The National Institutes of Health Clinical Center Digital Imaging Network, Picture Archival and Communication System, and Radiology Information System.

In this work, we describe the digital imaging network (DIN), picture archival and communication system (PACS), and radiology information system (RIS) currently being implemented at the Clinical Center, National Institutes of Health (NIH). These systems are presently in clinical operation. The DIN is a redundant meshed network designed to address gigabit density and expected high bandwidth requirements for image transfer and server aggregation. The PACS projected workload is 5.0 TB of new imaging data per year. Its architecture consists of a central, high-throughput Digital Imaging and Communications in Medicine (DICOM) data repository and distributed redundant array of inexpensive disks (RAID) servers employing fiber-channel technology for immediate delivery of imaging data. On demand distribution of images and reports to clinicians and researchers is accomplished via a clustered web server. The RIS follows a client-server model and provides tools to order exams, schedule resources, retrieve and review results, and generate management reports. The RIS-hospital information system (HIS) interfaces include admissions, discharges, and transfers (ATDs)/demographics, orders, appointment notifications, doctors update, and results.

NCI image archive management workshop: a preliminary report.

The National Cancer Institute organized a workshop entitled "Image Archive Management" that was presented on August 28 and 29, 2000, at the Natcher Conference Center on the National Institutes of Health (NIH) campus. The purpose of this workshop was to solicit expert input for the planned development of an archival system to make imaging databases readily accessible by the broad scientific community. The specific goals were to (a) define the technical requirements for a virtual archive of images used in oncology, (b) define the policy issues for access to these images, (c) recommend a process and phases for implementation of a robust imaging archival system, (d) review how this effort could be expanded and coupled with other ongoing efforts by NIH and other organizations interested in imaging, and (e) form an overall plan and policy to allow interoperability of image data archives. Representatives who attended the workshop came from academia, government agencies, and large and small businesses. A preliminary report was generated, as outlined herein, and additional reports are anticipated from the steering committee being organized as one of this workshop's recommendations, which is expected to be active by summer 2001. Additional information, including the list of participants in this workshop, is available at the Biomedical Imaging Program Web site (http://www.nci.nih.gov/bip/).

National Cancer Institute initiative: Lung image database resource for imaging research.

Preliminary clinical studies suggest that spiral computed tomography (CT) of the lungs can improve early detection of lung cancer in high-risk individuals. More clinical studies are needed, however, before public health recommendations can be proposed for population-based screening. Spiral CT generates large-volume data sets and thus poses problems in terms of implementation of efficient and cost-effective screening methods. Image processing algorithms such as computer assisted diagnostic (CAD) methods have the potential to assist in lesion (eg, nodule) detection on spiral CT studies. CAD methods may also be used to characterize nodules by either assessing the stability or change in size of lesions based on evaluation of serial CT studies, or quantitatively measuring the temporal parameters related to contrast dynamics when using contrast material-enhanced CT studies. CAD methods therefore have the potential to enhance the sensitivity and specificity of spiral CT lung screening studies. Lung cancer screening studies now under investigation create an opportunity to develop an image database that will allow comparison and optimization of CAD algorithms. This database could serve as an important national resource for the academic and industrial research community that is currently involved in the development of CAD methods. The National Cancer Institute request for applications (RFA) (CA-01-001) has already been announced (April 2000) to establish and support a consortium of academic centers to develop this database, the consortium to be referred to as the Lung Image Database Consortium (LIDC). This RFA is now closed. Five academic sites have been selected to be members of the LIDC, the first meeting of this consortium is planned for spring of 2001, and a public meeting is to be held in 2002. This report is abstracted from the previously published RFA to serve as an example of how an initiative is developed by the National Cancer Institute to support a research resource. For specific details of the RFA, please access the following Internet site: http://www. nci.nih.gov/bip/NCI-DIPinisumm.htm#a11.

New formula for quantification of pleural effusions from computed tomography.

The authors develop a method to accurately and easily estimate the volume of pleural effusions with computed tomography (CT). In 15 patients with either simple or loculated pleural effusions (14 right-sided and 11 left-sided), routine helical CT examinations of the thorax were obtained. Two experienced radiologists visually estimated the volume of the effusions. Three-dimensional reconstructions of the pleural effusions were performed from the helical CT examinations, and the volumes of the effusions were calculated. Effusion volumes were also estimated using the formula d2 x l (d = greatest depth of the effusion on a single CT image, l = greatest length of the effusion). The computer calculated the estimated volumes and they were then statistically compared. The coefficient of correlation between the estimation by measurement and calculated volumes of all effusions was 0.908 (p<0.0001) for the right side, and 0.849 for the left side (p<0.002). Excluding the loculated effusions, the coefficient of correlation was 0.969 for the right side and 0.949 for the left side (p<.001). The volume estimation by visual inspection was also correlated to the calculated volumes (0.84 on the left and 0.85 on the right, p<0.008), but resulted in a consistent overestimation by 300 ml to 500 ml. Although pleural effusion volumes can be estimated by visual inspection with good correlation, some overestimation is consistently seen. Use of the formula d2 x l readily enables estimation of pleural effusion volume from CT, from two simple measurements. This formula-based method of volume estimation provides an accurate and easily measured means of estimation that is readily obtained from routine CT of the chest.

Three-dimensional volumetric assessment of abnormally low attenuation of the lung from routine helical CT: inspiratory and expiratory quantification.

OBJECTIVE: The purpose of this investigation was to quantitatively assess abnormally low attenuation of the lung by use of three-dimensional volumetric reconstructions from routine helical CT and to assess their correlation with pulmonary function tests. MATERIALS AND METHODS: Helical CT was performed in 100 patients in full inspiration. Examination was also performed in full expiration in 53 of these patients. Three-dimensional volumetric reconstructions were performed for total lung volumes at inspiration and at expiration, with a threshold of -896 H on inspiratory CT and -790 H on expiratory CT, to quantify emphysematous change. Correlation was made with pulmonary function tests in 79 patients. RESULTS: CT volumetric assessments of abnormally low attenuation of the lung at inspiration and expiration had a high correlation (r2 = .84, p < or = .0001). In comparison with pulmonary function tests, both inspiratory low attenuation of the lung and expiratory low attenuation of the lung correlated well with the logarithm of the ratio of the forced expiratory volume in 1 sec (FEV1) to the forced vital capacity (r2 = .74, p < or = .0001 and r2 = .74, p < or = .0001, respectively) and with the percentage of predicted ratio of the FEV1 to the forced vital capacity (r2 = .69, p < or = .0001 and r2 = .69, p < or = .0001, respectively). Linear correlations were also seen with FEV1, residual volume, and forced residual capacity. CONCLUSION: Three-dimensional volumetric reconstructions of hypoattenuating lung correlate well with pulmonary function tests. In addition, inspiratory and expiratory data are also correlative, suggesting that a dedicated expiratory examination is not needed. This easily obtainable information will prove useful for patients with obstructive lung disease from emphysema, providing a measure of pulmonary function status in this population.

Current status of computed radiography in emergency departments.

This study reports the findings of a computed radiography (CR) imaging experience questionnaire sent to 35 emergency departments (ED) in North America. A total of 25 responses to the questionnaire were received corresponding to a return rate of 71%. The median daily workload was 71 patient examinations and the average number of films per patient examination for the 21 facilities was 3.0 +/- 0.8. A total of 91% of respondents printed to film all ED trauma images obtained with CR with only one ED claiming to be filmless. CR in the ED was easy to use and had significant benefits of reducing examination repeat rates, permitting a prompt availability of radiographic images, improving image quality, providing improved operational efficiency, and eliminating lost films. Major limitations of CR were deemed to be limited viewing stations, CR costs, and inefficient patient ID entry. Radiology departments were very happy with the introduction of CR into the ED setting with approximately half being highly satisfied and half somewhat satisfied. The degree of satisfaction by ED personnel was similar with about half being highly satisfied, 40% somewhat satisfied, and the remainder neutral. The fact that not a single respondent was in any way dissatisfied shows that CR can play a useful role in the ED setting.

Picture archiving and communication system bandwidth and storage requirements.

The purpose of this work was to determine the requirements for image storage and network bandwidth for a total digital department in a moderate sized academic radiology department. Data from the radiology information system was combined with image production information to produce a model of image acquisition. Destinations of images to reading rooms were studied to determine the final distributions of film. All findings were used to model the flow of data that would be expected if the images in the department were completely digital. Using today's standards, the department would produce approximately 15.7 Gbytes of data per day or 3.5 Tbytes of data per year if all acquisitions were digital. The peak acquisition rate would be 1.8 Gbytes per hour with a sustained rate greater than 1 Gbyte per hour for most of the working day. The anticipated bandwidth for the total digital department exceeded the capabilities of the existing picture archiving and communication system equipment. A distributed networked archive solution was shown to optimize access to images by radiologists and referring clinicians.

Consensus quest: reshaping the future of radiology. RSNA Hartman Lecture/Eugene P. Pendergrass New Horizons Lecture.

We are practicing medicine in an era of economic revolution and find ourselves re-examining virtually all of our beliefs, practices, rituals, habits, and ethics. This revolution is not driven by the government but--in large part--by economic factors that work to lower costs by restricting access to the health care system. Re-thinking all that we do and formulating responses in this new era should be viewed in a positive light; opportunities abound for radiologists to revive enthusiasm, stimulate creativity, deepen their involvement, and intensify their commitment. Are we capable of identifying what is valuable in our specialty and of safeguarding its stature? Can we develop a vision of what our future should be? Can we convince the majority of radiologists that change is inevitable and that a reshaping of our profession is critical? Can we develop a consensus about what we value most? Can we continue to entice bright individuals into our field? Can we identify our sundry missions and use them as guideposts? Can we create effective strategies for dealing with discontinuous change? Our rich heritage of innovation and our skills as consummate managers of medical information suggest that we can. The challenges will be great but the rewards are worth our best efforts.

Computed radiography and film digitizer inputs to an intensive care unit teleradiology system: an image quality comparison.

RATIONALE AND OBJECTIVES: We compared computed radiography (CR) with a film digitizer as an image input device for transmitting radiographs to intensive care unit (ICU) displays. METHODS: Limiting spatial resolution and low-contrast detectability performance were determined for a 600-speed screen-film combination and CR films. The same image data were transmitted to ICU displays directly from the CR or by digitizing the conventional film. RESULTS: CR resolution ranged from 2.5 to 3 line pairs per millimeter (Ip/ mm) depending on cassette size. Display station resolution for the CR image data was 1.5-1.9 lp/mm, but improved resolution could be achieved using display magnification modes. Film digitization resulted in a loss of resolution. Direct transmission of CR image data to display stations gave low-contrast detectability similar to that obtained with CR film. CONCLUSION: ICU teleradiology displays that use CR, rather than film digitizers, offer improved image quality and superior operational efficiency.

A cost-analysis of computed radiography and picture archiving and communication systems in portable radiography.

A total of 40,000 portable examinations are performed each year at Shands Hospital (Gainesville, FL), a 570-bed teaching hospital. Radiographs are obtained using a screen-film combination with the films digitized for transmission to displays in four intensive care units. A cost-analysis of replacing screen-film with computed radiography (CR) integrated into a filmless picture archiving and communication system (PACS) network was performed. Equipment requirements included two CR units, three high-resolution dual monitor displays, and an archive to store 3 months of image data. The capital costs were amortized over a 5-year period. Capital and operating costs of the proposed expansion to the existing PACS network, together with anticipated cost savings, were determined. The maximum data transfer rate for portable examinations was 150 MByte per hour and approximately 400 GByte of image data are generated each year. These figures were used to determine the hardware requirements for handling the acquisition, transfer, and display of the images. Annual costs of the proposed expansion were about $220,000. Cost savings were achieved by elimination of film, including its handling by technologists/library clerks, and amounted to about $200,000 per year.

Picture archiving and communications systems (PACS).

Although there has been a recent increase in interest in picture archiving and communications systems (PACS) topics, little has been published to assist the non-technical person in understanding the complexities of the technologies required for a PACS implementation. This issue of Current Problems in Radiology defines each PACS component and explains why each is important in a system design. PACS installations at the University of Florida are used as examples to tie the concepts together. The infrastructure required for PACS consists of the information system interfaces, networks, and databases. Information system interfaces guarantee consistent patient data across all platforms and reduce labor requirements by eliminating duplicate data entry. Data networks move information from the originating location to users around the hospital, clinic, campus, city, or world. In the PACS environment, the data consist of patient and study information as well as images and information about these images. Databases organize the data from multiple sources into a coherent package that can be queried for many different purposes, such as retrieving images, reviewing patient and study information, studying practice statistics, and performing outcomes analysis. PACS components consist of acquisition nodes, archives, and output devices. Acquisition nodes may include "digital modalities" such as CT, MRI, nuclear medicine, and computed radiography (CR), along with devices to convert from analog to digital, such as digitizers and frame grabbers. Options for archives are discussed along with configuration schemes. Output devices include both hard copy (film and paper prints) and soft copy (workstations for display and diagnosis). Finally, a description of the PACS installations at the University of Florida is presented, with comments on some of the difficulties and complexities encountered. A discussion of the cost and benefits of PACS is included, along with a forecast of the future of PACS.

infoRAD: computers for clinical practice and education in radiology. Teleradiology: fundamental considerations and clinical applications.

Teleradiology, the practice of radiology from a distance, involves transmission of images between two locations. Radiologists off site can be consulted easily and quickly whenever critical questions arise. Similarly, long-distance, real-time radiologic monitoring of patients is possible. Second opinions can be obtained from specialists in difficult cases. The result is improved and more efficient delivery of patient care. The large number of teleradiology products and features often confuses radiologists who are considering the incorporation of teleradiology into their practices. It is important that practice patterns and needs be evaluated prior to specification and acquisition of a teleradiology system.

Angelman and Prader-Willi syndrome: a magnetic resonance imaging study of differences in cerebral structure.

Recent improvements in magnetic resonance imaging techniques now allow the developing brain to be visualized in sufficient detail to perform "in vivo neuropathology." In this study we compared the cortical morphology in six children with Angelman and four with Prader-Willi syndrome. These two syndromes are of special interest because, although they are both caused by deletions in the same region of chromosome 15, Angelman children are far more severely affected, and do not speak. We measured the length of the banks of the Sylvian fissure in a gapless series of thin sagittal images. Angelman children had a significantly larger proportion (75%) of anomalous fissures than the Prader-Willi children (12%). Anomalous cortical growth could result from mistimed expression and recognition of macromolecules involved in axonal guidance, target recognition, and pruning. We hypothesize that misrouting of long projection axons may be related to the Sylvian fissure anomalies and the language disorder in Angelman syndrome.