Ramon Alfredo Moreno

Managing Medical Images and Clinical Information: InCor's Experience

Patients usually get medical assistance in several clinics and hospitals during their lifetime, archiving vital information in a dispersed way. Clearly, a proper patient care should take into account that information in order to check for incompatibilities, avoid unnecessary exams, and get relevant clinical history. The Heart Institute (InCor) of Satildeo Paulo, Brazil, has been committed to the goal of integrating all exams and clinical information within the institution and other hospitals. Since InCor is one of the six institutes of the University of Satildeo Paulo Medical School and each institute has its own information system, exchanging information among the institutes is also a very important aspect that has been considered. In the last few years, a system for transmission, archiving, retrieval, processing, and visualization of medical images integrated with a hospital information system has been successfully created and constitutes the InCors electronic patient record (EPR). This work describes the experience in the effort to develop a functional and comprehensive EPR, which includes laboratory exams, images (static, dynamic, and three dimensional), clinical reports, documents, and even real-time vital signals. A security policy based on a contextual role-based access control model was implemented to regulate users access to EPR. Currently, more than 10 TB of digital imaging and communications in medicine (DICOM) images have been stored using the proposed architecture and the EPR stores daily more than 11 GB of integrated data. The proposed storage subsystem allows 6 months of visibility for rapid retrieval and more than two years for automatic retrieval using a jukebox. This paper addresses also a prototype for the integration of distributed and heterogeneous EPR

Integration of heterogeneous systems via DICOM and the clinical image access service

In this paper we present a proposal for the integration of DICOM Structured Reporting (DICOM SR) and the Clinical Image Access Service (CIAS). The advantage of this approach is that the structure of information defined by the DICOM SR is well known and can be understood through institutions with little effort. It is also important because we believe that DICOM SR will be widely adopted together with the CORBAmed specification. The DICOM SR and the CIAS models were integrated by mapping the nodes of DICOM SR into observations and the relationships of DICOM SR (HAS PROPERTY, CONTAINS, etc.) to observation reference objects. For a more complete union of the standards the mapped information (in both directions) can be accessed by the Structured Reporting Object Model (SROM) which makes transparent the navigation in the tree nodes, supporting different implementations for the same interface. The union of these standards brings together the best side of each one, giving semantics to CIAS while improving DICOM through transparency of location, naming service and easier integration of legacy systems.

Cluster-based classification using self-organising maps for medical image databases

Images are a fundamental source of information in medicine. They can support doctors and students in diagnostic decisions besides providing research and didactic material. The images stored in a database and divided in categories are an important step for data mining and content-based image retrieval (CBIR). This work addresses a methodology which joins the use of discrete wavelet transforms to characterise images and self-organising maps (SOM) neural networks to cluster based classification of medical images. This data mining methodology can be used in categorisation and in computer-aided diagnostic decision.

A Contextual Medical Image Viewer

The conventional visualization of medical images is not enough for a rich and comprehensive electronic healthcare record. We believe that it is necessary to provide a viewer with more advanced capabilities than those of regular medical image viewers. In this paper, we propose an architecture that allows the use of contextual information to assist the healthcare professional in his regular tasks. The proposed architecture for the context is composed of an ontology describing the hospital and an inference engine. The result is a Java-based implementation of the architecture, named Dynamic N-dimensional Image Viewer, (ODIN in Portuguese) that is able to adapt its behavior according to the contextual information.

Integrating Medical Images and Clinical Information

The Heart Institute (InCor) of Sao Paulo has been committed to the goal of integrating all clinical information within the institution. In the last few years, InCor has successfully created a system for transmission, archiving, retrieval, processing and visualization of Medical Images and also a Hospital Information System that stores the administrative and clinical information. These integrated subsystems form InCors Electronic Patient Record (EPR). This work describes the experience in the effort to develop a functional and comprehensive EPR, which includes access control, lab exams, images (static, dynamic and 3D), clinical reports, documents and even real-time vital signals. Currently, more than 13TB of DICOM images have been stored using the proposed architecture. The EPR stores more than 5 GB/day of integrated data and presents more than 1400 hits per day. The proposed storage subsystem allows six months of visibility for rapid retrieval and more than two years for automatic retrieval using a jukebox.

Medical Image Categorization Based on Wavelet Transform and Self-Organizing Map

Nowadays, images are fundamental data source in modern medicine. The images stored in a database according with categories are an important step for data mining and content-based image retrieval (CBIR). These can support doctors and students in diagnostic decisions and provide research and didactic material. This work addresses the use of discrete wavelet transform and self-organizing map (SOM) to medical image categorization. Furthermore, extensive experiments to define map size, finetune using linear vector quantization and a contrastive study with another success approach of categorization are realized.

Implementation of an integrated hospital information system using an open-source three-tier architecture

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 InCors 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.

Contextual medical-image viewer

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 clients 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.

Combining wavelets transform and Hu moments with self-organizing maps for medical image categorization

Images are fundamental sources of information in modern medicine. The images stored in a database and divided in categories are an important step for image retrieval. For an automatic categorization process, detailed analysis is done regarding image representation and generalization method. The baseline method for this process, in the medical image context, is using thumbnails and K-nearest neighbor (KNN), which is easily implemented and has had satisfactory results in literature. This work addresses an alternative method for automatic categorization, which jointly uses discrete wavelet transform with Hus moments for image representation and self-organizing maps (SOM) neural networks combined with the KNN classifier (SOM-KNN), for medical image categorization. Furthermore, extensive experiments are conducted, to define the best wavelet family and to select the best coefficients set, to consider the remaining wavelet coefficients set (not selected as the best ones) through their Hus moments, and to carry out a contrastive study with other successful approaches for categorization. The categorization result from a database with 10,000 images in 116 categories yielded 81.8% of correct rate, which is much better than the 67.9% obtained by the baseline method; and the time consumed in classification processing with SOM-KNN is 100 times shorter than KNN.

BIRAM : A content-based image retrieval framework for medical images

In the medical field, digital images are becoming more and more important for diagnostics and therapy of the patients. At the same time, the development of new technologies has increased the amount of image data produced in a hospital. This creates a demand for access methods that offer more than text-based queries for retrieval of the information. In this paper is proposed a framework for the retrieval of medical images that allows the use of different algorithms for the search of medical images by similarity. The framework also enables the search for textual information from an associated medical report and DICOM header information. The proposed system can be used for support of clinical decision making and is intended to be integrated with an open source picture, archiving and communication systems (PACS). The BIRAM has the following advantages: (i) Can receive several types of algorithms for image similarity search; (ii) Allows the codification of the report according to a medical dictionary, improving the indexing of the information and retrieval; (iii) The algorithms can be selectively applied to images with the appropriated characteristics, for instance, only in magnetic resonance images. The framework was implemented in Java language using a MS Access 97 database. The proposed framework can still be improved, by the use of regions of interest (ROI), indexing with slim-trees and integration with a PACS Server.