The Society for Computer Applications in Radiology (SCAR) Transforming the Radiological Interpretation Process (TRIP) Initiative aims to spearhead research, education, and discovery of innovative solutions to address the problem of information and image data overload. The initiative will foster inter-disciplinary research on technological, environmental and human factors to better manage and exploit the massive amounts of data. TRIP will focus on the following basic objectives: improving the efficiency of interpretation of large data sets, improving the timeliness and effectiveness of communication, and decreasing medical errors. The ultimate goal of the initiative is to improve the quality and safety of patient care. Interdisciplinary research into several broad areas will be necessary to make progress in managing the ever-increasing volume of data. The six concepts involved include: human perception, image processing and computer-aided detection (CAD), visualization, navigation and usability, databases and integration, and evaluation and validation of methods and performance. The result of this transformation will affect several key processes in radiology, including image interpretation; communication of imaging results; workflow and efficiency within the health care enterprise; diagnostic accuracy and a reduction in medical errors; and, ultimately, the overall quality of care.

The radiology diagnostic reporting is a process that results in generating a diagnostic report to be made available outside the radiology department. The report captures the radiologist’s interpretations and impressions. It is an element of the patient healthcare record and represents important clinical information to assist in healthcare decisions. The reporting process is initiated by the existence of images or other radiology evidences to be interpreted. The work of individuals is controlled by systems that manage workflow. These systems may introduce delays or constraints on how and when tasks are performed. In order to design and implement efficient information systems that manage the reporting workflow, an accurate workflow modeling is needed. Workflow modeling consists in describing what is done by whom and in what sequence, that is the roles, tasks and sequences of tasks. The workflow model is very important and has major consequences. An inaccurate model introduces inefficiencies, frustrations and may result in a useless information system. In this paper, we will model several common reporting workflows by describing the roles, tasks and information flows involved.

As radiologists progress from reading images presented on film to modern computer systems with images presented on high-resolution displays, many new problems arise. Although the digital medium has many advantages, the radiologist’s job becomes cluttered with many new tasks related to image manipulation. This paper presents our solution for supporting radiologists’ interpretation of digital images by automating image presentation during sequential interpretation steps. Our method supports scenario based interpretation, which group data temporally, according to the mental paradigm of the physician. We extended current hanging protocols with support for “stages”. A stage reflects the presentation of digital information required to complete a single step within a complex task. We demonstrated the benefits of staging in a user study with 20 lay subjects involved in a visual conjunctive search for targets, similar to a radiology task of identifying anatomical abnormalities. We designed a task and a set of stimuli which allowed us to simulate the interpretation workflow from a typical radiology scenario - reading a chest computed radiography exam when a prior study is also available. The simulation was possible by abstracting the radiologist’s task and the basic workstation navigation functionality. We introduced “Stages,” an interaction technique attuned to the radiologist’s interpretation task. Compared to the traditional user interface, Stages generated a 14% reduction in the average interpretation.

The goal of the current study is to describe the experience of the Heart Institute (InCor) in the implementation of a patient-oriented Hospital Information System (HIS) integrated with the Radiology Information System (RIS) and the Picture Archiving and Communication System (PACS) in an open-source three-tier architecture. The system was designed in modules that permits patient admission, discharge and transfer (ADT), registration of medical activities, registration of diagnoses and therapy, order entry and access of all patient data, including vital signals, images and lab tests. The modules are integrated in a single Web-based application allowing easy and fast navigation through the application. In order to provide high quality of patient care in an efficient and cost-effective manner, thin clients workstations in a Linux environment were used. To access the patient information users have to perform an authentication procedure that uses LDAP protocol, which also defines a profile to the users. The system is fully integrated to the InCor's PACS, allowing instant access to the image database from applications that requires this information, such as diagnostic reports. For displaying the images a Java DICOM viewer was implemented. On the server side, a Java DICOM server was designed to allow communication with all DICOM modalities.

The JPEG2000 image compression standard provides many features to support interactive access to large images. These include efficient lossless and lossy compression, resolution scalability, quality scalability, region of interest coding and spatial random access. The ISO/IEC JPEG committee is in the process of finalizing the JPEG2000 Internet Protocol (JPIP) standard for interacting with JPEG2000 files. JPIP is tightly coupled to JPEG2000 and provides a powerful and flexible client-server architecture. By combining JPIP and JPEG2000 it is possible to store only one compressed file at the server and transmit the resolution, quality, and Window Of Interest (WOI) specified by the client, without having to transmit or decode the entire code-stream. This minimizes server computation, storage and bandwidth. It also reduces the latency for the client. By sending only the data that the client needs the effective compression ratio for client-server applications can be much higher than the file size compression ratio. A prototype JPIP client/server implementation is used to demonstrate resolution scalability, quality scalability and spatial random access of medical images. Efficiency, latency improvements and PSNR quality of the received images are measured as a function of the number of bytes transferred from the server to the client. The choice of JPEG2000 compression parameters, such as precincts and tiles, and the trade-offs of stateful and stateless sessions are discussed. The inclusion of radiology reports as metadata in image files is also demonstrated.

Although our hospital has an electronic order entry system, physician order entries must be performed using one of a limited number of terminals which are located in places like nurse stations. Due to this limitation, the mobility of the ultrasound machine cannot be fully exploited if a physician wants to perform a mobile ultrasound examination far from an entry terminal. In addition, this situation increases the possibility that a physician will fail to issue a paper voucher of the examination. In an attempt to resolve these problems, we have developed a mobile physician order entry system that incorporates ultrasound machines equipped with handheld wireless terminals, which use the 802.11b standard and the bandwidth is 11 Mbps. This is an efficient way to order ultrasound examination entries because physicians can register entries at any location in a hospital ward. In addition, the proposed system is a reliable method by which to attach images upon ultrasound examination entry.

The failures we have all experienced in attempts to fully integrate systems have long been a significant stumbling block to reaching the levels of department efficiency and convenience that we know are needed, desired and theoretically achievable in a modern Healthcare Enterprise. The problem is not a lack of Standards, but rather a lack of consistency and conformity in the way vendors apply the prevalent standards to solving specific problems. The Integrating the Healthcare Enterprise (IHE) Initiative has brought together key user and vendor representatives to document detailed technical descriptions, referred to as “Integration Profiles”, of how each system should participate in solutions to specific scenarios based on existing standards. This is followed up by intensive annual cross-vendor testing events called “Connectathons.” Now in its 6th year of development, the scope of IHE has expanded beyond its beginnings in Radiology to span additional domains including IT Infrastructure, Laboratory and Cardiology. IHE Connectathons are now being held in North America, Europe and Asia. IHE has proven itself to be an effective way to get vendors to cooperatively address the integration capabilities needed by healthcare facilities to reap the benefits of modern information technologies. The existing Integration Profiles, what they address, and an overview of current and upcoming IHE activities will be presented.

In 2003, the DICOM standard celebrated its 10th anniversary. Aside from the standard itself, also OFFIS’ open source DICOM toolkit DCMTK, which has continuously followed the development of DICOM, turns 10 years old. On this occasion, this article looks back at the main standardization efforts in DICOM and illustrates related developments in DCMTK. Considering the development of the DICOM standard, it is possible to distinguish several phases of progress. Within the first phase at the beginning of the 1990s, basic network services for image transfer and retrieval were being introduced. The second phase, in the mid 1990s, was characterized by advances in the specification of a file format and of regulations for media interchange. In the later but partly parallel third phase, DICOM predominantly dealt with the problem of optimizing the workflow within imaging departments. As a result of the fact that it was now possible to exchange images between different systems, efforts concerning image display consistency followed in a fourth phase at the end of the 1990s. In the current fifth phase, security enhancements are being integrated into the standard. In another phase of progress, which took place over a relatively long time period concurrently to the other mentioned phases, DICOM Structured Reporting was developed.

A number of clinical specialties routinely use images in treating patients, for example ophthalmology, dentistry, cardiology, endoscopy, and surgery. These images are captured by a variety of commercial digital image acquisition systems. The US Department of Veterans Affairs has been working for several years on advancing the use of the Digital Imaging and Communications in Medicine (DICOM) Standard in these clinical specialties. This is an effort that has involved several facets: (1) working with the vendors to ensure that they satisfy existing DICOM requirements, (2) developing interface software to the VistA hospital information system (HIS), (3) field testing DICOM systems, (4) deploying these DICOM interfaces nation-wide to all VA medical centers, (5) working with the healthcare providers using the system, and (6) participating in the DICOM working groups to improve the standard. The VA is now beginning to develop clinical applications that make use of the DICOM interfaces in the clinical specialties. The first of these will be in ophthalmology to remotely screen patients for diabetic retinopathy.

The DICOM standard, as all standards, specifies in generic way the management in network and storage media environments of digital medical images and their related information. However, understanding the specifications for particular implementation is not a trivial work. Thus, this work is about understanding and modelling parts of the DICOM standard using Object Oriented methodologies, as part of software development processes. This has offered different static and dynamic views, according with the standard specifications, and the resultant models have been represented through the Unified Modelling Language (UML). The modelled parts are related to network conformance claim: Network Communication Support for Message Exchange, Message Exchange, Information Object Definitions, Service Class Specifications, Data Structures and Encoding, and Data Dictionary. The resultant models have given a better understanding about DICOM parts and have opened the possibility of create a software library to develop DICOM conformable PACS applications.

The differentiating characteristics of text versus images and their impact on large medical image databases intended to allow content-based indexing and retrieval have recently been explored. For the design of powerful user interfaces, we propose a grouping of the various mechanisms into four classes: (i) output modules, (ii) parameter modules, (iii) transaction modules, and (iv) process modules, all of which being controlled by a detailed query logging. Relevance feedback by the physician loops the input, search, and output sequence and is commonly accepted to be most effective for query refinement. Our modular concept provides two additional loops of interaction. Based on the detailed logging of user interaction, an inner loop allows to step back to any previous answer that was given by the system during the interactive session. Boolean linkage of successive queries is provided by an outer loop. Nonetheless, the entire data flow is still controlled within a single web page by means of simple decision rules that are implemented only using push buttons, which can be handled most easily. Our approach is exemplified by means of an application for content-based access to medical images of similar modality, orientation, and body region using global features that model gray scale, texture, structure and global shape characteristics. The three nested loops for interaction provide a maximum of flexibility within a minimum of complexity. The resulting extended query refinement has a significant impact for content-based retrieval in medical applications (IRMA).

Content-based image retrieval systems (CBIRSs) have frequently been proposed for the use in medical image databases and PACS. Still, only few systems were developed and used in a real clinical environment. It rather seems that medical professionals define their needs and computer scientists develop systems based on data sets they receive with little or no interaction between the two groups. A first study on the diagnostic use of medical image retrieval also shows an improvement in diagnostics when using CBIRSs which underlines the potential importance of this technique. This article explains the use of an open source image retrieval system (GIFT - GNU Image Finding Tool) for the retrieval of medical images in the medical case database system CasImage that is used in daily, clinical routine in the university hospitals of Geneva. Although the base system of GIFT shows an unsatisfactory performance, already little changes in the feature space show to significantly improve the retrieval results. The performance of variations in feature space with respect to color (gray level) quantizations and changes in texture analysis (Gabor filters) is compared. Whereas stock photography relies mainly on colors for retrieval, medical images need a large number of gray levels for successful retrieval, especially when executing feedback queries. The results also show that a too fine granularity in the gray levels lowers the retrieval quality, especially with single-image queries. For the evaluation of the retrieval peformance, a subset of the entire case database of more than 40,000 images is taken with a total of 3752 images. Ground truth was generated by a user who defined the expected query result of a perfect system by selecting images relevant to a given query image. The results show that a smaller number of gray levels (32 - 64) leads to a better retrieval performance, especially when using relevance feedback. The use of more scales and directions for the Gabor filters in the texture analysis also leads to improved results but response time is going up equally due to the larger feature space. CBIRSs can be of great use in managing large medical image databases. They allow to find images that might otherwise be lost for research and publications. They also give students students the possibility to navigate within large image repositories. In the future, CBIR might also become more important in case-based reasoning and evidence-based medicine to support the diagnostics because first studies show good results.

Rheumatoid Arthritis (RA) is a common systemic disease predominantly involving the joints. Precise diagnosis and follow-up therapy requires objective quantification. For this purpose, radiological analyses using standardized scoring systems are considered to be the most appropriate method. The aim of our study is to develop a semi-automatic image analysis software, especially applicable for scoring of joints in rheumatic disorders. The X-Ray RheumaCoach software delivers various scoring systems (Larsen-Score and Ratingen-Rau-Score) which can be applied by the scorer. In addition to the qualitative assessment of joints performed by the radiologist, a semi-automatic image analysis for joint detection and measurements of bone diameters and swollen tissue supports the image assessment process. More than 3000 radiographs from hands and feet of more than 200 RA patients were collected, analyzed, and statistically evaluated. Radiographs were quantified using conventional paper-based Larsen score and the X-Ray RheumaCoach software. The use of the software shortened the scoring time by about 25 percent and reduced the rate of erroneous scorings in all our studies. Compared to paper-based scoring methods, the X-Ray RheumaCoach software offers several advantages: (i) Structured data analysis and input that minimizes variance by standardization, (ii) faster and more precise calculation of sum scores and indices, (iii) permanent data storing and fast access to the software’s database, (iv) the possibility of cross-calculation to other scores, (v) semi-automatic assessment of images, and (vii) reliable documentation of results in the form of graphical printouts.

Bone age assessment is a procedure frequently performed in pediatric patients to evaluate their growth disorder. A commonly used method is by atlas matching with a small reference set of Greulich-Pyle atlas that was developed in 1950s based on middle class white populations. The old Greulich-Pyle atlas is not fully applicable for today’s children especially regarding the standard development in other racial groups. We have collected over 1000 hand and wrist images for normal boys and girls of Caucasian, African, Hispanic and Asian descents from newborn to 18 years of age. In this paper we present an atlasing method that will automatically select the reference images from the normal collection database to form a digital atlas. Compared with the book-based Greulich-Pyle atlas, the digital atlas is easy to access and can always be replenished with new images reflecting more accurately the current population and skeletal development.

An efficient database is an essential component of organizing diverse information on image metadata and patient information for research in medical imaging. This paper describes the design, development and deployment of a large database system serving as a brain image repository that can be used across different platforms in various medical researches. It forms the infrastructure that links hospitals and institutions together and shares data among them. The database contains patient-, pathology-, image-, research- and management-specific data. The functionalities of the database system include image uploading, storage, indexing, downloading and sharing as well as database querying and management with security and data anonymization concerns well taken care of. The structure of database is multi-tier client-server architecture with Relational Database Management System, Security Layer, Application Layer and User Interface. Image source adapter has been developed to handle most of the popular image formats. The database has a user interface based on web browsers and is easy to handle. We have used Java programming language for its platform independency and vast function libraries. The brain image database can sort data according to clinically relevant information. This can be effectively used in research from the clinicians’ points of view. The database is suitable for validation of algorithms on large population of cases. Medical images for processing could be identified and organized based on information in image metadata. Clinical research in various pathologies can thus be performed with greater efficiency and large image repositories can be managed more effectively. The prototype of the system has been installed in a few hospitals and is working to the satisfaction of the clinicians.

Severe acute respiratory syndrome (SARS) is a respiratory illness that had been reported in Asia, North America, and Europe in last spring. Most of the China cases of SARS have occurred by infection in hospitals or among travelers. To protect the physicians, experts and nurses from the SARS during the diagnosis and treatment procedures, the infection control mechanisms were built in SARS hospitals. We built a Web-based interactive teleradiology system to assist the radiologists and physicians both in side and out side control area to make image diagnosis. The system consists of three major components: DICOM gateway (GW), Web-based image repository server (Server), and Web-based DICOM viewer (Viewer). This system was installed and integrated with CR, CT and the hospital information system (HIS) in Shanghai Xinhua hospital to provide image-based ePR functions for SARS consultation between the radiologists, physicians and experts inside and out side control area. The both users inside and out side the control area can use the system to process and manipulate the DICOM images interactively, and the system provide the remote control mechanism to synchronize their operations on images and display.

Presently most Nuclear Medicine physicians are well trained to report PET FDG studies. However, only a very limited number of them are able to diagnose difficult, unusual cases. For this reason, we developed an electronic lightbox called POSITOSCOPE onto which PET studies can be downloaded, displayed, reported and sent to remote sites for expert advice. To promote its use, we emphasized user-friendliness which is a keypoint of the prototype: the POSITOSCOPE looks like a classical lightbox equipped with a small touchscreen and a digital sound recorder. It is connected to local PET scanners and long distance high speed networks. Difficult studies can thus be sent to remote experts. The request consists of the whole image data set and a soundtrack explaining its nature. It may be sent to one or more experts. At this stage, only the local physician is responsible for reporting even though (s)he makes use of remote expertise. The prototype is being tested in two hospitals and the clinical evaluation involving four University hospitals and one private practice Nuclear Medicine center, started last September. Our goal is not to have PET studies acquired in a local center and to have them reported in a remote reference center, but to provide remote expertise when necessary to improve daily reporting of PET studies and to improve the expertise of local Nuclear Medicine physicians. The concept may be easily extended to unusual single photon studies for which local expertise is not always available, and to multimodality studies.

We present a scheme for compressed domain interactive rendering of large volume data sets over distributed environments. The scheme exploits the distortion scalability and multi-resolution properties offered by JPEG2000 to provide a unified framework for interactive rendering over low bandwidth networks. The interactive client is provided breadth in terms of scalability in resolution, position and progressive improvement by quality. The server exploits the spatial locality offered by the DWT and packet indexing information to transmit, in so far as possible, compressed volume data relevant to the clients query. Once the client identifies its volume of interest (VOI), the volume is refined progressively within the VOI. Contextual background information can also be made available having quality fading away from the VOI. The scheme is ideally suited for client-server setups with low bandwidth constraints, with the server maintaining the compressed volume data, to be browsed by a client with low processing power and/or memory. Rendering can be performed at a stage when the client feels that the desired quality threshold has been attained. We investigate the effects of code-block size on compression ratio, PSNR, decoding times and data transmission to arrive at an optimal code-block size for typical VOI decoding scenarios.

The existing image compression standards like JPEG and JPEG 2000, compress the whole image as a single frame. This makes the system simple but inefficient. The problem is acute for applications where lossless compression is mandatory viz. medical image compression. If the spatial characteristics of the image are considered, it can give rise to a more efficient coding scheme. For example, CT reconstructed images have uniform background outside the field of view (FOV). Even the portion within the FOV can be divided as anatomically relevant and irrelevant parts. They have distinctly different statistics. Hence coding them separately will result in more efficient compression. Segmentation is done based on thresholding and shape information is stored using 8-connected differential chain code. Simple 1-D DPCM is used as the prediction scheme. The experiments show that the 1st order entropies of images fall by more than 11% when each segment is coded separately. For simplicity and speed of decoding Huffman code is chosen for entropy coding. Segment based coding will have an overhead of one table per segment but the overhead is minimal. Lossless compression of image based on segmentation resulted in reduction of bit rate by 7%-9% compared to lossless compression of whole image as a single frame by the same prediction coder. Segmentation based scheme also has the advantage of natural ROI based progressive decoding. If it is allowed to delete the diagnostically irrelevant portions, the bit budget can go down as much as 40%. This concept can be extended to other modalities.

Cathode-ray tube (CRT) and liquid crystal display (LCD) are currently two main technologies for displaying medical images. LCDs possess a number of advantages, but their performance varies as a function of viewing angle. The purpose of this study was to delineate the impact of the angular response of LCDs on their DICOM grayscale display function (GSDF) compliance, and to develop a framework to define an angular acceptance for medical LCDs. Measurements were made on a calibrated dual-domain LCD (IBM T221). The on-axis luminance values were measured at all 8-bit driving levels three times using the TG18-LN test patterns and a baffled luminance meter and the results averaged. The luminance was also measured as a function of viewing angle at 17 discrete levels using a Fourier-optics-based luminance meter. The luminance data were analyzed according to the AAPM TG18 methodology. The on-axis results showed close conformance with the TG18 criteria with L_min, L_max, mean ▵JND/▵p, and maximum local deviation in ▵JND/▵p from GSDF, &kappa²⁵⁶, of 0.83 cd/m², 263 cd/m², 2.1, and 0.8, respectively. However, the values varied notably as a function of viewing angle. Overall, the luminance ratio remained greater than 175 within a ±20° viewing angle cone (β₁₇₅ = ±20°). Aiming to maintain κ¹⁷≥0.3, an acceptable viewing angle cone of ±35° was indicated (α_0.3 = ±35°). The findings demonstrate the significant impact of angular response on image contrast, and the utility of α_0.3 and β₁₇₅ quantities for defining the viewing angle cones within which a medical LCD device can be effectively utilized.

Last year we presented in RSNA an application to perform wireless remote control of PACS image distribution utilizing a handheld device such as a Personal Digital Assistant (PDA). This paper describes the clinical experiences including workflow scenarios of implementing the PDA application to route exams from the clinical PACS archive server to various locations for offsite distribution of clinical PACS exams. By utilizing this remote control application, radiologists can manage image workflow distribution with a single wireless handheld device without impacting their clinical workflow on diagnostic PACS workstations. A PDA application was designed and developed to perform DICOM Query and C-Move requests by a physician from a clinical PACS Archive to a CD-burning device for automatic burning of PACS data for the distribution to offsite. In addition, it was also used for convenient routing of historical PACS exams to the local web server, local workstations, and teleradiology systems. The application was evaluated by radiologists as well as other clinical staff who need to distribute PACS exams to offsite referring physician’s offices and offsite radiologists. An application for image workflow management utilizing wireless technology was implemented in a clinical environment and evaluated. A PDA application was successfully utilized to perform DICOM Query and C-Move requests from the clinical PACS archive to various offsite exam distribution devices. Clinical staff can utilize the PDA to manage image workflow and PACS exam distribution conveniently for offsite consultations by referring physicians and radiologists. This solution allows the radiologist to expand their effectiveness in health care delivery both within the radiology department as well as offisite by improving their clinical workflow.

In large institution PACS, patient data may often reside in multiple separate systems. While most systems tend to be DICOM compliant, none of them offer the flexibility of seamless integration of multiple DICOM sources through a single access point. We developed a generic portal system with a web-based interactive front-end as well as an application programming interface (API) that allows both web users and client applications to query and retrieve image data from multiple DICOM sources. A set of software tools was developed to allow accessing several DICOM archives through a single point of access. An interactive web-based front-end allows user to search image data seamlessly from the different archives and display the results or route the image data to another DICOM compliant destination. An XML-based API allows other software programs to easily benefit from this portal to query and retrieve image data as well. Various techniques are employed to minimize the performance overhead inherent in the DICOM. The system is integrated with a hospital-wide HIPAA-compliant authentication and auditing service that provides centralized management of access to patient medical records. The system is provided under open source free licensing and developed using open-source components (Apache Tomcat for web server, MySQL for database, OJB for object/relational data mapping etc.). The portal paradigm offers a convenient and effective solution for accessing multiple image data sources in a given healthcare enterprise and can easily be extended to multi-institution through appropriate security and encryption mechanisms.

Public databases today can be constructed with a wide variety of authoring and management structures. The widespread appeal of Internet search engines suggests that public information be made open and available to common search strategies, making accessible information that would otherwise be hidden by the infrastructure and software interfaces of a traditional database management system. We present the construction and organizational details for managing NOVA, the National Online Volumetric Archive. As an archival effort of the Visible Human Project for supporting medical visualization research, archiving 3D multimodal radiological teaching files, and enhancing medical education with volumetric data, our overall database structure is simplified; archives grow by accruing information, but seldom have to modify, delete, or overwrite stored records. NOVA is being constructed and populated so that it is transparent to the Internet; that is, much of its internal structure is mirrored in HTML allowing internet search engines to investigate, catalog, and link directly to the deep relational structure of the collection index. The key organizational concept for NOVA is the Image Content Group (ICG), an indexing strategy for cataloging incoming data as a set structure rather than by keyword management. These groups are managed through a series of XML files and authoring scripts. We cover the motivation for Image Content Groups, their overall construction, authorship, and management in XML, and the pilot results for creating public data repositories using this strategy.

Purpose: To present a multimedia presentation platform that allows retrieving data from any digital or analog imaging modality. The system includes an authoring environment allowing users to select key images and cine sequences and to prepare a script of a presentation stored in an XML file. Methods: The software was initially implemented for multidisciplinary cardiac conferences involving different cardiologists and cardiovascular surgeons. A typical presentation requires concise presentations of patient data and images and summary reports of data obtained from the different procedures. An XML-based scripting methodology was developed to allow for preparation of these presentations. The image display program uses the generated script for the sequential presentation of different documents in a pre-determined presentation settings. Results: A workstation equipped with the software platform was tested in an integrated electronic setting where selected images and data are gathered from different sources and transferred to the workstation. The XML script of a given presentation can further be saved as part of the patient record for subsequent review of the documents and images that supported a given medical or therapeutic decision. Discussion and conclusion: The script-based presentation mode allows for shorter and more effective clinical presentations that are much less time consuming than traditional conferences where analog data must be reviewed on different devices. This also constitutes a perfect documentation method for surgeons and physicians responsible of therapeutic procedures that were decided upon during the clinical conference. It allows them to review the relevant data that supported a given therapeutic decision.

We present an evaluation of methods for the automatic categorization of medical images. The properties of medical images render some otherwise very successful discriminate features for images (e.g. color) inapplicable. Therefore, we evaluate several feature types: texture, structure, and down-scaled representations. The classification is done using a nearest neighbor classifier with various distance measures as well as the automatic combination of classifier results. A corpus of 6,335 images selected arbitrarily from the clinical routine was encoded using a multi-axial, mono-hierarchical code. The reference categorization was done by experienced radiologists familiar with the code. The code's hierarchy allows the analysis of the automatic categorization performance (depending on the features and the classifier used) at different levels of differentiation. Experiments were done for 54 and 57 categories or 70 and 81 categories focussing on radiographs only or for all images, respectively. A maximum classification accuracy of 86% was obtained using the winner-takes-all rule and a one nearest neighbor classifier. Accuracy is increased to 93% and 95% if the correct category is only required to be within the 5 or 10 best matches, respectively. In this case, the best rate of 98% is obtained. This is sufficient for most applications in content-based image retrieval.

Acceptance Testing (AT) is a crucial step in the implementation process of a PACS within a clinical environment. AT determines whether the PACS is ready for clinical use and marks the official sign off of the PACS product. Most PACS vendors have Acceptance Testing (AT) plans, however, these plans do not provide a complete and robust evaluation of the full system. In addition, different sites will have different special requirements that vendor AT plans do not cover. The purpose of this paper is to introduce a protocol for AT design and present case studies of AT performed on clinical PACS. A methodology is presented that includes identifying testing components within PACS, quality assurance for both functionality and performance, and technical testing focusing on key single points-of-failure within the PACS product. Tools and resources that provide assistance in performing AT are discussed. In addition, implementation of the AT within the clinical environment and the overall implementation timeline of the PACS process are presented. Finally, case studies of actual AT of clinical PACS performed in the healthcare environment will be reviewed. The methodology for designing and implementing a robust AT plan for PACS was documented and has been used in PACS acceptance tests in several sites. This methodology can be applied to any PACS and can be used as a validation for the PACS product being acquired by radiology departments and hospitals. A methodology for AT design and implementation was presented that can be applied to future PACS installations. A robust AT plan for a PACS installation can increase both the utilization and satisfaction of a successful implementation of a PACS product that benefits both vendor and customer.

This work focuses on a general framework for image representation and image matching that may be appropriate for medical image archives. The proposed methodology is comprised of a continuous and probabilistic image representation scheme using Gaussian mixture modeling (GMM) along with information-theoretic image matching measures (KL). The GMM-KL framework is used for matching and categorizing x-ray images by body regions and orientation. A 4-dimensional feature space is used to represent the x-ray image input, including intensity, texture (contrast) and spatial information (x,y). Unsupervised clustering via the GMM is used to extract coherent regions in feature space, and corresponding coherent segments (“blobs”) in the image content. The blobs are used in the matching process. A dominant characteristic of the radiological images is their poor contrast and large intensity variations. This presents a challenge to matching between the images and is handled via a post-processing stage that provides an invariant blob-signature per image input. In a leave-one-out procedure, each image out of 851 is used once as a test-image, and is categorized by the remaining (labeled) images. The GMM-KL classifier was tested using 851 radiological images with error-rate of 1%. The classification results compare favorably with reported global representation schemes, such as histograms.

The sheer amount of digital data generated by the proliferation of filmless medical imaging, poses great scalability and manageability challenges to PACS systems. Manageability challenges are aggravated when weighing legislative requirements. An architecture for an enterprise level PACS should support the management of assorted medical objects (e.g., images and reports). Additionally, the architecture should allow services, including performance and reliability, to be tailored to classes of objects according to complex and possibly varying rules. The design should be flexible, allowing for on-demand cost-effective scaling, using a mix-and-match selection of hardware, operating systems, and storage devices. In light of the increased reliance on stored data, it should ensure 24x7 availability, even during system upgrade, and allow pluggable support for future formats. The Medical Object Management System (MOMS) presented in this paper, is an enterprise medical imaging solution architectured to meet the above demands. Flexible, configurable and scalable content and source based management of objects enables administrators to define and modify policies that govern various aspects of the objects' life-cycles, using either configuration files or a Web-based GUI. The modular architecture of MOMS includes (possibly multiple) instances of interface (DICOM, HL7 and Tivoli Storage Manager), storage management and administration agents. Agent instances are hot-pluggable, allowing for zero-downtime upgrades, and can be deployed on a heterogeneous and distributed infrastructure. Leveraging the expertise gained in the development and deployment of the IDMR research PACS project, combined with recent technological advances and modern middleware, MOMS delivers a solution for the present and future requirements of medical objects management.

Bone age assessment is a procedure performed in pediatric patients to quickly evaluate parameters of maturation and growth from a left hand and wrist radiograph. Pietka and Cao have developed a Computer-aided diagnosis (CAD) method of bone age assessment based on a digital hand atlas. The aim of this paper is to extend their work by automatically select the best representative image from a group of normal children based on specific bony features that reflect skeletal maturity. The group can be of any ethnic origin and gender from one year to 18 year old in the digital atlas. This best representative image is defined as the "average" image of the group that can be augmented to Piekta and Cao's method to facilitate in the bone age assessment process.

Medical image databases are growing at a rapid rate because of the increase in digital medical imaging modalities and the deployment of Picture Archiving and Communication Systems (PACS), Electronic Medical Records (EMR) and telemedicine applications. There is growing research interest in Content-Based Image Retrieval (CBIR) of medical images from such digital archives. A new distance function for CBIR is presented for measuring the similarity between two images. The distance function is a variant of relative entropy, or the Kullback-Liebler distance. The new distance is the sum of the relative entropy of the two images to each other. The latter is a symmetric non-negative function and is only zero when the two images have identical probability distributions. This method has been implemented in a prototype system and has been applied to a database of medical images. Initial results demonstrate improvements over L₁-norm and L₂-norm histogram matching. The method is computationally simple since it does not require image segmentation. It is invariant to translation, rotation and scaling. The method has also been extended to support retrieval based on Region-Of-Interest (ROI) queries.

Medical records are written mainly, in natural language. The focus of this study is narrative radiological reports written in natural Japanese. These reports cannot be used for advanced retrieval, data mining, and so on, unless they are stored, using a structured format such as DICOM-SR. The goal is to structure narrative reports progressively, using natural language processing (NLP). Structure has many different levels, for example, DICOM-SR has three established levels -- basic text, enhanced and comprehensive. At the enhanced level, it is necessary to use numerical measurements and spatial & temporal coordinates. In this study, the wording used in the reports was first standardized, dictionaries were organized, and morphological analysis performed. Next, numerical measurements and temporal coordinates were extracted, and the objects to which they referred, analyzed. 10,000 CT and MR reports were separated into 82,122 sentences, and 34,269 of the 36,444 numerical descriptions were tagged. Periods, slashes, hyphens, and parentheses are ambiguously used in the description of enumerated lists, dates, image numbers, and anatomical names, as well as at the end of sentences; to resolve this ambiguity, descriptions were processed, according to the order -- date, size, unit, enumerated list, and abbreviation -- then, the tagged reports were separated into sentences.

This paper presents a Cardiology oriented information system that provides permanent availability of all clinical history, including alphanumeric and image data, with time and cost-effective transmission (reduced download time), without loss of image diagnosis quality and based on a Web Multimedia Integrated Access Interface. This implies the integration of HIS and PACS in a unique access interface, providing on-line and fast access to authorized healthcare professionals. The benefits obtained from the HIS-PACS integration and from the availability of all historical patient data are unquestionable to practitioners but also to the patients. Moreover, the system includes a telematic platform capable of establishing cooperative telemedicine sessions where our most impressive utilization is a transcontinental work platform for cardiovascular ultrasound. The key point of our approach starts with the construction of a DICOM private transfer syntax that is prepared to support any video encoder installed on a Windows-based station. With this structure it is possible to select the best encoder to a specific modality and work scenario. Good trade-off between compression ratio and diagnostic quality, low network traffic load, backup facilities and data portability are other achievements of this system.

An Application Service Provider (ASP) archive model for disaster recovery for Saint John’s Health Center (SJHC) clinical PACS data has been implemented using a Fault-Tolerant Archive Server at the Image Processing and Informatics Laboratory, Marina del Rey, CA (IPIL) since mid-2002. The purpose of this paper is to provide clinical experiences with the implementation of an ASP model backup archive in conjunction with handheld wireless technologies for a particular disaster recovery scenario, an earthquake, in which the local PACS archive and the hospital are destroyed and the patients are moved from one hospital to another. The three sites involved are: (1) SJHC, the simulated disaster site; (2) IPIL, the ASP backup archive site; and (3) University of California, Los Angeles Medical Center (UCLA), the relocated patient site. An ASP backup archive has been established at IPIL to receive clinical PACS images daily using a T1 line from SJHC for backup and disaster recovery storage. Procedures were established to test the network connectivity and data integrity on a regular basis. In a given disaster scenario where the local PACS archive has been destroyed and the patients need to be moved to a second hospital, a wireless handheld device such as a Personal Digital Assistant (PDA) can be utilized to route images to the second hospital site with a PACS and reviewed by radiologists. To simulate this disaster scenario, a wireless network was implemented within the clinical environment in all three sites: SJHC, IPIL, and UCLA. Upon executing the disaster scenario, the SJHC PACS archive server simulates a downtime disaster event. Using the PDA, the radiologist at UCLA can query the ASP backup archive server at IPIL for PACS images and route them directly to UCLA. Implementation experiences integrating this solution within the three clinical environments as well as the wireless performance are discussed. A clinical downtime disaster scenario was implemented and successfully tested. Radiologists were able to successfully query PACS images utilizing a wireless handheld device from the ASP backup archive at IPIL and route the PACS images directly to a second clinical site at UCLA where they and the patients are located at that time. In a disaster scenario, using a wireless device, radiologists at the disaster health care center can route PACS data from an ASP backup archive server to be reviewed in a live clinical PACS environment at a secondary site. This solution allows Radiologists to use a wireless handheld device to control the image workflow and to review PACS images during a major disaster event where patients must be moved to a secondary site.

Managing medical digital information objects, and in particular medical images is an enterprise-grade problem. Firstly, there is the sheer amount of digital data that is generated in the proliferation of digital (and film-free) medical imaging. Secondly, the managing software ought to enjoy high availability, recoverability and manageability that are found only in the most business-critical systems. Indeed, such requirements are borrowed from the business enterprise world. Moreover, the solution for the medical information management problem should too employ the same software tools, middlewares and architectures. It is safe to say that all first-line medical PACS products strive to provide a solution for all these challenging requirements. The DICOM standard has been a prime enabler of such solutions. DICOM created the interconnectivity, which made it possible for a PACS service to manage millions of exams consisting of trillions of images. With the more comprehensive IHE architecture, the enterprise is expanded into a multi-facility regional conglomerate, which presents extreme demands from the data management system. HIPPA legislations add considerable challenges per security, privacy and other legal issues, which aggravate the situation. In this paper, we firstly present what in our view should be the general requirements for a first-line medical PACS, taken from an enterprise medical imaging storage and management solution perspective. While these requirements can be met by homegrown implementations, we suggest looking at the existing technologies, which have emerged in the recent years to meet exactly these challenges in the business world. We present an evolutionary process, which led to the design and implementation of a medical object management subsystem. This is indeed an enterprise medical imaging solution that is built upon respective technological components. The system answers all these challenges simply by not reinventing wheels, but rather reusing the best “wheels” for the job. Relying on such middleware components allowed us to concentrate on added value for this specific problem domain.

Similar to DICOM for PACS (Picture Archiving and Communication System), standards for radiotherapy (RT) information have been ratified with seven DICOM-RT objects and their IODs (Information Object Definitions), which are more than just images. This presentation describes how a DICOM-based RT Information System Server can be built based on the PACS technology and its data model for a web-based distribution. Methods: The RT information System consists of a Modality Simulator, a data format translator, a RT Gateway, the DICOM RT Server, and the Web-based Application Server. The DICOM RT Server was designed based on a PACS data model and was connected to a Web application Server for distribution of the RT information including therapeutic plans, structures, dose distribution, images and records. Various DICOM RT objects of the patient transmitted to the RT Server were routed to the Web Application Server where the contents of the DICOM RT objects were decoded and mapped to the corresponding location of the RT data model for display in the specially-designed Graphic User Interface. The non-DICOM objects were first rendered to DICOM RT Objects in the translator before they were sent to the RT Server. Results: Ten clinical cases have been collected from different hopsitals for evaluation of the DICOM-based RT Information System. They were successfully routed through the data flow and displayed in the client workstation of the RT information System. Conclusion: Using the DICOM-RT standards, integration of RT data from different vendors is possible.

The purpose of this study is to establish a reference for image acquisition for completing a standard brain for diverse Korean population, and to develop database management system that saves and manages acquired brain images and personal information of subjects. 3D MP-RAGE (Magnetization Prepared Rapid Gradient Echo) technique which has excellent Signal to Noise Ratio (SNR) and Contrast to Noise Ratio (CNR) as well as reduces image acquisition time was selected for anatomical image acquisition, and parameter values were obtained for the optimal image acquisition. Using these standards, image data of 121 young adults (early twenties) were obtained and stored in the system. System was designed to obtain, save, and manage not only anatomical image data but also subjects' basic demographic factors, medical history, handedness inventory, state-trait anxiety inventory, A-type personality inventory, self-assessment depression inventory, mini-mental state examination, intelligence test, and results of personality test via a survey questionnaire. Additionally this system was designed to have functions of saving, inserting, deleting, searching, and printing image data and personal information of subjects, and to have accessibility to them as well as automatic connection setup with ODBC. This newly developed system may have major contribution to the completion of a standard brain for diverse Korean population since it can save and manage their image data and personal information.

Content Based Color Cell Pathology Image Retrieval is one of the newest computer image processing applications in medicine. Recently, some algorithms have been developed to achieve this goal. Because of the particularity of cell pathology images, the result of the image retrieval based on single characteristic is not satisfactory. A new method for pathology image retrieval by combining color, texture and morphologic features to search cell images is proposed. Firstly, nucleus regions of leukocytes in images are automatically segmented by K-mean clustering method. Then single leukocyte region is detected by utilizing thresholding algorithm segmentation and mathematics morphology. The features that include color, texture and morphologic features are extracted from single leukocyte to represent main attribute in the search query. The features are then normalized because the numerical value range and physical meaning of extracted features are different. Finally, the relevance feedback system is introduced. So that the system can automatically adjust the weights of different features and improve the results of retrieval system according to the feedback information. Retrieval results using the proposed method fit closely with human perception and are better than those obtained with the methods based on single feature.

In a hospital, various kinds of medical images acquired from different modalities are generally used and stored in different department and each modality usually attaches several workstations to display or process images. To do better diagnosis, radiologists or physicians often need to retrieve other kinds of images for reference. The traditional image storage solution is to buildup a large-scale PACS archive server. However, the disadvantages of pure centralized management of PACS archive server are obvious. Besides high costs, any failure of PACS archive server would cripple the entire PACS operation. Here we present a new approach to develop the storage grid in PACS, which can provide more reliable image storage and more efficient query/retrieval for the whole hospital applications. In this paper, we also give the performance evaluation by comparing the three popular technologies mirror, cluster and grid.

We developed a Web-based system to interactively display image-based electronic patient records (EPR) for intranet and Internet collaborative medical applications. The system consists of four major components: EPR DICOM gateway (EPR-GW), Image-based EPR repository server (EPR-Server), Web Server and EPR DICOM viewer (EPR-Viewer). We have successfully used this system two times for the teleconsultation on Severe acute respiratory syndrome (SARS) in Shanghai Xinhua Hospital and Shanghai Infection Hospital. During the consultation, both the physicians in infection control area and the experts outside the control area could interactively study, manipulate and navigate the EPR of the SARS patients to make more precise diagnosis on images with this system assisting. This presentation gave a new approach to create and manage image-based EPR from actual patient records, and also presented a way to use Web technology and DICOM standard to build an open architecture for collaborative medical applications.

Purpose: Assessment of bone maturity is traditionally performed through visual comparison of hand and wrist radiograph with existing reference images in textbooks. Our goal was to develop a digital index based on idealized hand Xray images that can be incorporated in a hand held computer and used for visual assessment of bone age for patients. Material and methods: Due to the large variability in bone maturation in normals, we generated a set of “ideal” images obtained by computer combinations of images from our normal reference data sets. Software for hand-held PDA devices was developed for easy navigation through the set of images and visual selection of matching images. A formula based on our statistical analysis provides the standard deviation from normal based on the chronological age of the patient. The accuracy of the program was compared to traditional interpretation by two radiologists in a double blind reading of 200 normal Caucasian children (100 boys, 100 girls). Results: Strong correlations were present between chronological age and bone age (r > 0.9) with no statistical difference between the digital and traditional assessment methods. Determinations of carpal bone maturity in adolescents was slightly more accurate using the digital system. The users did praise the convenience and effectiveness of the digital Palm Index in clinical practice. Conclusion: An idealized digital Palm Bone Age Index provides a convenient and effective alternative to conventional atlases for the assessment of skeletal maturity.

Designing proper models for authorization and access control for the electronic patient record (EPR) is essential to wide scale use of the EPR in large health organizations. This work presents MAAC (Middleware for Authentication and Access Control), a tool that implements a contextual role-based access control (RBAC) authorization model. RBAC regulates user’s access to computers resources based on their organizational roles. A contextual authorization uses environmental information available at access-request time, like user/patient relationship, in order to decide whether a user has the right to access an EPR resource. The software architecture where MAAC is implemented uses Lightweight Directory Access Protocol, Java programming language and the CORBA/OMG standards CORBA Security Service and Resource Access Decision Facility. With those open and distributed standards, heterogeneous EPR components can request user authentication and access authorization services in a unified and consistent fashion across multiple platforms.

As PACS and related imaging technologies become more common in the healthcare setting, better methods are needed for training radiologists, allied healthcare and IT providers to understand and use them effectively. We presented a PACS Simulator in 2002 and a RIS integrated PACS Simulator in 2003. In this presentation we integrate a web-based image distribution and display system with the Simulator to complete the simulation of clinical PACS workflow. The Simulator consists of RIS Simulator, modality Simulator, DICOM gateway, PACS controller, clinical viewing workstations, PACS monitor system and a web server. The addition web server distributes images to browser-based display clients for display and manipulation using web technology. Using this Simulator, trainees can: (1) Observe RIS/PACS operation with web image distribution, component by component, (2) Trace image data flow through each component, (3) Query, retrieve and review images at web display clients, (4) Induce failure in a component to observe its impact on the entire RIS/PACS operation. The Simulator has been developed as a stand-alone tool for trainees to understand RIS/PACS concept and appreciate RIS/PACS workflow with hands-on experience. The Simulator has been applied in many training classes, and the addition of web server will significantly improve its training value.

One of the greatest difficulties of dealing with medical images is their distinct characteristics, in terms of generation process and noise that requires different forms of treatment for visualization and processing. Besides that, medical images are only a compounding part of the patient’s history, which should be accessible for the user in an understandable way. Other factors that can be used to enhance the user capability and experience are: the computational power of the client machine; available knowledge about the case; if the access is local or remote and what kind of user is accessing the system (physician, nurse, administrator, etc...). These information compose the context of an application and should define its behavior during execution time. In this article, we present the architecture of a viewer that takes into account the contextual information that is present at the moment of execution. We also present a viewer of X-Ray Angiographic images that uses contextual information about the client's hardware and the kind of user to, if necessary, reduce the image size and hide demographic information of the patient. The proposed architecture is extensible, allowing the inclusion of new tools and viewers, being adaptive along time to the evolution of the medical systems.

With the advent of the digital devices for medical diagnosis the use of the regular films in radiology has decreased. Thus, the management and handling of medical images in digital format has become an important and critical task. In Cardiology, for example, the main difficulty is to display dynamic images with the appropriated color palette and frame rate used on acquisition process by Cath, Angio and Echo systems. In addition, other difficulty is handling large images in memory by any existing personal computer, including thin clients. In this work we present a web-based application that carries out these tasks with robustness and excellent performance, without burdening the server and network. This application provides near-diagnostic quality display of cardiac images stored as DICOM 3.0 files via a web browser and provides a set of resources that allows the viewing of still and dynamic images. It can access image files from the local disks, or network connection. Its features include: allows real-time playback, dynamic thumbnails image viewing during loading, access to patient database information, image processing tools, linear and angular measurements, on-screen annotations, image printing and exporting DICOM images to other image formats, and many others, all characterized by a pleasant user-friendly interface, inside a Web browser by means of a Java application. This approach offers some advantages over the most of medical images viewers, such as: facility of installation, integration with other systems by means of public and standardized interfaces, platform independence, efficient manipulation and display of medical images, all with high performance.

Lung cancer is the most common cause, accounting for about 20% of all cancer deaths for males in Japan. Myocardial infarction is also known as a most fearful adult disease. Recently, multi-helical CT scanner advanced remarkably at the speed at which the chest CT images were acquired for screening examination. This screening examination requires a considerable number of images to be read. It is this time-consuming step that makes the use of multi-helical CT for mass screening. To overcome this problem, our group has developed a computer-aided diagnosis algorithm to automatically detect suspicious regions of lung cancer and coronary calcifications in chest CT images, so far. And in this time, our group has developed a newly computer-aided diagnosis workstation and database. These consist in three. First, it is an image processing system to automatically detect suspicious bronchial regions, pulmonary artery regions, plumonary vein regions and myocardial infarction regions at high speed. Second, they are two 1600 x 1200 matrix black and white liquid crystal monitor. Third, it is a terminal of image storage. These are connected mutually on the network. This makes it much easier to read images, since the 3D image of suspicious regions and shadow of suspicious regions can be displayed simultaneously on two 1600 x 1200 matrix liquid crystal monitor. The experimental results indicate that a newly computer-aided diagnosis workstation and database system can be effectively used in clinical practice to increase the speed and accuracy of routine diagnosis.

Medical image security in a PACS environment has become a pressing issue as communications of images increasingly extends over open networks, and hospitals are currently hard-pushed by Health Insurance Portability and Accountability Act (HIPAA) to be HIPPA complaint for ensuring health data security. Other security-related guidelines and technical standards continue bringing to the public attention in healthcare. However, there is not an infrastructure or systematic method to implement and deploy these standards in a PACS. In this paper, we first review DICOM Part15 standard for secure communications of medical images and the HIPAA impacts on PACS security, as well as our previous works on image security. Then we outline a security infrastructure in a HIPAA mandated PACS environment using a dedicated PACS image security server. The server manages its own database of all image security information. It acts as an image Authority for checking and certificating the image origin and integrity upon request by a user, as a secure DICOM gateway to the outside connections and meanwhile also as a PACS operation monitor for HIPAA supporting information.

Physicians are used to meet and discuss cases periodically and the objective of the discussion is not only finding diagnosis and better treatment options, but to share knowledge and teach residents. Although the discussions are very rich, they are also unstructured and not well documented. This work proposes the use of collaborative tools and proper archiving of the discussions, aiming to collect and store the knowledge exposed in medical meetings in a structured form for future retrieval. To achieve this goal, this research involves fields like CSCW (Computer Supported Collaborative Work), Natural Language Processing and Telemedicine. This research is a work in progress and is focused on the Brazilian health care context. A model for electronic medical discussion and archiving is presented.

In the current environment of medical information disclosure, the general-purpose image format such as JPEG/BMP which does not require special software for viewing, is suitable for carrying and managing medical image information individually. These formats have no way to know patient and study information. We have therefore developed two kinds of ID embedding methods: one is Bit-swapping method for embedding Alteration detection ID and the other is data-imposing method in Fourier domain using Discrete Cosine Transform (DCT) for embedding Original image source ID. We then applied these two digital watermark methods to four modality images (Chest X-ray, Head CT, Abdomen CT, Bone scintigraphy). However, there were some cases where the digital watermarked ID could not be detected correctly due to image degradation caused by image processing. In this study, we improved the detection rate in digital watermarked image using several techniques, which are Error correction method, Majority correction method, and Scramble location method. We applied these techniques to digital watermarked images against image processing (Smoothing) and evaluated the effectiveness. As a result, Majority correction method is effective to improve the detection rate in digital watermarked image against image degradation.