Articles catalogue 2010

Supervisor: Doc. RNDr. Ivan Mazura, CSc.

Abstract

Keywords: astrocytoma, glioblastoma, imunohistochemistry, molecular-biology methods, mutations, tumorsuppressor genes, amplifications, p53 gene, PTEN gene, EGFR amplification

 full article

 Česky

Supervisor: Doc. Ing. Lenka Lhotská, CSc.

Abstract

Utilization of procedural knowledge in the form of organizational processes and formalized medical guidelines can be useful in decision support systems (DSSs) in health care domain. The problem of using this form of knowledge arises when a multi-agent paradigm is to be applied in a DSS due to differences in specification of behavioural models of agents and process formalisms. In this work we continue in enhancing a novel process-based multi-agent architecture and demonstrate its integration into an existing DSS (K4care) focused on home care.

We analysed available documentation of the complex system K4Care and identified possible mutual common functionalities of implemented multi-agent system with the new architecture. These were the entry points, using which we further enhanced the K4Care platform with respect to the process-based multi-agent architecture.

The analysis proved that the integration is not only possible, but thanks to the general design of the process-based multi-agent architecture can be done with only small changes in the existing K4Care model. Immediate improvements in supporting human experts were identified and possible further improvements of the system were discussed.

Adopting the process-based multi-agent architecture can be beneficial even for existing DSSs and can open new possible features emerging from the multi-agent paradigm.

Keywords: multi-agent systems, multi-agent architecture, organizational processes, formalized medical guidelines, health care, home care, K4Care

 full article

 Česky

 Slovensky

Supervisor: Prof. MUDr. Taťjana Dostálová, DrSc., MBA

Abstract

The recent progress in understanding the biomedical basis for temporomandibular joint. (TMJ) therapy as well as in computer and imaging technologies is beginning to provide novel insights into TMJ reconstruction, rational diagnostic and therapeutic strategies. Several methods are available for imaging the TMJ. These include basic radiography (such as panorexes and corrected tomograms), ultrasonography, magnetic resonance imaging, and spiral or conebeam computed tomography. The latter technologies allow the joint to be visualized as sections in different planes and they can also be rendered as 3D volumetric reconstructions to enhance diagnostic capabilities. The aim of the study is to demonstrate patient’s therapy including 1 year recall.

This contribution deals with 3D analysis of the mathematical and computer modeling of biomechanics of the jaw. Case report monitors step by step treatment plan, preprosthetic surgery, mandibular reconstruction and process of healing. 3D-CT was used to monitor the patient with dental cysts.

The mathematical and computer TMJ models analysis using overlapping of 3D-CT before and after 1 year therapy shows the stability of TMJ partial prosthesis reconstruction.

Mathematical modeling and mathematical simulation of the reconstruction of patient jaw and/or patient TMJ may predict the joint loading of the reconstructed condyle. Our study demonstrated practical use of 3D images during TMJ therapy. Although many of the disorders involving the temporomandibular joint and associated structures can be diagnosed clinically on the basis of the history and physical findings, there are others that require the use of various imaging techniques to make an accurate diagnosis or to determine the extent of involvement. Imaging the hard and soft tissues of the temporomandibular joint can be used not only for treatment planning but also long-term treatment evaluation.

Keywords: dentistry, maxillofacial surgery, partial TMJ prosthesis, three-dimensional computed tomography, magnetic resonance imaging, temporomandibular joint

 full article

 Česky

Supervisor: Doc. Zdeněk Valenta, M.Sc., M.S., PhD.

Abstract

The survival analysis is a set of statistical methods dealing with time-to-event data. In biomedical applications the event of interest is usually relapse of the disease or death. A special feature of the survival analysis is censoring and truncation of data. When censoring or truncation occurs some information about the patients’ survival is lost, e.g. some patients are lost to follow-up or the study ends before all the patients die. The survival analysis methods are used for estimation of the survival time distribution, for identification of risk factors that affect the survival time, and also for predicting the survival time when risk factors are present. Survival analysis methods have been further developed by the means of counting processes and martingale theory. Univariate survival analysis methods have been extended to multivariate setting. The multivariate survival analysis covers the field where independence between survival times cannot be assumed. Multi-state models and frailty models represent the two main approaches of multivariate methods.

Keywords: survival analysis, survival function, hazard function, cumulative hazard function, censoring, truncation, Kaplan-Meier estimator, Nelson-Aalen estimator, Cox PH model, partial likelihood, counting process, history, filtration, martingale, competing risk, multi-state models, frailty models

 full article

 Česky

Supervisor: Prof. MUDr. Taťjana Dostálová, DrSc., MBA

Abstract

Our experimental work is engaged in creating a new type of Electronic Health Record (EHR) in the field of dentistry, especially for temporomandibular joint (TMJ) disorders (TMD) and the main goal is creation of the comprehensive system, which would be a good basis for the decision support in the TMD therapy.

Our EHR system is based on the Dental knowledge base, which is constructed in the application MUDR (Multimedia Distributed EHR) and then in the MUDRLite application. The MUDRLite architecture is built on two tiers (relational database (e.g. MS SQL server) and User Interface layer (MUDRLite UI), which is described by an XML configuration file). These applications are creating the real graphical user application DentCross, which is connected with the ASR (Automatic speech recognition) system and TTS (Text-to-speech) module. Special TMD EHR is a part of the DentCross and is created in the same applications. This part uses the RCD/TMD classification (Research diagnostic criteria for TMD).

We present the Integrated Dental-TMJ EHR system, which allows recording patient’s complex medical data in the field of dentistry. It affects the whole orthognathic system with its all components (especially TMD problematic) in the graphical form. The recording and data manipulation are enhanced with the ASR system and TTS module. This application has been constructed with an aim on ease of control and capability of covering medical information and should facilitate decision process in the therapy of TMD.

The integrated Dental-TMJ EHR system represents whole patient’s dental information, which is essential for the treatment plan creation. Connection with the ASR system and TTS module simplifies collecting of this information. Information is also well-arranged in the graphical form. Next step is continuation of clinical testing of the whole system and improvement of the ASR system and TTS module. The disadvantage of this system could be necessity of experience in operating with this system and ASR system functionality in the noisy background.

Keywords: electronic health record, automatic speech recognition, dental cross, temporomandibular joint, temporomandibular joint disorders, structured data entry, dentistry, data model, text-to-speech system, Research Diagnostic Criteria for TMD

 full article

 Česky

Supervisor: Doc. MUDr. Ing. Daniel Smutek, Ph.D.

Abstract

The objective has been to compare success of the texture classifier and a human observer in diagnosis of the autoimmune thyroiditis from B-mode ultrasound images and to determine inter- and intra-observer variability.

The data set of 161 subjects was classified by four human observers and by the Bayes classifier based on the texture features to three classes (healthy, border state, autoimmune thyroiditis).

Two observers had a higher success rate when classifying the healthy class (74.4 % and 83.3 %), the other two observers classified better cases with autoimmune thyroiditis (59.0 % and 77.4 %). The classifier gave the relatively high and balanced success rate for both classes (100.0 % for healthy and 875 % for thyroiditis). The different observers’ success rates resulted in the high inter-observer variability, showing only a fair agreement among the human observers. There was no significant difference among human observers in the intra-observer variability.

Due to the fair agreement among observers in the diagnosis of autoimmune thyroiditis from ultrasound images and good results of the classifier, the best way in establishing diagnosis is computer-aided diagnosis combined with observers’ clinical experience.

Keywords: thyroid gland, autoimmune thyroiditis, ultrasound image, B-mode ultrasound, texture analysis, computer-aided classification, inter-observer variability, intra-observer variability, Kappa Statistics, weighted Kappa

 full article

 Česky

Supervisor: Prof. RNDr. Jana Zvárová, DrSc.

Abstract

We deal with the relation between the generalized entropy (f-entropy, a family of functions that include several biodiversity measures) of a discrete random variable and the minimal probability of error (Bayes error) when the value of this random variable is estimated. Namely the tightness of their relation is studied.

Morales and Vajda [1] recently introduced a measure called average inaccuracy that aims to quantify the tightness of the relation between the posterior Bayes error and the power entropies. It is defined as a standardized average difference between the upper and the lower bound for the posterior Bayes error under given entropy. Their concept can be generalized to any strictly concave f-entropy and used to evaluate its relation to the Bayes probability of error. However, due to a complex form of the formula of the average inaccuracy, it is difficult to compare the average inaccuracies of most f-entropies analytically.

We propose a smooth approximation of the lower bound for the posterior Bayes error under given f-entropy that simplifies the formula of the average inaccuracy. We show that under this approximation, the quadratic entropy has the tightest relation to the posterior Bayes error among f-entropies.

The quadratic entropy has the tightest relation to the posterior Bayes error (in the sense described in this paper) than the Shannon’s entropy and other functions that belong to the family of f-entropies, like Emlen’s index, Ferreri’s index and Good’s index.

Keywords: generalized entropy, f-entropy, Bayes error, average inaccuracy, power entropy, quadratic entropy, Shannon’s entropy, Emlen’s index, Ferreri’s index, Good’s index

 full article

 Česky

Supervisor: Prof. RNDr. Jana Zvárová, DrSc.

Abstract

The assessment of the fetal size based on ultrasound biometry is important for proper pregnancy management. The means for evaluations of these measurements represent age-related reference intervals and centile charts which are used to detect the extreme values, possibly indicating pathology. The key features concerning the design of studies used for the construction of these intervals are given. The aim of this paper is to review possible statistical approaches, pointing out the methodology, goodness of fit, advantages and limitations.

Keywords: fetus, gestational age, ultrasound, biometry, pathology, reference interval, centile, centile chart, normal distribution, normalizing transformation, quantile regression

 full article

 Česky

Supervisor: Ing. Milan Šárek, CSc.

Abstract

Infrastructure as a service (infrastructure which is offered to a customer in the form of service of the provider) is a deployment model which allows utilize data and computing capacity of a cloud as a set of virtual devices and virtualized machines. Infrastructure as a service can be offered separately to each project. The same capacity of connected physical machines and devices can be shared. Currently, the concept of an Infrastructure as a service is tested on several projects within activity of the CESNET association, First Faculty of Medicine, Charles University, Prague and Musical and Dance Faculty of Academy of Performing Arts in Prague.

The current research in the field of computation physiology is demanding on a high computation capacity. The computation tasks are distributed to computers, which are provided by the infrastructure. The project in the field of the analysis of a human voice is demanding on high throughput of a computer network between an acoustic or video device on the local side and an analytic application on the remote high performance server side. This paper describes features and main challenges for infrastructure dedicated for such a type of an application. Infrastructure as a deployment model of cloud computing might be beneficial for a multi domain team and for collaboration and integration of a high specialized software application.

Keywords: cloud computing, infrastructure as a service, virtualization, computation physiology, identification of physiological systems, validation of physiological models, remote desktop protocol, grid computing, voice range profile

 full article

 Česky

Supervisor: Prof. RNDr. Jana Zvárová, DrSc.

Abstract

The objective of the paper is to compare Czech medical reports written in a free text and by means of a software application; to analyze the usability of international classification systems in the Czech healthcare environment. The analysis of medical reports was based on the attributes of the Minimal Data Model for Cardiology (MDMC). We have used medical reports written in a free text and medical reports from the ADAMEK software application where data are stored in a structured way. For our work SNOMED CT and ICD-10 have been used. We have focused on the language of Czech medical reports and the application of aforementioned international classification systems in MDMC. We have compared how well attributes of MDMC are recorded in textual medical reports and in medical reports recorded structurally by means of the ADAMEK software application. We have made the language analysis of the Czech textual medical reports. We compared how MDMC attributes are recorded in the ADAMEK application and in medical reports written in a free text. To conclude, using a free text in medical reports is very inhomogeneous and not standardized. The standardized terminology would bring benefits to physicians, patients, administrators, software developers and payers. It would help healthcare providers that it could provide complete and easily accessible information that belongs to the process of healthcare and it would result in better care of patients. The use of international classification systems is a necessary first step to enable interoperability of heterogeneous electronic health records.

Keywords: terminology, synonyms, classification systems, thesaurus, nomenclature, electronic health record, interoperability, semantic interoperability, cardiology, atherosclerosis

 full article

 Česky

Supervisor: Doc. Ing. Vojtěch Svátek, Dr.

Abstract

Modern technology offers a wide array of possibilities to publish almost any content freely on the Internet. Because of the importance and delicacy of medical information, the quality of such texts provided to general public seems to be a serious issue nowadays. Unfortunately, the only feasible way to approve the adequacy of the medical information content is human verification. Best practices in medicine are systematically captured by medical guidelines (MGL), which are provided by renowned medical societies and based on results of Evidence-Based Medicine (EBM).

We propose a simple approach exploiting MGL content as a benchmark for the assessment of a content quality in medical web sites (WS). It is based on the idea that the information content or at least the scope of a medical text is reflected in the domain terminology used. We discuss a possible use of this approach in semiautomatic human-based quality verification and various aspects related to its application.

Concept candidates discovered in a MGL and in the tested web pages are matched to UMLS, yielding sets of used medical terms and corresponding concepts. Several aggregation techniques for MGLs were proposed and tested. The two sets are analyzed for overall similarity at the term and concept level.

The method was applied on a selected medical topic employing relevant MGL and 100 WS. All the analyzed web pages fell into five distinct categories (corresponding to the target audience). Aggregations for the MGLs were proposed and tested. The average cosine similarity to MGL across all tested WS reached 0.69 whereas the average similarity calculated per each category varied up to 7,6 % against the overall number.

The research done is the first step towards automated evaluation of a medical web page content on the basis of MGLs as the quality standard. We describe further tasks which would improve the outputs of comparison and the possibility of its common application.

Keywords: information quality assessment, clinical vocabularies, Unified Medical Language System s (UMLS), Evidence-Based Medicine (EBM), Medical Guidelines (MGL), information quality, annotation, similarity, concept, content representation

 full article

 Česky

Supervisor: Doc. RNDr. Petr Lánský, CSc.

Abstract

One of the important tasks in the spike train analysis is to estimate the underlying firing rate function. The aim of this article is to improve the time performance of an algorithm which can be used for the estimation.

As there is no unique way how to infer the firing rate function, several different methods have been proposed. A popular method how to estimate this function is the convolution of a spike train with Gaussian kernel with appropriate kernel bandwidth. The definition of what “appropriate” means remains a matter of discussion and a recent paper [1] proposes a method how to exactly compute optimal bandwidth under certain conditions. For large sets of spike train data the elementary version of the algorithm is unfortunately too inefficient in terms of computational time complexity.

We present a refined version of the algorithm which in turn allows us to use the original method even for large data sets.

The achieved performance improvement is demonstrated on a particular results and shows usability of the proposed method.

Keywords: action potential, spike train, neural coding, firing rate, convolution, Gaussian kernel, kernel bandwidth, Brent's minimization, parallel computing, MPI

 full article

 Česky

Supervisor: Prof. MUDr. Štěpán Svačina, DrSc.

Abstract

Thyroid gland diseases are frequent in pregnant women. In the Czech Republic, there is not screening for this disease. Our objective was to evaluate the initial cost of the study of screening of thyroid gland disease in pregnant women.

We used simple statistical methods for comparing the cost of the study (examination of TSH and TPOAb) with the costs and efficiency when TSH-only or TPOAb-only screening would be performed.

Screening of TSH only is 2.7 times cost-effective than screening of TPOAb. Screening of TSH only may lead to miss 77 patients with TPOAb positive results and screening of TPOAb only may lead to miss 77 patients with TSH positive results.

To achieve better sensitivity, it is necessary to perform screening of TSH and TPOAb together.

Keywords: thyroid gland disease, thyreopathy, pregnancy, cost-effectivenes, screening, iq, qaly

 full article

 Česky

Supervisor: Doc. Ing. Vojtěch Svátek, Dr.

Abstract

Healthcare documentation in the Czech Republic usually has the form of a free text formatted just using spaces, tabs and line breaks. Extracting information from such a documentation is a challenge that if fulfilled would allow to use Czech medical reports by physicians with no knowledge of the Czech language as well as information transfer to a structured form. It is possible to approach this task as a task of finite-state machine, as a task of the linguistic analysis or as a task of statistics. This article summarizes our findings gained using finite-state machines and using commonly used code lists. Excerpts from real medical reports are translated to English in a way that demonstrates the same or similar problems as in the Czech language. Original Czech excerpts are available in the Czech version of this article.

Keywords: natural language processing, healthcare documentation, medical reports, EHR, finite-state machine, regular expression

 full article

 Česky

Supervisor: Doc. Ing. Arnošt Veselý, CSc.

Abstract

Since the nineties the publication activity of expert authorities in the field of Clinical practice guidelines (CPGs) has increased. During the first decade of the 21st century the Internet has become one of the most important publishing media for CPGs due to the popularity of sharing information in an electronic form. Nowadays, approximately 107 Czech expert societies publish about 731 CPGs on the Internet according to previous explorations of web sites of Czech expert medical societies and practical experience with previous operation of the Catalogue of Clinical Practice Guidelines. CPG documents can be used by physicians and other clinical employees in clinical practice and education, by administrative authorities, by health policy makers or by patients. Information in these documents is also essential in the development of clinical information systems and decision support systems.

The database of the Catalogue of Clinical Practice Guidelines takes a survey in the appearance of CPGs on free accessible web sites of expert societies or periodicals guaranteed by expert societies. The database stores bibliographical data, information about the recency and the quality of CPGs and information about the existence of web tools or formal models based on these CPGs.

The database contains 562 records of CPGs published by Czech expert medical societies. 212 records are accessible via the web interface of the Catalogue of Clinical Practice Guidelines.

Keywords: Internet, World Wide Web, database, clinical practice guideline, clinical practice, evidence-based medicine, formalisation, GLIF (Guideline Interchange Format), doctor of medicine, decision support systems

 full article

 Česky

Reprinted from: Medical Informatics in a United and Healthy Europe, K.P. Adlassnig et al. (Eds), IOS Press, 2009, 946-950, © 2009 European Federation for Medical Informatics.

Abstract

The paper presents an analysis of how EFMI disseminates new knowledge and the active medical informatics journals in EFMI member countries was carried out as an outcome of the EFMI Council meeting in London in 2008. The analysis identifies eight active major informatics journals and several other publications. Most are subscription-based and published at least quarterly. There is a possibility for the editors to meet regularly and form a community of practice with the aim of further improving their effectiveness in disseminating new knowledge and best practice in medical informatics. It is feasible to share expertise and it may be possible to harmonise several aspects of preparation and submission of manuscripts so that some of the identified barriers in publishing are reduced.

Keywords: EFMI, mission statement, task force of medical informatics journals

 full article

 Česky

 Hrvatski

Abstract

This paper is devoted to automatic localization of objects (eyes, mouth) in two-dimensional (2D) grey scale images of faces. Motivated by a practical problem in human genetics, the output of the localization of objects in the given database of images is needed for further tasks in the genetic research. A robust filter is applied on the image to ensure denoising. Templates are used as the main method. The mouth and both eyes are localized jointly using the weighted Pearson product-moment correlation coefficient or its robust analogy based on robust regression methods. In the database with 212 images of faces the method allows to locate the mouth and eyes correctly in 100 % of cases. Also the robust correlation coefficient based on the least weighted squares regression localizes the mouth and both eyes in 100 % of images of the given database. Robustness aspects of the method are examined with respect to rotation, noise, occlusion and asymmetry in the image. The joint localization of the mouth and both eyes produces the method invariant to rotation of any degree. This work is tailor made for the given images with expected usage of the methods in genetic applications.

Keywords: object localization, template matching, eye or mouth detection, robust correlation analysis, image denoising

 full article

 Česky

Reliability of measurement is a measure of its reproducibility under replicate conditions. The classical concept of reliability assumes that measurement Y is composed out of true value T and error term ε, two independent random variables, Y = T + ε . Reliability of measurement is defined as the ratio of the variance of the true scores to the variance of the observed scores. However, this concept is not applicable in models for dichotomous measurements which do not consider error terms and are instead defined via conditional probabilities. In this paper we examine a more general definition of reliability proposed in [1], which is based on decomposition of variance in mixed effects model. Proposed definition covers the classical definition of reliability and it is, moreover, appropriate for dichotomous measurements, too. Newly, for the proposed definition assumptions are derived, under which the reliability of composite measurement can be predicted by reliability of single measurement (Spearman-Brown formula) and approximate validity of Spearman-Brown formula is shown for the Rasch model. Finally, as a modification of the classical estimate of reliability based on Cronbach’s alpha, we examine its counterpart logistic alpha introduced in [2], which appears to be more appropriate for composite dichotomous measurements in some cases. Simulations show that the new estimate does not tend to underestimate reliability as often as the Cronbach’s alpha does. The new estimate is used in binary data of computerized process of myocardial perfusion diagnosis from cardiac single proton emission computed tomography (SPECT).

Keywords: reliability, binary data, logistic regression, Cronbach alpha, Rasch model, myocardial perfusion diagnosis

 full article

 Česky