Clustering patient mobility patterns to assess effectiveness of health-service delivery.

BACKGROUND: Analysis of patient mobility in a country not only gives an idea of how the health-care system works, but also can be a guideline to determine the quality of health care and health disparity among regions. Even though determination of patient movement is important, it is not often realized that patient mobility could have a unique pattern beyond health-related endowments (e.g., facilities, medical staff). This study therefore addresses the following research question: Is there a way to identify regions with similar patterns using spatio-temporal distribution of patient mobility? The aim of the paper is to answer this question and improve a classification method that is useful for populous countries like Turkey that have many administrative areas. METHODS: The data used in the study consist of spatio-temporal information on patient mobility for the period between 2009 and 2013. Patient mobility patterns based on the number of patients attracted/escaping across 81 provinces of Turkey are illustrated graphically. The hierarchical clustering method is used to group provinces in terms of the mobility characteristics revealed by the patterns. Clustered groups of provinces are analyzed using non-parametric statistical tests to identify potential correlations between clustered groups and the selected basic health indicators. RESULTS: Ineffective health-care delivery in certain regions of Turkey was determined through identifying patient mobility patterns. High escape values obtained for a large number of provinces suggest poor health-care accessibility. On the other hand, over the period of time studied, visualization of temporal mobility revealed a considerable decrease in the escape ratio for inadequately equipped provinces. Among four of twelve clusters created using the hierarchical clustering method, which include 64 of 81 Turkish provinces, there was a statistically significant relationship between the patterns and the selected basic health indicators of the clusters. The remaining eight clusters included 17 provinces and showed anomalies. CONCLUSIONS: The most important contribution of this study is the development of a way to identify patient mobility patterns by analyzing patient movements across the clusters. These results are strong evidence that patient mobility patterns provide a useful tool for decisions concerning the distribution of health-care services and the provision of health care equipment to the provinces.

Predicting early brain metastases based on clinicopathological factors and gene expression analysis in advanced HER2-positive breast cancer patients.

The overexpression or amplification of the human epidermal growth factor receptor 2 gene (HER2/neu) is associated with high risk of brain metastasis (BM). The identification of patients at highest immediate risk of BM could optimize screening and facilitate interventional trials. We performed gene expression analysis using complementary deoxyribonucleic acid-mediated annealing, selection, extension and ligation and real-time quantitative reverse transcription PCR (qRT-PCR) in primary tumor samples from two independent cohorts of advanced HER2 positive breast cancer patients. Additionally, we analyzed predictive relevance of clinicopathological factors in this series. Study group included discovery Cohort A (84 patients) and validation Cohort B (75 patients). The only independent variables associated with the development of early BM in both cohorts were the visceral location of first distant relapse [Cohort A: hazard ratio (HR) 7.4, 95 % CI 2.4-22.3; p < 0.001; Cohort B: HR 6.1, 95 % CI 1.5-25.6; p = 0.01] and the lack of trastuzumab administration in the metastatic setting (Cohort A: HR 5.0, 95 % CI 1.4-10.0; p = 0.009; Cohort B: HR 10.0, 95 % CI 2.0-100.0; p = 0.008). A profile including 13 genes was associated with early (≤36 months) symptomatic BM in the discovery cohort. This was refined by qRT-PCR to a 3-gene classifier (RAD51, HDGF, TPR) highly predictive of early BM (HR 5.3, 95 % CI 1.6-16.7; p = 0.005; multivariate analysis). However, predictive value of the classifier was not confirmed in the independent validation Cohort B. The presence of visceral metastases and the lack of trastuzumab administration in the metastatic setting apparently increase the likelihood of early BM in advanced HER2-positive breast cancer.

A non-parametric Bayesian model for joint cell clustering and cluster matching: identification of anomalous sample phenotypes with random effects.

BACKGROUND: Flow cytometry (FC)-based computer-aided diagnostics is an emerging technique utilizing modern multiparametric cytometry systems.The major difficulty in using machine-learning approaches for classification of FC data arises from limited access to a wide variety of anomalous samples for training. In consequence, any learning with an abundance of normal cases and a limited set of specific anomalous cases is biased towards the types of anomalies represented in the training set. Such models do not accurately identify anomalies, whether previously known or unknown, that may exist in future samples tested. Although one-class classifiers trained using only normal cases would avoid such a bias, robust sample characterization is critical for a generalizable model. Owing to sample heterogeneity and instrumental variability, arbitrary characterization of samples usually introduces feature noise that may lead to poor predictive performance. Herein, we present a non-parametric Bayesian algorithm called ASPIRE (anomalous sample phenotype identification with random effects) that identifies phenotypic differences across a batch of samples in the presence of random effects. Our approach involves simultaneous clustering of cellular measurements in individual samples and matching of discovered clusters across all samples in order to recover global clusters using probabilistic sampling techniques in a systematic way. RESULTS: We demonstrate the performance of the proposed method in identifying anomalous samples in two different FC data sets, one of which represents a set of samples including acute myeloid leukemia (AML) cases, and the other a generic 5-parameter peripheral-blood immunophenotyping. Results are evaluated in terms of the area under the receiver operating characteristics curve (AUC). ASPIRE achieved AUCs of 0.99 and 1.0 on the AML and generic blood immunophenotyping data sets, respectively. CONCLUSIONS: These results demonstrate that anomalous samples can be identified by ASPIRE with almost perfect accuracy without a priori access to samples of anomalous subtypes in the training set. The ASPIRE approach is unique in its ability to form generalizations regarding normal and anomalous states given only very weak assumptions regarding sample characteristics and origin. Thus, ASPIRE could become highly instrumental in providing unique insights about observed biological phenomena in the absence of full information about the investigated samples.

Biosorption of Cu(II) ions onto the litter of natural trembling poplar forest.

The litter of natural trembling poplar (Populus tremula) forest (LNTPF) was used for the biosorption of Cu(II) ions in a batch adsorption experiments. The sorption capacity of LNTPF was investigated as a function of pH, particle size, agitating speed, initial Cu(II) concentration, adsorbent concentration and temperature. The efficiency of copper uptake by the used LNTPF increases with a rise of solution pH, adsorbent concentration, agitating speed, temperature, and with a decline of particle size and initial Cu(II) concentration. The biosorption process was very fast; 94% of Cu(II) removal occurred within 5 min and equilibrium was reached at around 30 min. Batch adsorption models, based on the assumption of the pseudo-first order, pseudo-second order mechanism were applied to examine the adsorption kinetics. The pseudo-second order model was found to best fit the kinetic data. EPR studies combined with FTIR spectroscopy were used to represent the biosorption mechanism. Thermodynamic parameters such as DeltaH degrees, DeltaS degrees and DeltaG degrees were calculated. The adsorption process was found to be endothermic and spontaneous. Equilibrium data fitted well to Langmuir adsorption model. This study proved that the LNTPF can be used as an effective, cheap and abundant adsorbent for the treatment of Cu(II) containing wastewaters.

Photoacoustic tomography of intact human prostates and vascular texture analysis identify prostate cancer biopsy targets.

Prostate cancer is poorly visualized on ultrasonography (US) so that current biopsy requires either a templated technique or guidance after fusion of US with magnetic resonance imaging. Here we determined the ability for photoacoustic tomography (PAT) and US followed by texture-based image processing to identify prostate biopsy targets. K-means clustering feature learning and testing was performed on separate datasets comprised of 1064 and 1197 nm PAT and US images of intact, ex vivo human prostates. 1197 nm PAT was found to not contribute to the feature learning, and thus, only 1064 nm PAT and US images were used for final feature testing. Biopsy targets, determined by the tumor-assigned pixels' center of mass, located 100% of the primary lesions and 67% of the secondary lesions. In conclusion, 1064 nm PAT and US texture-based feature analysis provided successful prostate biopsy targets.

Automated Assessment of Disease Progression in Acute Myeloid Leukemia by Probabilistic Analysis of Flow Cytometry Data.

OBJECTIVE: Flow cytometry (FC) is a widely acknowledged technology in diagnosis of acute myeloid leukemia (AML) and has been indispensable in determining progression of the disease. Although FC plays a key role as a posttherapy prognosticator and evaluator of therapeutic efficacy, the manual analysis of cytometry data is a barrier to optimization of reproducibility and objectivity. This study investigates the utility of our recently introduced nonparametric Bayesian framework in accurately predicting the direction of change in disease progression in AML patients using FC data. METHODS: The highly flexible nonparametric Bayesian model based on the infinite mixture of infinite Gaussian mixtures is used for jointly modeling data from multiple FC samples to automatically identify functionally distinct cell populations and their local realizations. Phenotype vectors are obtained by characterizing each sample by the proportions of recovered cell populations, which are, in turn, used to predict the direction of change in disease progression for each patient. RESULTS: We used 200 diseased and nondiseased immunophenotypic panels for training and tested the system with 36 additional AML cases collected at multiple time points. The proposed framework identified the change in direction of disease progression with accuracies of 90% (nine out of ten) for relapsing cases and 100% (26 out of 26) for the remaining cases. CONCLUSIONS: We believe that these promising results are an important first step toward the development of automated predictive systems for disease monitoring and continuous response evaluation. SIGNIFICANCE: Automated measurement and monitoring of therapeutic response is critical not only for objective evaluation of disease status prognosis but also for timely assessment of treatment strategies.

Adsorption of chromium(VI) on pomace--an olive oil industry waste: batch and column studies.

The waste pomace of olive oil factory (WPOOF) was tested for its ability to remove chromium(VI) from aqueous solution by batch and column experiments. Various thermodynamic parameters, such as DeltaG degrees , DeltaH degrees and DeltaS degrees have been calculated. The thermodynamics of chromium(VI) ion onto WPOOF system indicates spontaneous and endothermic nature of the process. The ability of WPOOF to adsorb chromium(VI) in a fixed bed column was investigated, as well. The effect of operating parameters such as flow rate and inlet metal ion concentration on the sorption characteristics of WPOOF was investigated. The longest breakthrough time and maximum of Cr(VI) adsorption is obtained at pH 2.0. The total adsorbed quantities, equilibrium uptakes and total removal percents of chromium(VI) related to the effluent volumes were determined by evaluating the breakthrough curves obtained at different flow rates and different inlet chromium(VI) concentrations for adsorbent. The data confirmed that the total amount of sorbed chromium(VI) and equilibrium chromium(VI) uptake decreased with increasing flow rate and increased with increasing inlet chromium(VI) concentration. The Adams-Bohart model were used to analyze the experimental data and the model parameters were evaluated.

Towards large-scale histopathological image analysis: hashing-based image retrieval.

Automatic analysis of histopathological images has been widely utilized leveraging computational image-processing methods and modern machine learning techniques. Both computer-aided diagnosis (CAD) and content-based image-retrieval (CBIR) systems have been successfully developed for diagnosis, disease detection, and decision support in this area. Recently, with the ever-increasing amount of annotated medical data, large-scale and data-driven methods have emerged to offer a promise of bridging the semantic gap between images and diagnostic information. In this paper, we focus on developing scalable image-retrieval techniques to cope intelligently with massive histopathological images. Specifically, we present a supervised kernel hashing technique which leverages a small amount of supervised information in learning to compress a 10 000-dimensional image feature vector into only tens of binary bits with the informative signatures preserved. These binary codes are then indexed into a hash table that enables real-time retrieval of images in a large database. Critically, the supervised information is employed to bridge the semantic gap between low-level image features and high-level diagnostic information. We build a scalable image-retrieval framework based on the supervised hashing technique and validate its performance on several thousand histopathological images acquired from breast microscopic tissues. Extensive evaluations are carried out in terms of image classification (i.e., benign versus actionable categorization) and retrieval tests. Our framework achieves about 88.1% classification accuracy as well as promising time efficiency. For example, the framework can execute around 800 queries in only 0.01 s, comparing favorably with other commonly used dimensionality reduction and feature selection methods.

Batch Discovery of Recurring Rare Classes toward Identifying Anomalous Samples.

We present a clustering algorithm for discovering rare yet significant recurring classes across a batch of samples in the presence of random effects. We model each sample data by an infinite mixture of Dirichlet-process Gaussian-mixture models (DPMs) with each DPM representing the noisy realization of its corresponding class distribution in a given sample. We introduce dependencies across multiple samples by placing a global Dirichlet process prior over individual DPMs. This hierarchical prior introduces a sharing mechanism across samples and allows for identifying local realizations of classes across samples. We use collapsed Gibbs sampler for inference to recover local DPMs and identify their class associations. We demonstrate the utility of the proposed algorithm, processing a flow cytometry data set containing two extremely rare cell populations, and report results that significantly outperform competing techniques. The source code of the proposed algorithm is available on the web via the link: http://cs.iupui.edu/~dundar/aspire.htm.

Laser optical sensor, a label-free on-plate Salmonella enterica colony detection tool.

UNLABELLED: We investigated the application capabilities of a laser optical sensor, BARDOT (bacterial rapid detection using optical scatter technology) to generate differentiating scatter patterns for the 20 most frequently reported serovars of Salmonella enterica. Initially, the study tested the classification ability of BARDOT by using six Salmonella serovars grown on brain heart infusion, brilliant green, xylose lysine deoxycholate, and xylose lysine tergitol 4 (XLT4) agar plates. Highly accurate discrimination (95.9%) was obtained by using scatter signatures collected from colonies grown on XLT4. Further verification used a total of 36 serovars (the top 20 plus 16) comprising 123 strains with classification precision levels of 88 to 100%. The similarities between the optical phenotypes of strains analyzed by BARDOT were in general agreement with the genotypes analyzed by pulsed-field gel electrophoresis (PFGE). BARDOT was evaluated for the real-time detection and identification of Salmonella colonies grown from inoculated (1.2 × 10(2) CFU/30 g) peanut butter, chicken breast, and spinach or from naturally contaminated meat. After a sequential enrichment in buffered peptone water and modified Rappaport Vassiliadis broth for 4 h each, followed by growth on XLT4 (~16 h), BARDOT detected S. Typhimurium with 84% accuracy in 24 h, returning results comparable to those of the USDA Food Safety and Inspection Service method, which requires ~72 h. BARDOT also detected Salmonella (90 to 100% accuracy) in the presence of background microbiota from naturally contaminated meat, verified by 16S rRNA sequencing and PFGE. Prolonged residence (28 days) of Salmonella in peanut butter did not affect the bacterial ability to form colonies with consistent optical phenotypes. This study shows BARDOT's potential for nondestructive and high-throughput detection of Salmonella in food samples. IMPORTANCE: High-throughput screening of food products for pathogens would have a significant impact on the reduction of food-borne hazards. A laser optical sensor was developed to screen pathogen colonies on an agar plate instantly without damaging the colonies; this method aids in early pathogen detection by the classical microbiological culture-based method. Here we demonstrate that this sensor was able to detect the 36 Salmonella serovars tested, including the top 20 serovars, and to identify isolates of the top 8 Salmonella serovars. Furthermore, it can detect Salmonella in food samples in the presence of background microbiota in 24 h, whereas the standard USDA Food Safety and Inspection Service method requires about 72 h.

A Machine-Learning Approach to Detecting Unknown Bacterial Serovars.

Technologies for rapid detection of bacterial pathogens are crucial for securing the food supply. A light-scattering sensor recently developed for real-time identification of multiple colonies has shown great promise for distinguishing bacteria cultures. The classification approach currently used with this system relies on supervised learning. For accurate classification of bacterial pathogens, the training library should be exhaustive, i.e., should consist of samples of all possible pathogens. Yet, the sheer number of existing bacterial serovars and more importantly the effect of their high mutation rate would not allow for a practical and manageable training. In this study, we propose a Bayesian approach to learning with a nonexhaustive training dataset for automated detection of unmatched bacterial serovars, i.e., serovars for which no samples exist in the training library. The main contribution of our work is the Wishart conjugate priors defined over class distributions. This allows us to employ the prior information obtained from known classes to make inferences about unknown classes as well. By this means, we identify new classes of informational value and dynamically update the training dataset with these classes to make it increasingly more representative of the sample population. This results in a classifier with improved predictive performance for future samples. We evaluated our approach on a 28-class bacteria dataset and also on the benchmark 26-class letter recognition dataset for further validation. The proposed approach is compared against state-of-the-art involving density-based approaches and support vector domain description, as well as a recently introduced Bayesian approach based on simulated classes.

A two-level approach towards semantic colon segmentation: removing extra-colonic findings.

Computer aided detection (CAD) of colonic polyps in computed tomographic colonography has tremendously impacted colorectal cancer diagnosis using 3D medical imaging. It is a prerequisite for all CAD systems to extract the air-distended colon segments from 3D abdomen computed tomography scans. In this paper, we present a two-level statistical approach of first separating colon segments from small intestine, stomach and other extra-colonic parts by classification on a new geometric feature set; then evaluating the overall performance confidence using distance and geometry statistics over patients. The proposed method is fully automatic and validated using both the classification results in the first level and its numerical impacts on false positive reduction of extra-colonic findings in a CAD system. It shows superior performance than the state-of-art knowledge or anatomy based colon segmentation algorithms.