Risk stratification of patients with locally aggressive differentiated thyroid cancer. Results of the MSDS trial.

UNLABELLED: The Multicentre Study Differentiated Thyroid Cancer (MSDS) collective represents a well defined group of patients with locally aggressive thyroid carcinomas (pT4; AJCC/UICC 1997). The aim of the present study was to compare the survival of patients with minimum and extensive extrathyroidal growth according to the new AJCC/UICC TNM staging system 2009. PATIENTS, METHODS: The follow-up data of 347 patients were analysed. Patients were reclassified according to the current AJCC/UICC 2009 classification. The event-free and overall survival was evaluated using Kaplan-Meier analysis. In addition, postoperative complications and status of disease were documented. RESULTS: 327 patients were assigned to stage pT3 and 20 patients to stage pT4a, respectively. Median follow-up was 6.1 years (range 0.04-9.8 years). 92.5% of patients reached complete remission. There were 7.8 % recurrences in the thyroid bed, in locoregional lymph nodes and/or in distant sites. The overall survival was >98% both in pT3 and pT4a patients (p = n. s.). In contrast, the event-free survival was significantly less favourable in pT4a patients (p < 0.001). Using multivariate analysis the following parameters were significant predictors of event-free survival: histological tumour type, degree of extrathyroidal extension and nodal metastasis (p < 0.05). CONCLUSIONS: The MSDS patients with locally aggressive differentiated thyroid cancer showed an excellent overall survival during a median follow-up of 6.1 years. According to the current AJCC/UICC 2009 classification, pT3 patients with minimal extrathyroidal extension revealed a significantly better event-free survival than pT4a patients with extensive extrathyroidal growth.

Design of a web portal for interdisciplinary image retrieval from multiple online image resources.

OBJECTIVES: Images play an important role in medicine. Finding the desired images within the multitude of online image databases is a time-consuming and frustrating process. Existing websites do not meet all the requirements for an ideal learning environment for medical students. This work intends to establish a new web portal providing a centralized access point to a selected number of online image databases. METHODS: A back-end system locates images on given websites and extracts relevant metadata. The images are indexed using UMLS and the MetaMap system provided by the US National Library of Medicine. Specially developed functions allow to create individual navigation structures. The front-end system suits the specific needs of medical students. A navigation structure consisting of several medical fields, university curricula and the ICD-10 was created. The images may be accessed via the given navigation structure or using different search functions. Cross-references are provided by the semantic relations of the UMLS. RESULTS: Over 25,000 images were identified and indexed. A pilot evaluation among medical students showed good first results concerning the acceptance of the developed navigation structures and search features. CONCLUSION: The integration of the images from different sources into the UMLS semantic network offers a quick and an easy-to-use learning environment.

Navigation in medical Internet image databases.

The world wide web (WWW) changes common ideas of database access. Hypertext Markup Language allows the simultaneous presentation of information from different sources such as static pages, results of queries from) databases or dynamically generated pages. 'Therefore, the metaphor of the WWW itself as a database was proposed by Mendelzon and Nlilo in 1998. Against this background the techniques of navigation within WWW-databases and the semantic types of their queries has e been analysed. Forty eight image repositories of different types and content, but all concerning medical essence, have been found by search-engines. Many different techniques are offered to enable navigation ranging from simple HTML-link-lists to complex applets. The applets in particular promise an improvement for navigation. Within the meta-information for querying, only ACR- and UMLS-encoding were found, but not standardized vocabularies like ICD10 or Terminologia Anatomica. UMLS especially shows that a well defined thesaurus can improve navigation. However, of the analysed databases only the UMLS 'metathesaurus' is currently implemented without providing additional navigation support based on the UMLS 'semantic network'. Including the information about relationships between the concepts of the metathesaurus or using the UMLS semantic network could provide a much easier navigation within a network of concepts pointing to multimedia files stored somewhere in the WWW.

Graphical tool for navigation within the semantic network of the UMLS metathesaurus on a locally installed database.

Knowledge in the environment of information technologies is bound to structured vocabularies. Medical data dictionaries are necessary for uniquely describing findings like diagnoses, procedures or functions. Therefore we decided to locally install a version of the Unified Medical Language System (UMLS) of the U.S. National Library of Medicine as a repository for defining entries of a medical multimedia database. Because of the requirement to extend the vocabulary in concepts and relations between existing concepts a graphical tool for appending new items to the database has been developed: Although the database is an instance of a semantic network the focus on single entries offers the opportunity of reducing the net to a tree within this detail. Based on the graph theorem, there are definitions of nodes of concepts and nodes of knowledge. The UMLS additionally offers the specification of sub-relations, which can be represented, too. Using this view it is possible to manage these 1:n-Relations in a simple tree view. On this background an explorer like graphical user interface has been realised to add new concepts and define new relationships between those and existing entries for adapting the UMLS for specific purposes such as describing medical multimedia objects.

Workflow analysis and evidence-based medicine: towards integration of knowledge-based functions in hospital information systems.

The large extent and complexity of scientific evidence described in the concept of evidence-based medicine often overwhelms clinicians who want to apply best external evidence. Hospital Information Systems usually do not provide knowledge-based functions to support context-sensitive linking to external information sources. Knowledge-based components need specific data, which must be entered manually and should be well adapted to clinical environment to be accepted by clinicians. This paper describes a workflow-based approach to understand and visualize clinical reality as a preliminary to designing software applications, and possible starting points for further software development.

Multimedia explorer: image database, image proxy-server and search-engine.

Multimedia plays a major role in medicine. Databases containing images, movies or other types of multimedia objects are increasing in number, especially on the WWW. However, no good retrieval mechanism or search engine currently exists to efficiently track down such multimedia sources in the vast of information provided by the WWW. Secondly, the tools for searching databases are usually not adapted to the properties of images. HTML pages do not allow complex searches. Therefore establishing a more comfortable retrieval involves the use of a higher programming level like JAVA. With this platform independent language it is possible to create extensions to commonly used web browsers. These applets offer a graphical user interface for high level navigation. We implemented a database using JAVA objects as the primary storage container which are then stored by a JAVA controlled ORACLE8 database. Navigation depends on a structured vocabulary enhanced by a semantic network. With this approach multimedia objects can be encapsulated within a logical module for quick data retrieval.

Synthesis of pharmacokinetic parameters of vancomycin via bootstrap methods.

OBJECTIVE: For the adjustment of individual vancomycin dosages, we estimate the important pharmacokinetic quantities half-life, clearance, and volume of distribution. MATERIAL: To obtain reliable information 293 observations from 244 patients were extracted from 23 published studies on vancomycin. Information about vancomycin's pharmacokinetics out of different sources represents an increase in sample size and, therefore, interpretive power. METHODS: Once the whole of the data had been stratified into a small number of homogeneous clusters based on cofactors, different (robust) estimators (mean, median, Winsorized, and trimmed mean) were calculated for the expected value of the pharmacokinetic parameters of vancomycin within the clusters. Measures of the statistical accuracy such as standard error, bias, mean square error, and confidence interval were estimated via bootstrap methods from large bootstrap sample sizes to compare the quality of the estimators. RESULTS: Due to the homogenization of the data all individual estimator functions yield very similar results and the empirical mean works fairly well as an estimate. The most frequently used estimator with the smallest estimated mean square error was the Winsorized mean.

Standardized structure and modular design of a pharmacokinetic database.

BACKGROUND: The accumulated knowledge on drugs can be used for an individual drug dosage adjustment if it is placed at our disposal in an informatically structured form. THEORY AND METHODS: We have started building up a pharmacokinetic database aimed at adjusting drug dosages, in exemplary form, to patients with renal impairment. Parameters needed for the three dosage adjustment rules (Dettli, Kunin, Holford) and the most general concept of pharmacokinetics constituted the theoretical basis. TWO PROCESSES PERTAIN TO ALL DRUGS: Distribution and elimination. Total drug clearance and at least two parameters representing distribution and elimination processes are closely interdependent in mathematical terms (clearance = volume of distribution*rate of elimination). This relation yields the unifying concept that serves as a prerequisite for a structured recording of 30 assigned pharmacokinetic and pharmacodynamic parameters within an informatic database. SOLUTIONS AND RESULTS: The information is retrieved and referenced from 2383 original publications by means of a standardized input module. The complete database at present contains 15,397 records for 1573 drugs. A programmed meta-analytic algorithm is used to calculate the statistical measures for the central value and variance--as available--from the pooled values of primary records. The statistically standardized parameters are extracted for 6601 pharmacokinetic parameters, and placed at the users disposal with the output module. PRACTICAL UTILITY: Following meta-analysis, published pharmacokinetics can be used as statistical estimates of population parameters. The statistical estimates with variances permit an individual drug dosage adjustment by applying the Bayesian approach or neural networks.

Structured data entry for reliable acquisition of pharmacokinetic data.

A pharmacokinetic database was constructed that is as free of errors as possible. Pharmacokinetic parameters were derived from the literature using a text-processing system and a database system. A random data sample from each system was compared with the original literature. The estimated error frequencies using statistical methods differed significantly between the two systems. The estimated error frequency in the text-processing system was 7.2%, that in the database system 2.7%. Compared with the original values in the literature, the estimated probability of error for identical pharmacokinetic parameters recorded in both systems is 2.4% and is not significantly different from the error frequency in the database. Parallel data entry with a text-processing system and a database system is, therefore, not significantly better than structured data entry for reducing the error frequency.

Statistical analysis of heterogeneous pharmacokinetic data from the literature.

To obtain the reliable pharmacokinetic quantities necessary for the adjustment of individual drug dosages, the pharmacokinetic data in the literature must be analysed by various appropriate statistical methods. Generally, pharmacokinetic quantities have been calculated from small sample sizes and published in different ways (the coefficient of variation of an estimated pharmacokinetic quantity is often higher than 40 percent). In the present study data were therefore structured and clustered according to the important influence variables. The important influence variables were obtained by known multivariate statistical methods. To obtain optimal estimators (uniformly minimum variance unbiased estimators) for the quantities inside the clusters, goodness of fit tests must be performed. Skewed distributions and samples with heterogeneous variances must be subjected to transformation procedures to obtain normally distributed values with homogeneous variances. The population-derived estimate of the a priori knowledge of pharmacokinetic quantities can be obtained by a statistical algorithm proposed in this paper. Information on the quantities half-life, total body clearance and volume of distribution of vancomycin were extracted from over 200 publications. These data were analysed with the proposed statistical algorithm, yielding population-based estimates for half-life, clearance and volume of vancomycin, taking into consideration the detected influence variables--renal function and the age of the patient. Important dependencies between different variables are discussed.

Unifying concept of pharmacokinetics derived from drug distribution and elimination in renal failure.

We are in the process of establishing a pharmacokinetic database for drug dosage adjustment to impaired renal function. To meet these demands, the system needs a unifying pharmacokinetic concept. Drug clearance is the common parameter that allows for intersecting the different pharmacokinetic approaches. In addition, two parameters are essentially needed for the concept, either the area-derived volume and the dominant elimination half-life or the moment-derived volume and the mean residence time. For drugs where the area-derived volume decreases with renal impairment, it can be shown that compartment-derived parameters are explicitly convertible into moment-derived parameters and vice versa.