Department at Madigan Army Medical Center, the Wright Patterson Air Force Medical Center and the Brooke Army Medical Center from film- based operations to direct digital image capture and display has been completed. This presentation describes the planning process and the impact of the transition on radiology operations and clinical services. PACS implementation requires changes in both the physical plant and the human element of the Radiology departments as well as in the clinical areas where imaging workstations were installed. Equipment retrofit, utility upgrades, space trade- offs, quality control operations, work flow variations, and educational requirements were major considerations. An overview of the scope of departmental transitions is achieved.

This paper defines the roles of customers and suppliers, customer variables, and customer contributions to the system creation process, discusses expected outcomes for a customer- driven system development process, and illustrates how this approach was used in the development of digital radiology.

A wide area computer network has been installed at the Austin Hospital, Melbourne. The network consists of optic fiber segments between buildings, an ethernet spine through the main buildings and ethernet segments throughout each department. The network is connected to Internet via an ISDN link to the University of Melbourne computer network. The Austin hospital network is used for intra-hospital image distribution, external image distribution, internal and external electronic mail via Internet, electronic information access, file transfer via Internet, and remote login to Internet networked computers. Present and future developments include secure patient record access for internal users, confidential information transmission using public key encryption techniques, external dial-up connectivity for teleradiology, and research and development into medical image processing and analysis.

A workstation for radiographic images, known as the Arizona Viewing Console (AVC), was developed at the University of Arizona Health Sciences Center in the Department of Radiology. This workstation has been in use as a research tool to aid us in investigating how a radiologist interacts with a workstation, to determine which image processing features are required to aid the radiologist, to develop user interfaces and to support psychophysical and clinical studies. Results from these studies have show a need to increase the current image memory's available storage in order to accommodate high resolution images. The current triple-ported image memory can be allocated to store any number of images up to a combined total of 4 million pixels. Over the past couple of years, higher resolution images have become easier to generate with the advent of laser digitizers and computed radiology systems. As part of our research, a larger 32 million pixel image memory for AVC has been designed to replace the existing image memory.

We have developed and installed a Medical Image Access System in an intensive care unit. Images are acquired and transmitted automatically to this system, thus expanding on the previous results of Shile et. al. It is our goal to determine what effect regular, sustained availability of image data in the clinic has on the Intensive Care Unit and the Department of Radiology. Our system is installed and has been in regular use in the hospital since late August of 1993. Since the time of installation we have been collecting usage information from both the manual and automated systems. From this data we are performing the standard measures established by DeSimone et. al. Our initial results support the original findings that image availability in the clinic leads to earlier patient care decision based on the image data. However, our findings do not seem to indicate that there is a breakdown of communication between the clinician and the radiologist as a result of the use of the clinical display system. In addition to the established measure we are investigating other criteria to measure time saved by both the clinician and radiologist. The results are reported in this paper.