DOI: 10.1148/rg.266065707
RadioGraphics 2006;26:1877-1885
© RSNA, 2006
Informatics in Radiology (infoRAD)
Vendor-Neutral Case Input into a Server-based Digital Teaching File System1
Aaron W. C. Kamauu, MD, MS, MPH,
Scott L. DuVall,
Reid J. Robison, MD, MBA,
Andrew P. Liimatta,
Richard H. Wiggins, III, MD and
David E. Avrin, MD, PhD
1 From the Department of Biomedical Informatics, University of Utah, 26 South 2000 East HSEB 5700, Salt Lake City, UT 84112-5750 (A.W.C.K., S.L.D.); the Departments of Psychiatry (R.J.R.) and Radiology (R.H.W., D.E.A.), University of Utah Health Sciences Center, Salt Lake City; and the Utah Center for Advanced Imaging Research, Center for Advanced Medical Imaging, Salt Lake City (A.P.L.). Received February 15, 2006; revision requested June 2; final revision received August 15; accepted August 16. Supported in part by training grant 2T15 LM07124 from the National Library of Medicine, National Institutes of Health; the Educational Computing Committee at the University of Utah School of Medicine; and Amirsys, Salt Lake City. S.L.D., A.P.L., R.H.W., and D.E.A. are consultants to Amirsys; R.H.W. and D.E.A. are also stockholders in Amirsys.
Address correspondence to A.W.C.K. (e-mail: Aaron.Kamauu{at}hsc.utah.edu).
 |
Abstract
|
|---|
Although digital teaching files are important to radiology education, there are no current satisfactory solutions for export of Digital Imaging and Communications in Medicine (DICOM) images from picture archiving and communication systems (PACS) in desktop publishing format. A vendor-neutral digital teaching file, the Radiology Interesting Case Server (RadICS), offers an efficient tool for harvesting interesting cases from PACS without requiring modifications of the PACS configurations. Radiologists push imaging studies from PACS to RadICS via the standard DICOM Send process, and the RadICS server automatically converts the DICOM images into the Joint Photographic Experts Group format, a common desktop publishing format. They can then select key images and create an interesting case series at the PACS workstation. RadICS was tested successfully against multiple unmodified commercial PACS. Using RadICS, radiologists are able to harvest and author interesting cases at the point of clinical interpretation with minimal disruption in clinical work flow.
© RSNA, 2006
|
Introduction
|
|---|
Digital teaching file (DTF) systems play an important role in radiology education (1,2). Although some picture archiving and communication system (PACS) vendors enable images to be exported in desktop publishing formats, none currently offer a satisfactory solution for the creation of DTFs (3–7). Some academic institutions have struggled with a variety of methods to design in-house vendor-specific DTF solutions, while many others lack the resources to do so (8–12). The Radiology Interesting Case Server (RadICS) is a vendor-neutral DTF and a server-driven, Web-based application that provides a solution without requiring any modification of PACS configurations.
This article describes innovative use of Web technology and information management in solving problems commonly encountered in the development of a DTF based on ideal characteristics (3). RadICS is a solution to the problem of efficiently harvesting key images from an interesting study in any Digital Imaging and Communications in Medicine (DICOM)–compliant PACS with minimal radiologist work flow disruption.
|
Materials and Methods
|
|---|
The hardware for the RadICS server consists of a Pentium 4 (Intel, Santa Clara, Calif) 2.0-GHz processor, 512 MB of double-data-rate (DDR) random-access memory (RAM), and 80 GB of 7200-rpm hard drive storage. Linux Red Hat 8.0 was chosen as the operating system, and Apache Tomcat version 4.1.27 was installed and configured as a stand-alone server running on port 80. The server is protected behind the radiology department PACS firewall and is HIPAA (Health Insurance Portability and Accountability Act) compliant (13).
RadICS is composed of three principal components: an image server, a Web server, and a relational database. The image server receives DICOM images from the PACS, manages the initial image processing, and moves the images to temporary file storage. The Web server displays the images for key image selection, interacts with the temporary and permanent file storages and the relational database, and manages the creation of interesting cases (Fig 1).
View larger version (25K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. Imaging studies are pushed from the PACS to RadICS. Automatic image processing is handled by the image server, and the original DICOM files and newly generated Joint Photographic Experts Group (JPEG) files are moved to temporary (Temp) file storage on the server. The case authoring process, including key image selection, case description input, and key image optimization, is managed by the Web server. Selected key images are copied to permanent file storage, and case information is stored in a MySQL (Cupertino, Calif) relational database. HTML = hypertext markup language, JSP = Java Server Pages.
|
|
|
Image Server
|
|---|
A DTF could query a DICOM-compliant PACS and still be vendor neutral, but it would require additional information about the context of a case prior to the query. To achieve vendor neutrality, RadICS does not query the PACS archive directly. Instead, studies are pushed by the radiologist from a client workstation, such as one in a reading room. RadICS does not require additional information, as the case context is transferred from the PACS when a case is sent. Some PACS workstations support direct DICOM SEND, where the workstation is capable of sending images stored in local memory. Other workstations support indirect DICOM SEND, where the workstation relays a command and images are sent from a separate PACS server component. A minority of PACS vendors support DICOM SEND of user-selected key image virtual series. This simplifies key image selection in case creation. However, most PACS support only DICOM SEND of entire studies. Both direct and indirect DICOM SEND of whole studies or virtual series are accepted by RadICS (14).
The image server was written in the Perl programming language and manages the DICOM Receive and DICOM-to-JPEG conversion process (Fig 2). A DICOM Storage Service Class Provider (SCP) handles all incoming DICOM files and moves them to an incoming directory. The DICOM SCP was obtained from the Olden-burger Forschungs- und Entwicklungsinstitut für Informatik-Werkzeuge und –Systeme (OFFIS) DICOM Toolkit (DCMTK), an open-source collection of libraries and applications for constructing and converting DICOM image files and sending and receiving images over a network connection (15). A Perl script constantly monitors the incoming directory for incoming DICOM files. The presence of a DICOM file in the incoming directory initiates the rest of the image server processes. Each image is handled independently.
View larger version (28K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2. The image server manages the DICOM Receive and DICOM-to-JPEG conversion processes. A DICOM Receive Perl script runs the Oldenburger Forschungs- und Entwicklungs-institut für Informatik-Werkzeuge und –Systeme (OFFIS) DICOM toolkit (DCMTK) DICOM Service Class Provider (SCP) process. The image server parses the DICOM Header and determines appropriate window and level settings. All magnetic resonance (MR) images are processed through the Automatic Window and Level Calculator (AWLC), while all other images use preset window and level (WL) settings. The Perl script passes the window and level settings as parameters to the DCMTK, which generates full-size and thumbnail JPEG files in temporary (Temp) file storage.
|
|
The image server first parses an image’s DICOM header using the DCMTK and retrieves DICOM-specific elements including PatientsName, AccessionNumber, SeriesNumber, ImageNumber, Modality, StudyDescription, SeriesDescription, PatientsSex, PatientsAge, Row, Column, and PhotometricInterpretation (14). This information is used by other image server processes, also behind department firewalls to maintain HIPAA compliance (13).
Next, appropriate window and level settings are determined. In the conversion of medical images from DICOM to desktop publishing format, contrast resolution may be lost. While it is common for DICOM images to allocate 16 bits of memory to describe each gray-scale pixel, desktop publishing formats usually use only 8 bits. The loss of contrast resolution increases the importance of appropriate window and level settings.
In the transfer of interesting cases to RadICS, two obstacles became apparent. First, the window and level values optimized by a radiologist at the PACS workstation were not transferred with the DICOM images. Second, the values for window and level found in the DICOM header were often not appropriate for key image selection, especially for MR images. To resolve these window and level issues, two adjustments were implemented.
If the Modality value of an image transferred to RadICS is "MR," an automatic calculation of window and level settings is made for each image transferred to RadICS. This is done by creating a brightness histogram from the DICOM pixel data. The median brightness is calculated from a portion of gray-scale values and is defined as the level. The top and bottom 15% of the histogram are excluded in this calculation to reduce the effect that large spikes corresponding to near-black or near-white pixels surrounding the anatomic location have on the median. The standard deviation of brightness is also calculated, and four times this standard deviation is defined as the window for images with total contrast resolution greater than 1000 gray-scale values. For images with less than 1000 gray-scale values, six times the standard deviation is used. The algorithm and values for the cutoffs and standard deviation multipliers were determined empirically.
Window and level settings for non-MR images are determined within the Perl script. Preset window and level settings are applied to computed tomographic (CT) images based on anatomic locations: 100 and 50 for brain CT, 400 and 50 for body CT, 4000 and 400 for bone CT, and 1500 and –500 for lung CT (16). For computed radiography images, the level is preset at 1500, but the window setting is determined by the image BitsStored value: 3000 when BitsStored is 10 and 4000 otherwise. The computed radiography values and thresholds were determined empirically. For images from other modalities, the BitsStored value is used to calculate the total number of possible pixel values. The BitsStored DICOM element represents the number of bits stored for each pixel sample. From the number of bits per pixel, the total number of possible gray-scale values can be determined. A BitsStored value of 8 implies there are 256 possible values, while a BitsStored value of 12 implies there are 4096 possible values. This total range of pixel values becomes the window, and half the total range becomes the level.
The Perl script passes the window and level settings as parameters to the DCMTK, which generates full-size and thumbnail JPEG files. The spatial resolution of the full-size JPEG files is equal to the Row and Column values of the original DICOM image; however, the thumbnail JPEG images are proportionally constrained to a resolution of 128 x 128 pixels. For CT images, full-size and thumbnail JPEG image files are generated for all four anatomic locations: brain, body, bone, and lung. Color medical images do not require window and level settings for generation of JPEG files (16). If the PhotometricInterpretation DICOM element contains a value of RGB (red-green-blue), the DICOM file contains a color medical image and the window and level settings are ignored.
Finally, the original DICOM and newly generated JPEG files are moved from the incoming directory to temporary file storage on the server and are available for use by the Web server. Images in studies not used in interesting cases are automatically deleted from the temporary storage directory 1 month after they were received from the PACS.
|
Web Server
|
|---|
The RadICS Web application was developed by using the Java Server Pages (JSP), hypertext markup language (HTML), and JavaScript and Java programming languages and runs on the RadICS server. It provides the user interface for creating and viewing interesting cases. The radiologist logs into RadICS via an Internet Explorer (Microsoft, Redmond, Wash) Web browser on the control monitor of the PACS workstation where JPEG images from studies recently sent from the PACS are accessible for key image selection (Fig 3). When an image is selected, the original DICOM and associated JPEG files are copied to permanent file storage and specific image information is stored in a MySQL relational database. The image name, path, and associated case identifier are used to correlate images to their appropriate case and to provide pointers to image locations for display. In addition, the window, level, Row, and Column values of the image are stored for later use.
View larger version (38K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. At the PACS workstation, the radiologist opens an Internet Explorer (Microsoft) Web browser and logs into RadICS on the control monitor (left), while the study DICOM images are still viewable on the second PACS monitor (right). Thumbnail JPEG images of the original DICOM images are generated by using the window and level settings obtained from the automatic calculator. These images are displayed in RadICS for key image selection (left).
|
|
Once all key images are selected, the radiologist is presented with a Web form for input of relevant case information, including diagnosis, anatomy, pathologic findings, patient age and gender, clinical history, and case description. The diagnosis, anatomy, and coarse pathologic findings fields are coded data presented in dynamic multitier drop-down selection lists. The diagnosis selection box has the added capability of allowing a new diagnosis to be manually entered. The patient age and gender fields are autopopulated with values captured from the DICOM header. All input case information is stored in the relational MySQL database.
As the automatic calculation of window and level is primarily for key image selection, the radiologist may further optimize image window and level settings and add cropping parameters to further distinguish an area of educational focus through a Java applet for user-defined, interactive optimization for gray-scale images of all modalities. Communication between this applet and Java servlets allows manual adjustment to further distinguish the educational focus or to fine-tune the visual presentation of an image for publication. Although images are presented and can be exported in desktop publishing format, RadICS also stores the original DICOM images. The applet in the Web client viewer user interface actually operates on the DICOM image data stored on the server. Window and level adjustments are always performed on these original DICOM images and then converted into desktop publishing format to avoid serial deterioration. After manual adjustments are made to the window and level, the case creation is complete.
A robust search engine was implemented in the RadICS Web server to provide an efficient tool for locating and viewing interesting cases on the basis of diagnosis, anatomy, pathologic findings, case author, patient gender, patient age range, or keywords found in the title, clinical history, or case description, or any combination thereof. Searches on coded terms as well as free text were allowed.
|
Evaluation of Vendor Neutrality
|
|---|
The vendor neutrality of RadICS was tested against six unmodified commercial PACS at five different institutions. They were selected based on their ability to perform a standard DICOM SEND and that their workstations are able to support Microsoft Internet Explorer (14). The RadICS server was installed on site and connected to the PACS internal network behind the radiology or radiology informatics department firewall. Imaging studies of various sizes and from multiple modalities were sent from the PACS to RadICS. The presence of functional DICOM images on the RadICS Web server signified a successful result. DICOM images were deemed functional if Web-viewable JPEG images were successfully generated.
|
Results
|
|---|
The teaching file system was tested on multiple PACS workstations that supported either direct or indirect DICOM SEND. The teaching file successfully accepted DICOM images from workstations running the following PACS: (a) Rad-works (Marconi, London, England) (tested at University of Utah Health Sciences Center, Salt Lake City); (b) EasyVision (Philips Medical Systems, Andover, Mass) (tested at University of Utah Health Sciences Center) (Fig 4); (c) Stentor 3.2.2 (Stentor, Brisbane, Calif) (tested at University of Pittsburgh Medical Center, Pittsburgh, Pa) (Fig 5); (d) Centricity (GE Healthcare Technologies, Waukesha, Wis) (tested at Northwestern Radiology, Chicago, Ill); (e) Impax 4.0 (Agfa, Ridgefield Park, NJ) (tested at Salt Lake City Veterans Administration Medical Center); and (f) Sectra IDS5 (Sectra Medical Systems, Shelton, Conn) (tested at Utah Center for Advanced Imaging Research, Salt Lake City).
Submissions of functional DICOM images to RadICS were successful regardless of study type. The RadICS server was able to handle CT, MR, computed radiography, and digital radiography images adhering to the DICOM format, regardless of the number of series or the number of images per study. In addition, RadICS handled mammographic, fluoroscopic, nuclear medicine, positron emission tomography, secondary capture, and three-dimensional reconstructed DICOM images sent from the Philips Medical Systems and Sectra PACS (these were not tested with the other PACS). RadICS was successfully developed and implemented and is currently in use in a large academic radiology department (Fig 6).
|
Discussion
|
|---|
Unfortunately, the important educational functionality of DTFs has not been a high priority for PACS vendors (9–12). Therefore, academic institutions around the world struggle to find the ideal digital radiology teaching file solution and work at collaborating teaching and research endeavors. The DTF project at Northwestern University (4) used a vendor function that e-mailed JPEG converted images, which is not optimal work flow. The University of California, San Francisco, DTF project is tightly integrated, but only with vendor cooperation and assistance and works only on that vendor’s PACS (8).
The Radiological Society of North America has developed a standard and created open-source software for a Medical Imaging Resource Center (MIRC) (17,18). The standard defines methods for configuring and managing a MIRC site. It also describes a set of extensible markup language (XML) schemas for case archiving and case sharing among MIRC-compliant DTFs worldwide (19). Although the MIRC standard addresses the authoring and sharing of interesting cases, it has only recently begun addressing the problem of harvesting selected images. The current version of MIRC does not include any automated window and level functionality. RadICS, with its automatic window and level calculation algorithm, could serve as an effective front end to MIRC authoring tools, enabling efficient export of images from PACS. The version of MIRC currently under development will contain the Java applet used in RadICS for manual window and level adjustment.
In April 2005, the Integrating the Healthcare Enterprise (IHE) Radiology Technical Committee introduced the Teaching File and Clinical Trial Export (TCE) profile as an IHE Radiology Technical Framework Supplement (20). The TCE profile proposes guidelines to enable export of select images from PACS to a DTF. It also defines the information exported, the actors involved in the export process, various image processing capabilities, and the methods for the export process. The authors support the IHE TCE profile and call for wider acceptance by PACS vendors. Widespread compliance with the TCE profile will provide all radiologists a method for harvesting key images from PACS into a DTF, including MIRC, without disruption in clinical work flow. However, at this point only a minimal number of PACS are IHE TCE compliant. RadICS provides an efficient intermediary solution that provides IHE TCE profile functionality without relying on vendor compliance. Future work will ensure that RadICS is completely IHE TCE compliant.
The RadICS project was focused on solving the image harvesting problem for any DICOM-compliant PACS, which to date is an unfulfilled objective of MIRC and the IHE TCE profile. RadICS was designed to implement and demonstrate the functionality that should and will exist in MIRC when vendors implement the IHE TCE profile. The work flow of a radiologist is such that interesting cases are encountered during primary interpretation of studies. For efficient collection of education materials, it is essential that the radiologist is able to transfer images from PACS to a DTF when a case is identified (4,21). It is also important that the transfer process be seamless, with minimal effort and with minimal work flow disruption. The key work flow elements are selection of key images, rendering at proper window and level to optimally demonstrate the abnormality, conversion to desktop publishing format (such as JPEG), capture of any available clinical information, categorization by presumed or proved image findings and diagnosis, and transfer to a HIPAA-sterilized repository separate from the clinical archive. To our knowledge, no single system has been described in the literature that accomplishes these tasks for virtually all PACS vendors.
As interesting cases are identified at the PACS workstation, the radiologist uses the PACS to immediately push the study to RadICS. This does not require the physician to change or log in to a different application. As images are transferred and automatically processed by the RadICS server, the radiologist can continue with his or her clinical tasks. At a time convenient to the radiologist, either during the current reading session or later, RadICS can be opened to allow key image selection and case creation. The entire process from DICOM SEND to display of the JPEG images for key image selection is automated and does not significantly interrupt the clinical work flow.
There are two philosophies as to the appropriate time and location for the authoring process. One describes the process of selecting key images at the time of case interpretation; however, the rest of the case authoring occurs at a later time (4). The second and more widely accepted philosophy describes the process of authoring a first-draft teaching file case at the point of interpretation (21). The authors believe that it is most efficient to generate the case, with selected key images, coded diagnosis, and case history stored in a relational database, at the PACS workstation before moving on to another study. At this moment, the clinical history and case information are at the forefront of the radiologist’s mind, as well as familiarity with the image set, and additional related information is still readily available.
The radiologist could decide to immediately log into RadICS and complete the case creation or choose to create the case at a later time or day. By opening the RadICS Web application on the PACS workstation at the beginning of the work day, radiologists can quickly create cases and easily review existing teaching files. In addition, the RadICS Web application is accessible from any personal computer, laptop, or workstation with access behind the department firewall, including the radiologist’s office or home via virtual private network (VPN) access. RadICS is an effective tool, as study images are stored on the RadICS server for 1 month after having been received from PACS and are accessible for key image selection at any time during this period.
Although some PACS allow creation of JPEG files from key images that represent the current presentation state, the process often involves multiple steps exporting images from PACS to a DTF. Functionality varies between PACS, from saving the image to the local workstation to copying the image to the clipboard and pasting into the desired application. In many situations, the creation of these JPEG files is not accomplished in a batch process, so each image must be saved or copied one at a time. Secondary nonclinical applications, including image processing and local teaching file software, are often prohibited from being installed on PACS workstations by vendor regulation or in an attempt to prevent conflict with clinical applications. Moving images from PACS workstations to network directories or separate systems for use in a DTF presents additional steps and may depend on department policies or network infrastructure.
Some PACS do not provide any means for exporting images. In this case, the radiologist may be able to use the Print Screen function on the keyboard to copy the visible desktop contents to the clipboard and then paste them into another application (4). This is also a nonbatch process where each image would be copied separately and could also be limited by the software allowed to be installed on the PACS workstations. Although these manual methods of image capture may be effective for harvesting a few desired images, they are not ideal for harvesting images in a DTF. In RadICS, all of the initial image processing and management is handled automatically and the original DICOM image file is archived on the server for future real-time optimization of the JPEG image.
|
Conclusions
|
|---|
RadICS provides an effective solution for automating the harvesting and export process of images from a PACS. The RadICS system can be used with any DICOM-compliant PACS without requiring system modifications and allows case creation with minimal radiologist work flow disruption.
|
TAKE-HOME POINTS
|
|---|
- RadICS is a solution to the problem of efficiently harvesting key images from an interesting study in any DICOM–compliant PACS with minimal radiologist work flow disruption.
- To achieve vendor neutrality, RadICS does not query the PACS archive directly. Instead, studies are pushed by the radiologist from a client workstation, such as one in a reading room.
- The authors support the IHE TCE profile and call for wider acceptance by PACS vendors. Widespread compliance with the TCE profile will provide all radiologists a method for harvesting key images from PACS into a DTF, including MIRC, without disruption in clinical work flow.
- The authors believe that it is most efficient to generate the case, with selected key images, coded diagnosis, and case history stored in a relational database, at the PACS workstation before moving on to another study.
- In RadICS, all of the initial image processing and management is handled automatically and the original DICOM image file is archived on the server for future real-time optimization of the JPEG image.
|
Acknowledgments
|
|---|
The authors thank Paul J. Chang, MD, Barton F. Branstetter IV, MD, David S. Channin, MD, Pattanasak Mongkolwat, PhD, and their supportive staffs for assistance with testing RadICS at their respective institutions.
|
Footnotes
|
|---|
Abbreviations: DCMTK = DICOM Toolkit, DICOM = Digital Imaging and Communications in Medicine, DTF = digital teaching file, HIPAA = Health Insurance Portability and Accountability Act, IHE = Integrating the Healthcare Enterprise, JPEG = Joint Photographic Experts Group, MIRC = Medical Imaging Resource Center, PACS = picture archiving and communication system, RadICS = Radiology Interesting Case Server, TCE = Teaching File and Clinical Trial Export
|
References
|
|---|
- Seshadri SB, Arenson RL. The impact of PACS on research and education. Int J Biomed Comput 1992;30(3–4):263–266.[CrossRef][Medline]
- Macura KJ, Macura RT, Morstad BD. Digital case library: a resource for teaching, learning, and diagnosis support in radiology. RadioGraphics 1995;15(1):155–164.[Abstract]
- Scarsbrook AF, Foley PT, Perriss RW, Graham RN. Radiological digital teaching file development: an overview. Clin Radiol 2005;60(8):831–837.[CrossRef][Medline]
- Mongkolwat P, Bhalodia P, Makori A, Gehl JA, Kogan A, Channin DS. Integrating MIRC-compliant semiautomated teaching files into PACS work flow. RadioGraphics 2005;25(2):543–548.[Abstract/Free Full Text]
- Halsted MJ, Moskovitz J, Johnson N, Perry L. A simple method of capturing PACS and other radiographic images for digital teaching files or other image repositories. AJR Am J Roentgenol 2002; 178(4):817–819.[Abstract/Free Full Text]
- Siegel E, Reiner B. Electronic teaching files: 7-year experience using a commercial picture archiving and communication system. J Digit Imaging 2001; 14(2 suppl 1):125–127.[Medline]
- Maldjian JA, Listerud J. Automated teaching file and slide database for digital images. AJR Am J Roentgenol 2000;175(5):1249–1251.[Abstract/Free Full Text]
- Tellis WM, Andriole KP, Avrin DE, Arenson RL. Web technology in the integration of a digital teaching file at the diagnostic workstation. J Digit Imaging 1998;11(3 suppl 1):117–119.[Medline]
- Chronaki CE, Zabulis X, Orphanoudakis SC. I2Cnet medical image annotation service. Med Inform (Lond) 1997;22(4):337–347.[Medline]
- Goldberg DJ, DeMarco JK, Parikh T. Internet-based interactive teaching file for neuroradiology. AJR Am J Roentgenol 2000;175(5):1371–1373.[Abstract/Free Full Text]
- Piraino D, Recht M, Richmond B. Implementation of an electronic teaching file using Web technology. J Digit Imaging 1997;10(3 suppl 1):190–192.[Medline]
- Lim CC, Yang GL, Nowinski WL, Hui F. Medical Image Resource Center: making electronic teaching files from PACS. J Digit Imaging 2003; 16(4):331–336.[CrossRef][Medline]
- Cao F, Huang HK, Zhou XQ. Medical image security in a HIPAA mandated PACS environment. Comput Med Imaging Graph 2003;27(2–3):185–196.[CrossRef][Medline]
- Digital Imaging and Communications in Medicine (DICOM). Rosslyn, Va: National Electrical Manufacturers Association, 2001; PS 3.1-2001: iii–iv, 4–12.
- OFFIS DICOM Software, DCMTK—DICOM Toolkit. http://dicom.offis.de/dcmtk.php.en. Accessed July 1, 2006.
- Dreyer KJ, Mehta A, Thrall JH. PACS: a guide to the digital revolution. New York, NY: Springer, 2002.
- Radiological Society of North America. Medical Imaging Resource Center. http://www.rsna.org/mirc. Accessed February 7, 2006.
- Siegel E, Channin D, Perry J, Carr C, Reiner B. Medical Image Resource Center 2002: an update on the RSNA’s Medical Image Resource Center. J Digit Imaging 2002;15(1):2–4.[CrossRef][Medline]
- Radiological Society of North America. MIRC documentation index. http://mirc.rsna.org/mircstorage/documents/documentation/MIRCdocumentationindex.xml. Accessed February 7, 2006.
- Committee IRT. Teaching File and Clinical Trial Export (TCE). http://www.ihe.net/Technical_Framework/upload/IHE_TF_Suppl_Teaching_File_Clinical_Trial_Export_TI_2005-04-22.pdf. Accessed February 1, 2006.
- Avrin DE, Andriole KP, Arenson RL. Integration of a digital teaching file at the diagnostic workstation. In: Kilcoyne RF, Lear JL, Rowberg AH, eds. Proceedings of the Annual Meeting of the Society for Computer Applications in Radiology. Lees-burg, Va: Society for Computer Applications in Radiology, 1996; 56–61.
This article has been cited by other articles:
|
|
|
|
 
A. W. C. Kamauu, J. J. Whipple, S. L DuVall, K. M. Siddiqui, E. L. Siegel, and D. Avrin
Informatics in Radiology: IHE Teaching File and Clinical Trial Export Integration Profile: Functional Examples
RadioGraphics,
July 1, 2008;
28(4):
933 - 945.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C.-S. Yam
An Alternative for Presenting Interactive Dynamic Data Sets in Electronic Presentations: A Scrollable Flash Movie Loop
Am. J. Roentgenol.,
November 1, 2007;
189(5):
W295 - W300.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
