The Use of Combined Classifiers in Medical Functional Diagnostics

In otological and ophthalmological functional diagnostics it is typical to record for a certain individual several similar biopotentials under different conditions of stimulation and/or to collect various potentials for the same individual. Since the combination of these different data collected from a patient is of interest, coupling of the classification results obtained for each single data set seems desirable. Based on sample classifiers, obtained by linear discriminance analysis, of acoustically or visually evoked potentials from selected patients an algorithm is proposed for their coupling, which logically links the various rules of allocation and is based on the principle of assignment by highest likelihood. Furthermore, the procedure may be modified by inclusion of corresponding confidence intervals for the determined allocation frequencies. The results obtained for the groups of patients investigated show a considerable improvement of classification as judged from the error rates being distinctly reduced after the coupling.     

Optimal number of features as a function of sample size for various classification rules

MOTIVATION: Given the joint feature-label distribution, increasing the number of features always results in decreased classification error; however, this is not the case when a classifier is designed via a classification rule from sample data. Typically (but not always), for fixed sample size, the error of a designed classifier decreases and then increases as the number of features grows. The potential downside of using too many features is most critical for small samples, which are commonplace for gene-expression-based classifiers for phenotype discrimination. For fixed sample size and feature-label distribution, the issue is to find an optimal number of features. RESULTS: Since only in rare cases is there a known distribution of the error as a function of the number of features and sample size, this study employs simulation for various feature-label distributions and classification rules, and across a wide range of sample and feature-set sizes. To achieve the desired end, finding the optimal number of features as a function of sample size, it employs massively parallel computation. Seven classifiers are treated: 3-nearest-neighbor, Gaussian kernel, linear support vector machine, polynomial support vector machine, perceptron, regular histogram and linear discriminant analysis. Three Gaussian-based models are considered: linear, nonlinear and bimodal. In addition, real patient data from a large breast-cancer study is considered. To mitigate the combinatorial search for finding optimal feature sets, and to model the situation in which subsets of genes are co-regulated and correlation is internal to these subsets, we assume that the covariance matrix of the features is blocked, with each block corresponding to a group of correlated features. Altogether there are a large number of error surfaces for the many cases. These are provided in full on a companion website, which is meant to serve as resource for those working with small-sample classification. AVAILABILITY: For the companion website, please visit http://public.tgen.org/tamu/ofs/ CONTACT: e-dougherty@ee.tamu.edu.     

EEG correlates of the change in information processing strategy in visual imagery

Cnanger of the spatial organization of biopotentials (spatial synchronization and disorder, spectral power and coherence) were analyzed during mental creating of visual images from two simple elements: the angle the oblique line. With the transition from the first to the fourth task, the total number of used elements increased from the number suitable for simultaneous presentation and conscious processing (less than 7 +/- 2) to a much higher number. Changes in the characteristics of the spatial organization of biopotentials associated with the increase in the number of the elements can be explained by a change in the information processing strategy, transformation of information processing strategy, i.e. the transition from the left-hemispheric successive (conscious) analyses to the right-hemispheric simultaneous (unconscious) processing. It was shown that this change in the information processing strategy was accompanied by an increase in the index of spatial disorder sensitive to complicated nonlinear processes. Subjects were divided in two groups with different forms of the reorganization of interhemispheric and fronto-occipital relationships of biopotentials. These data are interpreted in terms of different involvement of the unconscious intellectual processes of different layers (subconsciousness and superconsciousness) in the change in information processing strategy.     

Classification with High‐Dimensional Genetic Data: Assigning Patients and Genetic Features to Known Classes

A major task in the statistical analysis of genetic data such as gene expressions and single nucleotide polymorphisms (SNPs) is to predict whether a patient has a certain disease, or from which of several known subtypes of a disease a patient suffers. A large number of discrimination methods have been proposed in the literature and have been applied to genetic data to tackle this task. In this paper, we give an overview on the most popular of these procedures in the analysis of genetic data. Moreover, we describe how these methods for supervised classification can be combined with variable selection approaches to reduce the number of genetic features from several thousands to as few as possible to form a concise classification rule. Finally, we show how the resulting statistical models can be validated. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Prediction of various grades of cervical neoplasia on plastic-embedded cytobrush samples. Discriminant analysis with qualitative and quantitative predictors.

The purpose of this study was to investigate whether discrimination into five groups of various grades of cervical preneoplasia and neoplasia is possible using discriminant analysis models. Data were analyzed for 242 cases diagnosed as either slight dysplasia (n = 50), moderate dysplasia (n = 50), severe dysplasia (n = 50), carcinoma in situ (n = 50) or invasive carcinoma (n = 42) and consisted of qualitative and quantitative features of cells derived from a repeat sample taken from the ectocervix as well as the endocervix using Cytobrushes. The samples were embedded in plastic, and thin sections were prepared, resulting in a monolayer of cut nuclei. The percentages of expected correct prediction were obtained by using 10,000 double cross-validation samples; the mean percentage of correct prediction into five groups using cross-validation was 65% (in the original analysis, 72%) and into two groups (dysplasia versus carcinoma in situ and invasive carcinoma) was 91% (93%). The results reflect group discrimination potential; we do not claim reliability of prediction for an individual patient. The patients were not a representative sample of the population; to investigate whether groups of patients could be discriminated on the basis of both qualitative and quantitative features, the data analyzed contain an almost equal number of observations in each of the five groups. The results indicate that features do not classify the cases in the same way; the discriminant analyses suggest that quantitative features play an important role in the discrimination of dysplasia from carcinoma cases, while the majority of the qualitative features are important in discrimination within the three dysplasia groups.     

The intercentral relations of the frontal cortex and limbic structures of the brain in dogs]

In 3 dogs with implanted electrodes, in conditioned experiments correlation of the bioelectrical processes was studied by coherence function calculation of the hippocampus, hypothalamus, amygdala and frontal cortex biopotentials. It was shown, that the level of maximum values of coherence function of bioelectrical oscillations, led from various pairs of the studied brain structures significantly differed both in magnitude and frequency at which the greatest synchronization of biopotentials was noticed. In one dog with a high degree of connection between the hippocampus and hypothalamus biopotentials oscillations, a low synchronization of the frontal cortex and amygdala oscillations was found; in two other animals with a higher level of coherence between the oscillations of the frontal cortex and amygdala biopotentials, a lower degree of connection between the oscillations led from the hippocampus and hypothalamus was revealed. Synchronization of the biopotentials of the hippocampus and frontal cortex and also of the hippocampus and amygdala biopotentials proved to be low in all experimental dogs, what additionally testifies to different role of these structures in organization of the behaviour.     

Molecular diagnostics of clinically important staphylococci

Bacterial species of the genus Staphylococcus known as important human and animal pathogens are the cause of a number of severe infectious diseases. Apart from the major pathogen Staphylococcus aureus, other species until recently considered to be nonpathogenic may also be involved in serious infections. Rapid and accurate identification of the disease-causing agent is therefore prerequisite for disease control and epidemiological surveillance. Modern methods for identification and typing of bacterial species are based on genome analysis and have many advantages compared to phenotypic methods. The genotypic methods currently used in molecular diagnostics of staphylococcal species, particularly of S. aureus, are reviewed. Attention is also paid to new molecular methods with the highest discriminatory power. Efforts made to achieve interlaboratory reproducibility of diagnostic methods are presented.     

On Model-Choice in Discrimination with Categorical Variables

An algorithm for model selection in discrimination with categorical variables is presented. It is based on four models applied hierarchically and linked with a build-up procedure of feature-selection. The choice of models and features is ensured by a consequent cross-validation. Results of an application in medical diagnostics are described.     

Morphohistochemical and electrophysiological characteristics of the transverse striated muscles of the rabbit esophagus

A specific histochemical profile of the muscular tunic of the stomach has been revealed; its myons demonstrate a high succinate dehydrogenase and myofibrillar ATPase activity. Electrical activity of the oesophageal musculature has common features with biopotentials of the skeletal muscles and the smooth musculature of the stomach. The similiarity with the skeletal muscle biopotentials is evident as a peculiar spike activity, especially at the time of feeding. In electrograms of the striated oesophageal muscles slow waves are seen; they change at deprivation and feeding. Therefore, it is possible to compare them with the smooth musculature capable to generate slow waves.     

Massive parallele Sequenzierung in der Diagnostik hereditärer BRCA1-/-2-Mutationen

During the past few years, application of massively-parallel sequencing (MPS) in molecular diagnostics of hereditary cancer has increased significantly. The potential advantages of MPS, compared for example to Sanger sequencing-based methods, are higher sample capacities, shorter turnaround times, and decreased costs. Adoption in routine diagnostics is simplified due to the availability of certified kits for sample preparation. Here, the validation of two MPS systems for routine BRCA1/2 sequencing diagnostics are describe. Users should keep in mind the technical advantages and disadvantages of the individual sequencing machines. The results achieved with MPS are equal to those from Sanger sequencing.     

Rectal Swabs for Analysis of the Intestinal Microbiota

The composition of the gut microbiota is associated with various disease states, most notably inflammatory bowel disease, obesity and malnutrition. This underlines that analysis of intestinal microbiota is potentially an interesting target for clinical diagnostics. Currently, the most commonly used sample types are feces and mucosal biopsy specimens. Because sampling method, storage and processing of samples impact microbiota analysis, each sample type has its own limitations. An ideal sample type for use in routine diagnostics should be easy to obtain in a standardized fashion without perturbation of the microbiota. Rectal swabs may satisfy these criteria, but little is known about microbiota analysis on these sample types. In this study we investigated the characteristics and applicability of rectal swabs for gut microbiota profiling in a clinical routine setting in patients presenting with various gastro-intestinal disorders. We found that rectal swabs appeared to be a convenient means of sampling the human gut microbiota. Swabs can be performed on demand, whenever a patient presents; swab-derived microbiota profiles are reproducible, whether they are gathered at home by patients or by medical professionals in an outpatient setting and may be ideally suited for clinical diagnostics and large-scale studies.     

Fluorescence spectroscopy as a potential metabonomic tool for early detection of colorectal cancer

Fluorescence spectroscopy Excitation Emission Matrix (EEM) measurements were applied on human blood plasma samples from a case control study on colorectal cancer. Samples were collected before large bowel endoscopy and included patients with colorectal cancer or with adenomas, and from individuals with other non malignant findings or no findings (N = 308). The objective of the study was to explore the possibilities for applying fluorescence spectroscopy as a tool for detection of colorectal cancer. Parallel Factor Analysis (PARAFAC) was applied to decompose the fluorescence EEMs into estimates of the underlying fluorophores in the sample. Both the pooled score matrix from PARAFAC, holding the relative concentrations of the derived components, and the raw unfolded spectra were used as basis for discrimination models between cancer and the various controls. Both methods gave test set validated sensitivity and specificity values around 0.75 between cancer and controls, and poor discriminations between the various controls. The PARAFAC solution gave better options for analyzing the chemical mechanisms behind the discrimination, and revealed a blue shift in tryptophan emission in the cancer patients, a result that supports previous findings. The present findings show how fluorescence spectroscopy and chemometrics can help in cancer diagnostics, and with PARAFAC fluorescence spectroscopy can be a potential metabonomic tool.     

A coupling procedure for the discrimination of mixed data.

A procedure is described for coupling different discriminators to a new (common) decision rule using the corresponding allocation vectors only. This enables one to cope jointly with data of different structure and/or scales of measurement but without strong restrictions on the number of (especially categorical) features. The method is combined with a consequent cross-validation process securing the results reached. Examples from medical diagnostics demonstrate the usefulness of the proposed procedure, especially in comparison with the known linear discriminant analysis as judged from the error rates obtained.     

Streitkultur and the governance of genetic testing and insurance in Germany

Rapid developments in genetic testing have given rise to fundamental ethical, legal, and social questions that need to be dealt with in society. Results of genetic tests may be of interest to third parties such as private insurance companies, leading to fears of genetic discrimination. In Germany, the Government adopted the Genetic Diagnosis Act (Gendiagnostikgesetz, GenDG) in 2009 to protect people from, inter alia, genetic discrimination in obtaining life or health insurance. Given the sensitivity of the topic, this legislation was continually revised between 2001 and 2009. In this article, we reconstruct the process of formulating the GenDG with regard to genetics and insurance. The article begins with the parliamentary Enquete Commission in 2000 to develop a strategy and recommendations for the governance of genetic diagnostics, and analyzes how these recommendations were applied during the legislative process. We demonstrate that the legislative process of GenDG was largely determined by conventional methods of governance, rather than Streitkultur called for by the Enquete Commission in 2002. We conclude that though Streitkultur was defined as a mechanism to develop a robust approach to the governance of genetic diagnostics, it failed to influence a crucial element in genetic testing and insurance; namely, to fully protect insurees from genetic discrimination.     

Properties of Markov Chain Monte Carlo Performance across Many Empirical Alignments

Nearly all current Bayesian phylogenetic applications rely on Markov chain Monte Carlo (MCMC) methods to approximate the posterior distribution for trees and other parameters of the model. These approximations are only reliable if Markov chains adequately converge and sample from the joint posterior distribution. Although several studies of phylogenetic MCMC convergence exist, these have focused on simulated data sets or select empirical examples. Therefore, much that is considered common knowledge about MCMC in empirical systems derives from a relatively small family of analyses under ideal conditions. To address this, we present an overview of commonly applied phylogenetic MCMC diagnostics and an assessment of patterns of these diagnostics across more than 18,000 empirical analyses. Many analyses appeared to perform well and failures in convergence were most likely to be detected using the average standard deviation of split frequencies, a diagnostic that compares topologies among independent chains. Different diagnostics yielded different information about failed convergence, demonstrating that multiple diagnostics must be employed to reliably detect problems. The number of taxa and average branch lengths in analyses have clear impacts on MCMC performance, with more taxa and shorter branches leading to more difficult convergence. We show that the usage of models that include both Γ-distributed among-site rate variation and a proportion of invariable sites is not broadly problematic for MCMC convergence but is also unnecessary. Changes to heating and the usage of model-averaged substitution models can both offer improved convergence in some cases, but neither are a panacea.     

Automatic tumor-stroma separation in fluorescence TMAs enables the quantitative high-throughput analysis of multiple cancer biomarkers

The upcoming quantification and automation in biomarker based histological tumor evaluation will require computational methods capable of automatically identifying tumor areas and differentiating them from the stroma. As no single generally applicable tumor biomarker is available, pathology routinely uses morphological criteria as a spatial reference system. We here present and evaluate a method capable of performing the classification in immunofluorescence histological slides solely using a DAPI background stain. Due to the restriction to a single color channel this is inherently challenging. We formed cell graphs based on the topological distribution of the tissue cell nuclei and extracted the corresponding graph features. By using topological, morphological and intensity based features we could systematically quantify and compare the discrimination capability individual features contribute to the overall algorithm. We here show that when classifying fluorescence tissue slides in the DAPI channel, morphological and intensity based features clearly outpace topological ones which have been used exclusively in related previous approaches. We assembled the 15 best features to train a support vector machine based on Keratin stained tumor areas. On a test set of TMAs with 210 cores of triple negative breast cancers our classifier was able to distinguish between tumor and stroma tissue with a total overall accuracy of 88%. Our method yields first results on the discrimination capability of features groups which is essential for an automated tumor diagnostics. Also, it provides an objective spatial reference system for the multiplex analysis of biomarkers in fluorescence immunohistochemistry.     

Comparison of various procedures for estimation of the classification error in discriminance analysis

The practical use of classification methods depends, for the majority of cases, on whether it is possible to obtain evidence how newly appearing objects of unknown classes-membership can be allocated on the basis of a classifier established by a training sample. In the present paper several error estimation methods are presented and compared with each other. It is shown by means of a practical example in which way the number of characteristics used in the discrimination affects the classification error. Discriminance functions calculated from many characteristics yield a higher error rate than those from fewer characteristics. From the comparison of the error rates and computing times involved in the individual methods recommendations for the selection of practicable estimation methods are given.     

Pain Intensity Recognition Rates via Biopotential Feature Patterns with Support Vector Machines

Background

The clinically used methods of pain diagnosis do not allow for objective and robust measurement, and physicians must rely on the patient’s report on the pain sensation. Verbal scales, visual analog scales (VAS) or numeric rating scales (NRS) count among the most common tools, which are restricted to patients with normal mental abilities. There also exist instruments for pain assessment in people with verbal and / or cognitive impairments and instruments for pain assessment in people who are sedated and automated ventilated. However, all these diagnostic methods either have limited reliability and validity or are very time-consuming. In contrast, biopotentials can be automatically analyzed with machine learning algorithms to provide a surrogate measure of pain intensity.

Methods

In this context, we created a database of biopotentials to advance an automated pain recognition system, determine its theoretical testing quality, and optimize its performance. Eighty-five participants were subjected to painful heat stimuli (baseline, pain threshold, two intermediate thresholds, and pain tolerance threshold) under controlled conditions and the signals of electromyography, skin conductance level, and electrocardiography were collected. A total of 159 features were extracted from the mathematical groupings of amplitude, frequency, stationarity, entropy, linearity, variability, and similarity.

Results

We achieved classification rates of 90.94% for baseline vs. pain tolerance threshold and 79.29% for baseline vs. pain threshold. The most selected pain features stemmed from the amplitude and similarity group and were derived from facial electromyography.

Conclusion

The machine learning measurement of pain in patients could provide valuable information for a clinical team and thus support the treatment assessment.     

Erratum to: Fluorescence spectroscopy as a potential metabonomic tool for early detection of colorectal cancer

Fluorescence spectroscopy Excitation Emission Matrix (EEM) measurements were applied on human blood plasma samples from a case control study on colorectal cancer. Samples were collected before large bowel endoscopy and included patients with colorectal cancer or with adenomas, and from individuals with other non malignant findings or no findings (N = 308). The objective of the study was to explore the possibilities for applying fluorescence spectroscopy as a tool for detection of colorectal cancer. Parallel Factor Analysis (PARAFAC) was applied to decompose the fluorescence EEMs into estimates of the underlying fluorophores in the sample. Both the pooled score matrix from PARAFAC, holding the relative concentrations of the derived components, and the raw unfolded spectra were used as basis for discrimination models between cancer and the various controls. Both methods gave test set validated sensitivity and specificity values around 0.75 between cancer and controls, and poor discriminations between the various controls. The PARAFAC solution gave better options for analyzing the chemical mechanisms behind the discrimination, and revealed a blue shift in tryptophan emission in the cancer patients, a result that supports previous findings. The present findings show how fluorescence spectroscopy and chemometrics can help in cancer diagnostics, and with PARAFAC fluorescence spectroscopy can be a potential metabonomic tool.