Antoine Geissbuhler

A review of content-based image retrieval systems in medical applications-clinical benefits and future directions.

Content-based visual information retrieval (CBVIR) or content-based image retrieval (CBIR) has been one on the most vivid research areas in the field of computer vision over the last 10 years. The availability of large and steadily growing amounts of visual and multimedia data, and the development of the Internet underline the need to create thematic access methods that offer more than simple text-based queries or requests based on matching exact database fields. Many programs and tools have been developed to formulate and execute queries based on the visual or audio content and to help browsing large multimedia repositories. Still, no general breakthrough has been achieved with respect to large varied databases with documents of differing sorts and with varying characteristics. Answers to many questions with respect to speed, semantic descriptors or objective image interpretations are still unanswered. In the medical field, images, and especially digital images, are produced in ever-increasing quantities and used for diagnostics and therapy. The Radiology Department of the University Hospital of Geneva alone produced more than 12,000 images a day in 2002. The cardiology is currently the second largest producer of digital images, especially with videos of cardiac catheterization ( approximately 1800 exams per year containing almost 2000 images each). The total amount of cardiologic image data produced in the Geneva University Hospital was around 1 TB in 2002. Endoscopic videos can equally produce enormous amounts of data. With digital imaging and communications in medicine (DICOM), a standard for image communication has been set and patient information can be stored with the actual image(s), although still a few problems prevail with respect to the standardization. In several articles, content-based access to medical images for supporting clinical decision-making has been proposed that would ease the management of clinical data and scenarios for the integration of content-based access methods into picture archiving and communication systems (PACS) have been created. This article gives an overview of available literature in the field of content-based access to medical image data and on the technologies used in the field. Section 1 gives an introduction into generic content-based image retrieval and the technologies used. Section 2 explains the propositions for the use of image retrieval in medical practice and the various approaches. Example systems and application areas are described. Section 3 describes the techniques used in the implemented systems, their datasets and evaluations. Section 4 identifies possible clinical benefits of image retrieval systems in clinical practice as well as in research and education. New research directions are being defined that can prove to be useful. This article also identifies explanations to some of the outlined problems in the field as it looks like many propositions for systems are made from the medical domain and research prototypes are developed in computer science departments using medical datasets. Still, there are very few systems that seem to be used in clinical practice. It needs to be stated as well that the goal is not, in general, to replace text-based retrieval methods as they exist at the moment but to complement them with visual search tools.

Fully three-dimensional reconstruction for a PET camera with retractable septa

A fully 3-D reconstruction algorithm has been developed to reconstruct data from a 16 ring PET camera (a Siemens/CTI 953B) with automatically retractable septa. The tomograph is able to acquire coincidences between any pair of detector rings and septa retraction increases the total system count rate by a factor of 7.8 (including scatter) and 4.7 (scatter subtracted) for a uniform, 20 cm diameter cylinder. The reconstruction algorithm is based on 3-D filtered backprojection, expressed in a form suitable for the multi-angle sinogram data. Sinograms which are not measured due to the truncated cylindrical geometry of the tomograph, but which are required for a spatially invariant response function, are obtained by forward projection. After filtering, the complete set of sinograms is backprojected into a 3-D volume of 128*128*31 voxels using a voxel-driven procedure. The algorithm has been validated with simulation, and tested with both phantom and clinical data from the 953B. >

Learning from imbalanced data in surveillance of nosocomial infection

OBJECTIVE An important problem that arises in hospitals is the monitoring and detection of nosocomial or hospital acquired infections (NIs). This paper describes a retrospective analysis of a prevalence survey of NIs done in the Geneva University Hospital. Our goal is to identify patients with one or more NIs on the basis of clinical and other data collected during the survey. METHODS AND MATERIAL Standard surveillance strategies are time-consuming and cannot be applied hospital-wide; alternative methods are required. In NI detection viewed as a classification task, the main difficulty resides in the significant imbalance between positive or infected (11%) and negative (89%) cases. To remedy class imbalance, we explore two distinct avenues: (1) a new re-sampling approach in which both over-sampling of rare positives and under-sampling of the noninfected majority rely on synthetic cases (prototypes) generated via class-specific sub-clustering, and (2) a support vector algorithm in which asymmetrical margins are tuned to improve recognition of rare positive cases. RESULTS AND CONCLUSION Experiments have shown both approaches to be effective for the NI detection problem. Our novel re-sampling strategies perform remarkably better than classical random re-sampling. However, they are outperformed by asymmetrical soft margin support vector machines which attained a sensitivity rate of 92%, significantly better than the highest sensitivity (87%) obtained via prototype-based re-sampling.

Three dimensional reconstruction of PET data from a multi-ring camera

In order to assess the potential of a multiring camera for three-dimensional PET (positron emission tomography) the septa were removed from a commercially produced PET camera (ECAT 931/08-12). An approximate reconfiguration was made to the data acquisition subsystem to acquire the full set of 64 sinograms containing all permitted lines of response between pairs of detectors. A number of different phantoms were imaged with this camera configuration, and one fluorodopa study was performed with a healthy volunteer. The authors discuss the results of these studies, and describe the back-projection and filtering procedure that has been adopted in order to incorporate the maximum amount of cross-plane data in the reconstruction. The results are compared with the same data reconstructed using the standard two-dimensional approach (i.e. for the direct and cross planes only), and with data collected for the same phantoms with the septa in place. >

The RAFT network: 5 years of distance continuing medical education and tele-consultations over the Internet in French-speaking Africa

Continuing education of healthcare professionals is a key element for the quality and efficiency of a health system. In developing countries, this activity is usually limited to capitals, and delocalized professionals do not have access to such opportunities, or to didactic material adapted to their needs. This limits the interest of such professionals to remain active in the periphery, where they are most needed to implement effective strategies for prevention and first-line healthcare. Telemedicine tools enable the communication and sharing of medical information in electronic form, and thus facilitate access to remote expertise. A physician located far from a reference center can consult its colleagues remotely in order to resolve a difficult case, follow a continuous education course over the Internet, or access medical information from digital libraries or knowledge bases. These same tools can also be used to facilitate exchanges between centers of medical expertise: health institutions of a same country as well as across borders. Since 2000, the Geneva University Hospitals have been involved in coordinating the development of a network for eHealth in Africa (the RAFT, Réseau en Afrique Francophone pour la Télémédecine), first in Mali, and now extending to 10 French-speaking African countries. The core activity of the RAFT is the webcasting of interactive courses. These sessions put the emphasis on knowledge sharing across care professionals, usually in the form of presentations and dialogs between experts in different countries. The technology used for the webcasting works with a slow (25 kbits/s) internet connection. Other activities of the RAFT network include visioconferences, teleconsultations based on the iPath system, collaborative knowledge bases development, support for medical laboratories quality control, and the evaluation of the use of telemedicine in rural areas (via satellite connections) in the context of multisectorial development. Finally, a strong emphasis is put on the development of capacities for the creation, maintenance, and publication of quality medical didactic contents. Specific courses are organized for the national coordinators of the network to develop these competencies, with the help of the Health-On-the-Net Foundation. The richness of the plurality of knowledge and know-how must be steered towards emulation and sharing, respectful of each partners identity and culture. Collaborations with UNESCO and WHO have been initiated to address these challenges.

Building a reference multimedia database for interstitial lung diseases

This paper describes the methodology used to create a multimedia collection of cases with interstitial lung diseases (ILDs) at the University Hospitals of Geneva. The dataset contains high-resolution computed tomography (HRCT) image series with three-dimensional annotated regions of pathological lung tissue along with clinical parameters from patients with pathologically proven diagnoses of ILDs. The motivations for this work is to palliate the lack of publicly available collections of ILD cases to serve as a basis for the development and evaluation of image-based computerized diagnostic aid. After 38 months of data collection, the library contains 128 patients affected with one of the 13 histological diagnoses of ILDs, 108 image series with more than 41l of annotated lung tissue patterns as well as a comprehensive set of 99 clinical parameters related to ILDs. The database is available for research on request and after signature of a license agreement.

Automatic medical encoding with SNOMED categories

BackgroundIn this paper, we describe the design and preliminary evaluation of a new type of tools to speed up the encoding of episodes of care using the SNOMED CT terminology.MethodsThe proposed system can be used either as a search tool to browse the terminology or as a categorization tool to support automatic annotation of textual contents with SNOMED concepts. The general strategy is similar for both tools and is based on the fusion of two complementary retrieval strategies with thesaural resources. The first classification module uses a traditional vector-space retrieval engine which has been fine-tuned for the task, while the second classifier is based on regular variations of the term list. For evaluating the system, we use a sample of MEDLINE. SNOMED CT categories have been restricted to Medical Subject Headings (MeSH) using the SNOMED-MeSH mapping provided by the UMLS (version 2006).ResultsConsistent with previous investigations applied on biomedical terminologies, our results show that performances of the hybrid system are significantly improved as compared to each single module. For top returned concepts, a precision at high ranks (P0) of more than 80% is observed. In addition, a manual and qualitative evaluation on a dozen of MEDLINE abstracts suggests that SNOMED CT could represent an improvement compared to existing medical terminologies such as MeSH.ConclusionAlthough the precision of the SNOMED categorizer seems sufficient to help professional encoders, it is concluded that clinical benchmarks as well as usability studies are needed to assess the impact of our SNOMED encoding method in real settings.AvailabilitiesThe system is available for research purposes on: http://eagl.unige.ch/SNOCat.

Implementation of three-dimensional image reconstruction for multi-ring positron tomographs

In view of the number of PET studies involving low count rate acquisitions, there has been increasing interest recently in the development of positron cameras capable of fully three-dimensional acquisition and reconstruction. This interest has given impetus to the study of algorithms for 3D reconstruction, including those algorithms suitable for application to multi-ring PET scanners. While 2D reconstruction methods can often be generalised to 3D, a number of implementation problems arise which are unique to the 3D approach. This paper examines some of the difficulties associated with the generalisation of the filtered backprojection algorithm to 3D, paying particular attention to the approximations and variable transformations required for application to data from a multi-ring scanner.

Using lexical disambiguation and named-entity recognition to improve spelling correction in the electronic patient record

In this article, we show how a set of natural language processing (NLP) tools can be combined to improve the processing of clinical records. The study concentrates on improving spelling correction, which is of major importance for quality control in the electronic patient record (EPR). As first task, we report on the design of an improved interactive tool for correcting spelling errors. Unlike traditional systems, the linguistic context (both semantic and syntactic) is used to improve the correction strategy. The system is organized along three modules. Module 1 is based on a classical spelling checker, it means that it is context-independent and simply measures a string-edit-distance between a misspelled word and a list of well-formed words. Module 2 attempts to rank more relevantly the set of candidates provided by the first module using morpho-syntactic disambiguation tools. Module 3 processes words with the same part-of-speech (POS) and apply word-sense (WS) disambiguation in order to rerank the set of candidates. As second task, we show how this improved interactive spell checker can be cast as a fully automatic system by adjunction of another NLP module: a named-entity (NE) extractor, i.e. a tool able to identify words as such patient and physician names. This module is used to avoid replacement of named-entities when the system is not used in an interactive mode. Results confirm that using the linguistic context can improve interactive spelling correction, and justify the use of named-entity recognizer to conduct fully automatic spelling correction. It is concluded that NLP is mature enough to help information processing in EPR.

Near-Affine-Invariant Texture Learning for Lung Tissue Analysis Using Isotropic Wavelet Frames

We propose near-affine-invariant texture descriptors derived from isotropic wavelet frames for the characterization of lung tissue patterns in high-resolution computed tomography (HRCT) imaging. Affine invariance is desirable to enable learning of nondeterministic textures without a priori localizations, orientations, or sizes. When combined with complementary gray-level histograms, the proposed method allows a global classification accuracy of 76.9% with balanced precision among five classes of lung tissue using a leave-one-patient-out cross validation, in accordance with clinical practice.