Search

[ES] Hoy en día, el bienestar de la sociedad es una prioridad, y parte de este bienestar viene dado por un servicio esencial como es el acceso al agua, para lo cual se requiere una red de tuberías a lo largo de las ciudades, algunas de las cuales se encuentran bajo sus carreteras. Este Trabajo de Fin de Máster considera los datos de una ciudad del Mediterráneo con una población de tamaño medio. Esta ciudad requiere un suministro regular, para lo cual el mantenimiento de las tuberías es de gran importancia, ya que un fallo causaría molestias a los habitantes y pérdidas económicas. Este estudio propone aplicar modelos de supervivencia clásicos y de machine learning, que permiten no sólo predecir sino también aportar información sobre las características que provocan las roturas de las tuberías. La implementación y comparación de los modelos se ha realizado con el software libre R. Una vez ajustados los modelos, se validaron mediante las respectivas matrices de confusión y el área bajo la curva ROC (AUC). El mejor modelo obtenido fue el modelo Random Survival Forest (RSF) con un AUC de 0,85 y una precisión del 84%, seguido del modelo lineal generalizado clásico (GLM) con un AUC y una precisión de 0,83 y 77%, respectivamente., [EN] Nowadays, the welfare of society is a priority, and part of this welfare is given by an essential service such as access to water, which requires a network of pipes along the cities, some of which are located under their roads. This Master's thesis considers data from a Mediterranean city with a medium-sized population. This city requires a regular supply, for which the maintenance of the pipelines is of great importance, since a failure would cause inconvenience to the inhabitants and economic losses. This study proposes to apply classical and machine learning survival models, which allow not only to predict but also to provide information about the characteristics that cause pipe breaks. The implementation and comparison of the models was carried out with the free software R. Once the models were adjusted, they were validated by means of the respective confusion matrices and the area under the ROC curve (AUC). The best model obtained was the Random Survival Forest (RSF) model with an AUC of 0.85 and an accuracy of 84%, followed by the classical generalized linear model (GLM) with an AUC and accuracy of 0.83 and 77%, respectively.

International audience; Cetacean Distribution Modeling (CDM) is used to quantify mobile marine species distributions and densities. It is essential to better understand and protect whales and their relatives. Current CDM approaches often fail in capturing general species-environment relationships, which would be valid within a broader range of environmental conditions that characterize the surveyed regions. This paper aims at investigating the usefulness of deep learning based schemes, namely multi-task and transfer learning, in CDM. Co-training of a stochastic presence-background model on a classification task and a deterministic rule-based model on a regression task was performed. Whale presence-only records were used for the first task, and index outputs of a feeding habitat occurrence model for the second one. This new approach has been experimented through the study case of fin whales in the western Mediterranean Sea. To evaluate our approach, a new metric called True Positive rate per unit of Surface Habitat (TPSH) and an original multimodal fully-connected neural networks were developed. A Generalized Additive Model (GAM)—a standard CDM method—was also used as a reference for performance. Results show that our multi-task learning model improves both the feeding habitat model by 10.8% and data-driven models such as GAM by 16.5% on our TPSH metric in relative terms, revealing a higher accuracy of our approach in estimating whale presence. Such trends in results have been further supported by the use of two other independent datasets that forced models to generalize beyond their training dataset of species-environment relationships. Performance could be further improved by adopting more optimal thresholds as observed from Receiver Operating Characteristic curves, e.g. the multi-task learning model could reach absolute gains up to 10% in the median of the True Positive Rate while maintaining its habitat spatial spreading. Globally, our work confirmed our working hypothesis that expert information on whale behaviour represent a good knowledge base for model generalization. This result can be further improved by a concurrent learning of more local species-environment relationships from in-situ presence data.

Although revascularization rates after endovascular thrombectomy for large vessel acute ischemic stroke (AIS) are high (71%), only 46% of patients achieve functional independence at 90 days. The present study was designed to explore a new method for predicting the functional prognosis of AIS patients after endovascular recanalization. A total of 200 anterior circulation stroke patients who received endovascular therapy were enrolled. Logistic regression analysis of clinical characteristics on functional independence were performed. The predictive power of sub-items in National Institute of Health stroke scale (NIHSS) and the combination of NIHSS consciousness and Alberta Stroke Program Early CT Score (ASPECTS) on functional independence were assessed by Receiver Operating Characteristic (ROC) curves and the latter was compared with 3 previously published prediction models by AUC (the area under ROC curve). The AUC for the NIHSS consciousness score to predict functional independence was higher than whole NIHSS and other sub-items (0.716 v 0.705, 0.586, 0.573, 0.552 and 0.559). Low NIHSS consciousness score, high ASPECTS score, short time from onset to recanalization, and high rate of successful recanalization were demonstrated to be significantly associated with the functional independence (OR 0.697, 2.226, 0.994 and 28.643). The prediction power of the combination was significantly better than NIHSS and ASPECTS alone (AUC 0.793 v 0.705 and 0.752). Compared with 3 other prediction models, the combination was found to be the strongest predictor for functional independence (AUC 0.793 v 0.791, 0.671 and 0.564). NIHSS which has been shown to be a strong predictor of functional outcomes after endovascular recanalization is largely dependent on the consciousness component. NIHSS consciousness score combined with ASPECTS appears to be a favorable predictor of functional independence. These findings may have broad reaching effects for isolated centers around the world without advanced imaging for triage and prognostication.

In many real-world settings, individuals rarely present another person’s ID, which increases the likelihood that a screener will fail to detect it. Three experiments examined how within-person variability (i.e., differences between two images of the same person) and feedback may have influenced criterion shifting, thought to be one of the sources of the low-prevalence effect (LPE). Participants made identity judgments of a target face and an ID under either high, medium, or low mismatch prevalence. Feedback appeared after every trial, only error trials, or no trials. Experiment 1 used two controlled images taken on the same day. Experiment 2 used two controlled images taken at least 6 months apart. Experiment 3 used one controlled and one ambient image taken at least 1 year apart. Importantly, receiver operating characteristic curves revealed that feedback and greater within-person variability exacerbated the LPE by affecting both criterion and discriminability. These results carry implications for many real-world settings, such as border crossings and airports, where identity screening plays a major role in securing public safety.

Receiver operating characteristic (ROC) curves are an established method for assessing the predictive capacity of a continuous biomarker for a binary outcome. However, in some cases, outcomes are time dependent. Although the literature has proposed packages for performing ROC analysis of time-independent outcomes, a package is not yet available for analyzing the predictive capacity of continuous biomarkers when the binary outcome is time dependent. In this article, we present stroccurve, a new command for performing ROC analysis within a survival framework.

Due to the critical role of inflammation in nasopharyngeal carcinoma, we aim to investigate the correlation between nasopharyngeal carcinoma prognosis and the levels of tumor necrosis factor α and macrophages for the development of new prognostic models. The levels of tumor necrosis factor-α and CD68-positive macrophages were measured in 111 primary nasopharyngeal carcinoma specimens by immunohistochemistry. Kaplan-Meier analysis showed that, compared with nonelevated tumor necrosis factor-α levels, elevated tumor necrosis factor α levels were correlated with poorer 10-year distant metastasis-free survival (24.5% vs 5.2%, P = .004) and bone metastasis-free survival (17.0% vs 0.0%, P = .001). Multivariate analysis revealed that tumor necrosis factor α level was an independent prognostic factor for distant metastasis-free survival (hazard ratio = 16.765, P = .001), while the level of CD68-positive macrophages was a favorable independent prognostic factor for cancer-specific survival (hazard ratio = 0.481, P = .023) and disease-free survival (hazard ratio = 0.403, P = .010). Additionally, several prognostic models that considered tumor-node-metastasis stage alone or in combination with tumor necrosis factor α and/or CD68-positive macrophage levels were compared by receiver operating characteristic curve analysis. Interestingly, the T_score model, which considered the tumor necrosis factor α level alone, could better predict the distant metastasis-free survival and bone metastasis-free survival, whereas the MT model, which considered the combination of T stage and CD68-positive macrophage level, could better predict the cancer-specific survival and disease-free survival of patients with nasopharyngeal carcinoma. Elevated tumor necrosis factor-α levels and decreased CD68-positive macrophage levels in primary nasopharyngeal carcinoma tissues are unfavorable prognostic indicators in nasopharyngeal carcinoma. The T_score model or the MT model could be better prognostic models than those currently available for nasopharyngeal carcinoma and could be used to select high-risk patients and aid in the design of individualized immunotherapy.

Controlling an stochastic nonlinear system with a small amplitude signal is a fundamental problem with many practical applications. Quantifying locking is challenging, and current methods, such as spectral or correlation analysis, do not provide a precise measure of the degree of locking. Here we study locking in an experimental system, consisting of a semiconductor laser with optical feedback operated in the regime where it randomly emits abrupt spikes. To quantify the locking of the optical spikes to small electric perturbations, we use two measures, the success rate (SR) and the false positive rate (FPR). The SR counts the spikes that are emitted shortly after each perturbation, while the FPR counts the additional extra spikes. We show that the receiver operating characteristic (ROC) curve (SR versus FPR plot) uncovers parameter regions where the electric perturbations fully control the laser spikes, such that the laser emits, shortly after each perturbation, one and only one spike. To demonstrate the general applicability of the ROC analysis we also study a stochastic bistable system under square-wave forcing and show that the ROC curve allows identifying the parameters that produce best locking.

Introduction: Posterior Urethral Valves (PUVs) are a major cause of short and long term morbidity in male children. They cause recurrent urinary tract infections and lead to chronic renal failure in a significant proportion of boys. The diagnosis of posterior urethral valves has historically been based off voiding cystourethrograms (VCUGs) and confirmed and treated by cystoscopy, which is the gold standard. To the best of our knowledge the literature is sparse on the role of definitive diagnosis of PUV based on VCUG and no data on the urethral ratios in the normal population. Aim: To determine the ratio of posterior to anterior urethral diameter as an objective means to predict presence of Posterior Urethral Valves (PUV) using cystoscopy as the gold standard and to determine normal urethral ratios in paediatric age groups. Materials and Methods: The study was conducted prospectively in 493 male children over 18 months. Urethral diameters were measured on VCUG and Posterior Urethra/Anterior Urethra (PU/ AU) ratios derived. Cystoscopy was the reference standard. Diagnostic accuracy of ratios was calculated using Receiver Operating Characteristic (ROC) curves, best threshold values determined with corresponding estimates of sensitivity and specificity to decide optimal significant ratios. Age stratified normative ratios were calculated. Results: MCU has 97.96% sensitivity, 86.27% specificity, 87.27% positive predictive value (PPV) and 97.78% negative predictive value (NPV) value when compared with cystoscopy which is the current gold standard in the diagnosis of PUV. The Posterior Urethra/Anterior Urethra (PU/AU) ratio was found to be a valid indicator of the presence of PUV. The diagnostic ability of PU/average AU is the best of the three ratios (AUC 0.9055) in predicting the presence of a PUV on cystoscopy. PU/average AU value ≥2.11 suggests that the patient most likely has a PUV with an 83% sensitivity and 91% specificity, within 95% Confidence Interval of 0.8458-0.96527. With regard to calculation of the normal PU/AU ratio for each age group, we found a general trend of decreasing ratios with increasing patient age. Conclusion: VCUG is useful in predicting presence of PUV. Age group specific standard reference ratios can be used for identifying PUVs.

Early Warning Systems (EWS) and Early Warning Indictors (EWI) have recently emerged as an attractive domain for states and school districts interested in predicting student outcomes using data that schools already collect with the intention to better time and tailor interventions. However, current diagnostic measures used across the domain do not consider the dual issues of sensitivity and specificity of predictors, key components for considering accuracy. We apply signal detection theory using Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) analysis adapted from the engineering and medical domains, and using the pROC package in R. Using nationally generalizable data from the Education Longitudinal Study of 2002 (ELS:2002) we provide examples of applying ROC accuracy analysis to a variety of predictors of student outcomes, such as dropping out of high school, college enrollment, and postsecondary STEM degrees and careers. Keywords: ROC, AUC, Early Warning System, Early Warning Indicator, signal detection theory, dropout, college enrollment, Postsecondary STEM Degree, hard STEM career, soft STEM career

We designed and jointly optimized an integrated signal processing chain for detection and classification of long-range passive-acoustic underwater signals recorded by the global geophysical monitoring network of the Comprehensive Nuclear-Test-Ban Treaty Organization. Starting at the level of raw waveform data, a processing chain of signal detection, feature extraction, and signal classification was designed and jointly optimized to the task. Relevant waveform segments were in a first step identified by a generic, flexibly parameterized detection algorithm on a long-to short-term averages' ratio of the spectral energy. For representation, general-purpose sound processing features, with an added focus on spectral and cepstral features, were extracted from the detected segments. As classifiers, support vector machines with different kernel functions were employed alongside other baseline learning algorithms. The free parameters of the overall toolchain (i.e., trigger algorithm parameters and classifier hyperparameters) were jointly optimized in a cross-validation setting, either according to the cross-validation classification error or the cross-validation area under the receiver operating characteristic curve. Experiments demonstrate that our method outperforms machine learning algorithms task-tailored to a previous, human-expert-designed preprocessing chain. The presented approach can be adapted to a wide range of problems that can benefit from jointly optimizing parameters of preprocessing and classification algorithm. © 2016 IEEE.

Athletes may suffer from low back pain (LBP). Some studies have considered restricted range of motion (ROM) as a risk factor for LBP in athletes, however, ROM has not been evaluated in equestrian athletes (EAs) with LBP. The aims of this research in EAs were (I) to analyze the relationship between lower limb ROM (tightness and asymmetry) and LBP and (II) to determine the reference values for lower limb ROM indicating high risk of LBP. Forty-three young (9&ndash, 18 years old) EAs were voluntarily recruited in technical training camps. The EAs were assessed for 11 passive lower limb ROMs using the ROM-SPORT I Battery. LBP data were taken at the end of a 1-year period through a questionnaire. Binary logistic regression and receiver operating characteristic curves were calculated. Sixty-seven percent of EA had suffered LBP. Hip adduction (OR = 1.347, medium, p = 0.015) and knee flexion (OR = 1.109, small, p = 0.023) were predictors of LBP in EAs. A hip adduction of 26&deg, and knee flexion of 128&deg, were found to be the optimal cutoff values for predictive screening of EAs at high risk of LBP.

This paper addresses the use of a data analysis tool, known as robust principal component analysis (RPCA), in the context of change detection (CD) in ultrawideband (UWB) very high-frequency (VHF) synthetic aperture radar (SAR) images. The method considers image pairs of the same scene acquired at different time instants. The CD method aims to maximize the probability of detection (PD) and minimize the false alarm rate (FAR). Such aim fits into a multiobjective optimization problem, since maximizing the probability of detection generally implies an increase in the number of false alarms. In that sense, varying the RPCA regularization parameter leads to PD variation with respect to FAR, which is known as receiver operating characteristic (ROC) curve. To evaluate the proposed method, the CARABAS-II data set was considered. The experimental results show that RPCA via principal component pursuit (PCP) can provide a good trade-off between PD and FAR. A comparison between the results obtained with the proposed method and a classical CD algorithm based on the likelihood ratio test provides the pros and cons of the proposed method. © 2020 by the authors. Open access

Introduction A picture version of the Free and Cued Selective Reminding Test (FCSRT) would assist in the assessment of memory function in patients with low levels of schooling. A shortened version would improve the test's applicability. Objectives To analyse the diagnostic usefulness of a shortened picture version of the FCSRT for distinguishing patients with amnestic mild cognitive impairment (aMCI) from controls, without excluding participants with a low level of schooling. Methods Phase I study of a diagnostic evaluation (convenience sampling; pre-test prevalence 50%). A blinded researcher independently administered the FCSRT to 30 patients with aMCI and 30 controls matched for age, sex, level of schooling and literacy, using images and omitting the usual 30-minutes delayed recall item. Three variables were recorded: free recall, total recall, and cue efficiency. Diagnostic accuracy was calculated using receiver operating characteristic curves and the area under the curve. The Youden index was used to identify optimal cut-off points. Results Of all participants, 41.7% had not completed primary education. There were no differences between groups as regards sociodemographic variables. Area under the curve was excellent for free recall (0.99), total recall (0.95), and cue efficiency (0.93). The optimal cut-off points were 21/22, 43/44, and Conclusions This preliminary analysis shows that a shortened picture version of the FCSRT may be useful and applicable for the diagnosis of aMCI without excluding individuals with a low level of schooling.

I assess and forecast the probability of deflation in the EA at different horizons using a binomial probit model. I select the best predictors among more than one-hundred variables adopting a two-step combinatoric approach and exploiting parallel computation in Julia language. I show that the best-selected variables coincide to those standardly included in a small New Keynesian model. Also, I assess the goodness of the models using three different loss functions: the Mean Absolute Error (MAE), the Root Mean Squared Error (RMSE) and the Area Under the Receiver Operating Characteristics (AUROC). The results are reasonably consistent among the three criteria. Finally, I compute an index averaging the forecasts to assess the probability of being in a deflation state in the next two years. The index shows that having inflation above the 2% level before March 2019 is extremely unlikely., N/A

Introduction. Several medical physicists have moved to a statistical process control-oriented approach to learn more about the intrinsic performance of radiotherapy equipment. Our aim is to report how gauge repeatability and reproducibility (R&R) and sensitivity and specificity studies can provide optimal specifications for radiotherapy-beam calibration checks. Methods. A team of three medical physicists performed gauge R&R studies on 6 megavolt (MV) photon, and 4, 9, and 15-megaelectron volt electron-beams. The operator order was randomized to reduce time-related biases. Checks were performed accurately using ionization chambers with a water phantom. These gauge R&R studies allowed us to determine the proportion of variation due to the repeatability, operator, and beam-to-beam components associated with repeated beam-calibration checking. We then calculated the conditional probabilities of misclassifying the check results as a false accept (beta = 1-sensitivity) and a false reject (delta = 1-specificity) by changing the beam calibration specifications. Using this information we plotted the receiver operating characteristic curves in order to identify the sensitivity and specificity settings that maximize detection. Results. The main component of variation was due to the operator; the repeatability component was less than 30%. The intrinsic variation component for 6 MV photon-beams was approximately 20%, whereas it was over 34% for electron-beams. The optimal specification bands expressed as a percentage of the mean were +/- 0.7% for 6 MV photons and +/- 0.9% for electron-beams. Optimal sensitivity was over 50%, whereas specificity was not less than 0.995. The very low d indicates that the system is unlikely to reject a correct beam calibration. However, the high beta shows that should we require high levels of precision and therefore adopt such optimized specifications, the probability of accepting a drifted beam calibration would be fairly high. Therefore, the combination of 'medical physicist, water phantom, and ionization chamber' is an excellent tool for correctly adjusting beam calibrations to their specifications; moreover, this procedure could also be further improved in order to detect drifts in beam calibrations if the specifications required were to tighten in the future. Conclusion. Calculating the optimal specifications is feasible and leads to a very high detection specificity, but with low sensitivity. The operator-associated variation component contributed the most to the overall variation in check results. The beam calibration procedure might need to be adapted if the demand for precision in radiotherapy increases.

The Hand Arm Risk assessment Method (HARM) is a simplified risk assessment method for determining musculoskeletal symptoms to the arm, neck and/or shoulder posed by hand-arm tasks of the upper body. The purpose of this study was to evaluate the predictive validity of HARM using data collected from a three year prospective cohort Study on Musculoskeletal disorders, Absenteeism and Health (SMASH). Structured observations of video recordings using HARM were made of 88 participants performing hand-arm tasks. These video recordings were made at baseline of the SMASH study, whereas self-reported musculoskeletal symptoms were assessed both at baseline and during a three year follow-up. The receiver operating characteristic (ROC) curves showed an area under the curve (AUC) of 0.75 (CI: 0.65-0.86) for symptoms at baseline and 0.70 (CI: 0.58-0.83) for symptoms during follow-up. For a HARM score of 50, the sensitivity and specificity were 64% and 79% at baseline and 53% and 82% during follow up. Participants with a HARM score of 50 or more (considered a high risk) had a significantly higher risk of upper limb symptoms than those with a HARM score of less than 50 at baseline (OR=5.31; 95%CI: 2.10-13.39) and 3 year follow-up (OR=5.11; (95%CI: 1.61-16.27). This study has shown that HARM can be used to assess work tasks of the hands and arms and discriminate between tasks that pose a high or low risk of neck or upper limb symptoms. Relevance to industry: The association of HARM results and self-reported neck and upper limb symptoms was studied. The association found both cross sectionally and at follow up shows that the method can be used in industry to distinguish hazardous hand arm tasks from non-hazardous ones. Thus, HARM can support OSH practitioners and industrial engineers in designing healthy workplaces. © 2013 Elsevier B.V.

Aim : To evaluate the usefulness of acoustic radiation force impulse (ARFI) in predicting hepatocellularcarcinoma (HCC) in patients with chronic hepatic disease (CHD).Methods : A total of 230 patients participated in this study. They were subsequently classifiedinto the “HCC group” and the “non-HCC group”. We measured their liver stiffness and calculatedthe mean shear wave velocity (SWV) using ARFI.Results : The mean SWV was significantly higher in the HCC group than in the non-HCCgroup. The cut-off value for the mean SWV with the best discrimination between the two groupswas 1.36 (m/s). The area under the receiver operating characteristic curve (AUROC) was0.807. The AUROCs of the aspartate-aminotransferase-to-platelet ratio and Fibrosis 4 scorewere 0.780, 0.728, respectively. The independent risk factors of HCC included the mean SWVand age. A further analysis, based on the individual causes of CHD, found that among the HCVand non-HBV and non-HCV (nBnC) cases, the mean SWV was significantly higher in the HCCgroup than in the non-HCC group.Conclusion : Measuring the SWV using ARFI was a reliable method for predicting HCC in CHDpatients, especially in HCV and nBnC patients.

We designed and jointly optimized an integrated signal processing chain for detection and classification of long-range passive-acoustic underwater signals recorded by the global geophysical monitoring network of the Comprehensive Nuclear-Test-Ban Treaty Organization. Starting at the level of raw waveform data, a processing chain of signal detection, feature extraction, and signal classification was designed and jointly optimized to the task. Relevant waveform segments were in a first step identified by a generic, flexibly parameterized detection algorithm on a long-to short-term averages' ratio of the spectral energy. For representation, general-purpose sound processing features, with an added focus on spectral and cepstral features, were extracted from the detected segments. As classifiers, support vector machines with different kernel functions were employed alongside other baseline learning algorithms. The free parameters of the overall toolchain (i.e., trigger algorithm parameters and classifier hyperparameters) were jointly optimized in a cross-validation setting, either according to the cross-validation classification error or the cross-validation area under the receiver operating characteristic curve. Experiments demonstrate that our method outperforms machine learning algorithms task-tailored to a previous, human-expert-designed preprocessing chain. The presented approach can be adapted to a wide range of problems that can benefit from jointly optimizing parameters of preprocessing and classification algorithm. © 2016 IEEE.

Background and Purpose: A large number of subjective experiences and beliefs with some degree of affinity with psychotic symptoms can be found in the general population. However, the appraisal of these psychotic-like experiences in terms of associated distress, raised preoccupation, and the conviction with which the experience is held can be more discriminative in distinguishing people in need for care from those who simply hold unusual or uncommon beliefs because of cultural reasons. Method: In this study, 81 patients with a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, diagnosis of schizophrenia or an affective disorder with psychotic features were compared on the Peters et al Delusions Inventory (PDI) to 210 people from the same local area, who had never received a formal diagnosis of a mental disorder. Results: Patients scored higher than controls on the PDI total score and on its distress, preoccupation, and conviction subscales. A stepwise logistic regression model showed PDI-preoccupation (odds ratio, 2.46; 95% confidence interval, 1.52-3.98) and, marginally, PDI-distress (odds ratio = 1.58; 95% confidence interval, 0.93-2.58) adding discriminative power to PDI total score in distinguishing patients from controls. Conclusions: The evaluation of the severity of delusion-like experiences and beliefs is important in discriminating patients diagnosed with psychosis from people who are not in need of care. Keywords: Delusions, psychosis-proneness, diagnostic accuracy, receiver operating characteristic curves, Peters et al Delusions Inventory

Background: Systemic inflammation is a whole body reaction having an infection-positive (i.e., sepsis) or infection-negative origin. It is important to distinguish between these two etiologies early and accurately because this has significant therapeutic implications for critically ill patients. We hypothesized that a molecular classifier based on peripheral blood RNAs could be discovered that would (1) determine which patients with systemic inflammation had sepsis, (2) be robust across independent patient cohorts, (3) be insensitive to disease severity, and (4) provide diagnostic utility. The goal of this study was to identify and validate such a molecular classifier., Methods and findings: We conducted an observational, non-interventional study of adult patients recruited from tertiary intensive care units (ICUs). Biomarker discovery utilized an Australian cohort (n = 105) consisting of 74 cases (sepsis patients) and 31 controls (post-surgical patients with infection-negative systemic inflammation) recruited at five tertiary care settings in Brisbane, Australia, from June 3, 2008, to December 22, 2011. A four-gene classifier combining CEACAM4, LAMP1, PLA2G7, and PLAC8 RNA biomarkers was identified. This classifier, designated SeptiCyte Lab, was validated using reverse transcription quantitative PCR and receiver operating characteristic (ROC) curve analysis in five cohorts (n = 345) from the Netherlands. Patients for validation were selected from the Molecular Diagnosis and Risk Stratification of Sepsis study (ClinicalTrials.gov, NCT01905033), which recruited ICU patients from the Academic Medical Center in Amsterdam and the University Medical Center Utrecht. Patients recruited from November 30, 2012, to August 5, 2013, were eligible for inclusion in the present study. Validation cohort 1 (n = 59) consisted entirely of unambiguous cases and controls; SeptiCyte Lab gave an area under curve (AUC) of 0.95 (95% CI 0.91-1.00) in this cohort. ROC curve analysis of an independent, more heterogeneous group of patients (validation cohorts 2-5; 249 patients after excluding 37 patients with an infection likelihood of "possible") gave an AUC of 0.89 (95% CI 0.85-0.93). Disease severity, as measured by Sequential Organ Failure Assessment (SOFA) score or Acute Physiology and Chronic Health Evaluation (APACHE) IV score, was not a significant confounding variable. The diagnostic utility of SeptiCyte Lab was evaluated by comparison to various clinical and laboratory parameters available to a clinician within 24 h of ICU admission. SeptiCyte Lab was significantly better at differentiating cases from controls than all tested parameters, both singly and in various logistic combinations, and more than halved the diagnostic error rate compared to procalcitonin in all tested cohorts and cohort combinations. Limitations of this study relate to (1) cohort compositions that do not perfectly reflect the composition of the intended use population, (2) potential biases that could be introduced as a result of the current lack of a gold standard for diagnosing sepsis, and (3) lack of a complete, unbiased comparison to C-reactive protein., Conclusions: SeptiCyte Lab is a rapid molecular assay that may be clinically useful in managing ICU patients with systemic inflammation. Further study in population-based cohorts is needed to validate this assay for clinical use., peer-reviewed

Background: Vitamin D deficiency/Vitamin D Insufficiency (VDI) is now recognized as a pandemic. Cutaneous exposure to ultraviolet-B causes photolysis of epidermal 7-dehydrocholesterol, converting it to pre-Vitamin D3 (precholecalciferol), which then undergoes isomerization to form the stable hormone Vitamin D3. Objectives: The objectives of the study is to determine the status of Vitamin D level among the adults residing in a rural area of West Bengal and to find out the optimal cutoff of the sun exposure for Vitamin D sufficiency. Methods: This study was a rural community based cross-sectional study. It was done from May 2016 to April 2017 among 197 adults residing in a rural block of West Bengal. Data were collected by interviewing the respondents with the help of a structured pre-tested pre-designed schedule. After interviewing, every individual was examined for height and weight and blood was collected for serum Vitamin 25-(OH) D. Receiver Operating Characteristic (ROC) curves were utilized to find out optimum cut-off for sun exposure using Youden's index. Binary logistic regression was performed to find out the associates of high sun exposure. All analysis was done using R. Results: Overall VDI was found in 133 (67.5%). Among them, 102 (51.8%) were female. On ROC curve for veritable sun exposure for Vitamin D sufficiency, area under curve was 0.7841 which signifies veritable sun exposure as a good screening tool. Conclusion: Proper information, education, and communication material regarding various aspects of sun exposure and Vitamin D should be prepared with due consultation of field experts and disseminated to increase awareness among the community.

Aims/Introduction The aims of the present study were to evaluate the diagnostic value of fasting plasma glucose, 2-h postload plasma glucose and glycosylated hemoglobin (HbA1c) measurements in the screening of diabetes and prediabetes, and to determine the cut-off point of HbA1c in the diagnosis of diabetes and prediabetes in a Chinese population. Materials and Methods A total of 7,611 individuals aged 40 years or older, who did not have a prior history of diabetes, were randomly selected in the Changchun area. For each participant, a questionnaire was completed, and a physical examination and an oral glucose tolerance test were carried out. The values of fasting plasma glucose, 2-h postload plasma glucose and HbA1c were compared by area under the receiver operating characteristic curves. The Youden index was used to identify the optimal cut-off point of HbA1c in the diagnosis of diabetes and prediabetes. Results The prevalence of newly diagnosed diabetes and prediabetes was 12.71% and 29.39%, respectively. In participants with newly diagnosed diabetes, the area under the receiver operating characteristic curve was 0.8368 for fasting plasma glucose, 0.9330 for 2-h postload plasma glucose and 0.8064 for HbA1c; whereas for prediabetes, these values were 0.8022, 0.9288 and 0.6895, respectively. In addition, an HbA1c threshold of 6.3% showed the highest Youden index (0.4799) for detecting diabetes; furthermore, an HbA1c threshold of 5.8% showed the highest Youden index (0.2866) for detecting prediabetes. Conclusions HbA1c ≥6.3% (45 mmol/mol) and between 5.8% and 6.2% (40–44 mmol/mol) were the optimal cut-off values for the diagnosis of diabetes and prediabetes, respectively, in a Chinese population.

The Hand Arm Risk assessment Method (HARM) is a simplified risk assessment method for determining musculoskeletal symptoms to the arm, neck and/or shoulder posed by hand-arm tasks of the upper body. The purpose of this study was to evaluate the predictive validity of HARM using data collected from a three year prospective cohort Study on Musculoskeletal disorders, Absenteeism and Health (SMASH). Structured observations of video recordings using HARM were made of 88 participants performing hand-arm tasks. These video recordings were made at baseline of the SMASH study, whereas self-reported musculoskeletal symptoms were assessed both at baseline and during a three year follow-up. The receiver operating characteristic (ROC) curves showed an area under the curve (AUC) of 0.75 (CI: 0.65-0.86) for symptoms at baseline and 0.70 (CI: 0.58-0.83) for symptoms during follow-up. For a HARM score of 50, the sensitivity and specificity were 64% and 79% at baseline and 53% and 82% during follow up. Participants with a HARM score of 50 or more (considered a high risk) had a significantly higher risk of upper limb symptoms than those with a HARM score of less than 50 at baseline (OR=5.31; 95%CI: 2.10-13.39) and 3 year follow-up (OR=5.11; (95%CI: 1.61-16.27). This study has shown that HARM can be used to assess work tasks of the hands and arms and discriminate between tasks that pose a high or low risk of neck or upper limb symptoms. Relevance to industry: The association of HARM results and self-reported neck and upper limb symptoms was studied. The association found both cross sectionally and at follow up shows that the method can be used in industry to distinguish hazardous hand arm tasks from non-hazardous ones. Thus, HARM can support OSH practitioners and industrial engineers in designing healthy workplaces. © 2013 Elsevier B.V.

Screening early renal failure: Cut-off values for serum creati- creatinine. Given the effort required to obtain accurately nine as indicator of renal impairment. timed urine specimens, the glomerular filtration rate is Background. The aim of this study was to define cut-off often estimated from serum creatinine concentration alone values for serum creatinine as an indicator of several levels of or by standard formulas estimating creatinine clearance. renal impairment. The diagnosis of renal failure is usually suspected when Methods. To identify the suitable values, receiver operating characteristic curves were constructed based on the data of serum creatinine is greater than the upper limit of the 984 laboratory assessments of renal function. The glomerular "normal" interval. These limits are defined as the 97.5th filtration rate was measured with inulin clearance. Three levels percentile established by the distribution of the value of of renal impairment were analyzed. An index that gave the serum creatinine in a large sample of population consid- same weight to false positive and false negative results was ered healthy. One study, based on 18,000 serum creati- used to determine the thresholds. Robustness of the results was tested using a "bootstrap" technique. nine measurements (Jaffereaction), provided a normal Results. Considering an inulin clearance of less than 80 ml/ value in men of 63 to 112 mmol/liter and in women the min/1.73 m 2

We propose a method to analyze bilateral asymmetry in mammograms based upon systematic comparison of paired mammographic strips of the right and left breasts of a patient. Similarity between corresponding directional structures was quantified by applying measures of similarity to the masked strips. A novel application of Moran's index was designed to measure the angular covariance between rose diagrams related to the phase and magnitude responses of multidirectional Gabor filters. A set of 128 mammograms from the DDSM database, including 32 normal and 32 asymmetric pairs, was used to validate the procedure. The leave-one-patient-out method was used for cross-validation of the results. The best result, with an area under the receiver operating characteristic curve of 0.8435, was achieved using similarity measures on craniocaudal views and Fisher-linear discriminant analysis. The results indicate that the proposed techniques can be applied for computer-aided detection of bilateral asymmetry.

We present new feature descriptors specifically designed to quantify angular nonstationarity and angular dependence of pixel values in sectors of mammographic lesions. A key novelty of this work is that the proposed measures characterize the texture of masses without relying on accurate determination of their contours. An artificial neural network based on radial basis functions was used to predict the diagnosis of 120 benign masses and 26 malignant tumors in a database of full-field digital mammograms. Features were selected using stepwise logistic regression and the leave-one-patient-out method was used for cross-validation of results. An area under the receiver operating characteristic curve of 0.9890 ± 0.0114 was obtained using randomly selected centroids and an expected size of the masses. Results indicate that the use of the proposed contour-independent features can be an effective approach for computer-aided classification of mammographic lesions.

In this paper, a novel approach for classification of breast masses is presented that quantifies the texture of masses without relying on accurate extraction of their contours. Two novel feature descriptors based on 2D extensions of the reverse arrangement (RA) and Mantel's tests were designed for this purpose. Measures of radial correlation and radial trend were extracted from the original gray-scale values as well as from the Gabor magnitude response of 146 regions of interest, including 120 benign masses and 26 malignant tumors. Four classifiers, Fisher-linear discriminant analysis, Bayesian, support vector machine, and an artificial neural network based on radial basis functions (ANN-RBF), were employed to predict the diagnosis, using stepwise logistic regression for feature selection and the leave-one-patient-out method for cross-validation. The ANN-RBF resulted in an area under the receiver operating characteristic curve of 0.93. The experimental results show the effectiveness of the proposed approach.

Introduction Acute appendicitis (AA) is common surgical problem associated with acute-phase reaction. Blood tests role in decision-making process is unclear. This retrospective study aimed to determine diagnostic value of preoperative evaluation of white blood cells (WBCs) and neutrophils and its value in predicting AA severity. Methods Medical records of 456 patients who underwent appendectomy during 4-years period were retrospectively reviewed. Patients were subdivided according to histological finding into: normal appendix (n = 29), uncomplicated inflamed appendix (n = 350), complicated appendicitis (n = 77). Diagnostic performances of WBCs and neutrophils were analyzed using receiver operating characteristic (ROC) curves. Results WBCs and neutrophils counts were higher in patients with inflamed and complicated appendix than normal appendix and in complicated than inflamed appendix. In patients, WBCs count 9.400 × 103/mL had sensitivity of 76.81%, specificity of 65.52%, positive predictive value (PPV) of 97.0%, negative predictive value (NPV) of 16.1%, positive likelihood ratio [LR(+)] of 2.23, negative LR(−) of 0.35. Neutrophil count 7.540 × 103/mL had sensitivity of 70.96%, specificity of 65.52%, PPV of 96.8%, NPV of 13.3%, LR(+) of 2.06, LR(−) of 0.44. Areas under ROC curve were 0.701, 0.680 for elevated WBCs and neutrophils count. Conclusions Clinicians should not rely on either elevated WBCs or neutrophils count as appendicitis indicator as clinical data are superior in decision-making appendectomy.

This paper deals with the problem of voice activity detection in adverse acoustic conditions, namely high and varying noise scenarios. For robotic applications, we need the voice activity detector to be computationally light, robust to varying levels of background noise, and have a low latency, especially if we are tracking moving speakers. We analyze three voice activity detectors—two model the discrete Fourier transform coefficients by Gaussian and generalized Gaussian distribution, while the third models the spectral envelope as having either Rayleigh or Rice distribution—and we present them in a unifying and consistent manner, with respect to a statistical hypotheses ratio measure and a joint noise spectrum estimation algorithm. Moreover, we compare the performance under various noise conditions; three types of noises, three different signal-to-noise ratios and six different speakers, by means of receiver operating characteristic curves and area under a curve score. The results showed that the Rayleigh-Rice model had on average better results and medium computational demand.

We propose a generalization of the matched subspace filter for the detection of unknown signals in a background of non-Gaussian and nonindependent noise. The generalization is based on a modification of the Rao test by including a linear transformation derived from independent component analysis (ICA). Receiver operating characteristic (ROC) curves computed for simulated examples illustrate the significant improvement achieved with the generalized solution. © 2011 IEEE., This work has been supported by the Spanish Administration and the FEDER Programme of the EC under Grant TEC 2008-02975; and by the Generalitat Valenciana under Grant PROMETEO/2010/040.

The localisation and calibration of damage in a structure are often difficult, time consuming, subjective and error prone. The importance of a simple, fast and relatively inexpensive non-destructive technique (NDT) with reliable measurements is thus greatly felt. The usefulness and the efficiency of any such technique are often affected by environmental conditions. The definition of damage and the subsequent interpretation of the possible consequences due to the damage introduce subjectivity into an NDT technique and affect its performance. It is of great importance in terms of practical application to find out the efficiency of an NDT technique in a probabilistic way for various damage definitions and environmental conditions through the use of receiver operating characteristic (ROC) curves. Such variations of performance of an NDT tool can be predicted through simulation processes, and the test conditions conducive to good detections can be isolated and ranked according to their relative efficiency. This paper considers a camera based image analysis technique to identify, quantify and classify damage in structures at various levels of scale. The general method has been applied to identify the affected areas on aluminium due to pitting corrosion. The method depends on the optical contrast of the corroded region with respect to its surroundings, performs intelligent edge detection through image processing techniques and computes each affected and closed region to predict the total area of the affected part, together with its spatial distribution on a two-dimensional plane. The effects of various environmental factors on the quality of such images are simulated from an original photograph. The objectivity and the amount of available information, quantification and localisation and the extent of pitting corrosion are observed, together with the various constructed ROC curves. The method provides the engineer, the owner of the structure and the end-user of the NDT technique with a tool to assess the performance of the structure in an as-built condition and decide on the appropriateness of a certain NDT, under a given environmental condition and a certain definition of damage. Moreover, it allows the findings of the NDT results to be introduced in the decision chain and risk analysis.

Oral submucous fibrosis (OSF) is a precancerous condition which leads to the development of fibrous band and loss of oral mucosa elasticity. In the absence of medical treatment OSF can also lead to oral cancer. In the present work, in vivo fluorescence data obtained from the oral tissues of the pre-treated OSF subjects, post-treated OSF subjects, and the normal healthy volunteers were subjected to principal component analysis (PCA) and partial least square discriminate analysis (PLS-DA). PCA and PLS-DA are unsupervised and supervised pattern recognition methods, respectively. PCA and PLS-DA models were found to be highly sensitive and specific. The receiver operating characteristic (ROC) curve of the PLS-DA models mainly consisted of horizontal and vertical lines with an area under the curve of greater than 0.9 which shows that PLS-DA models are genuine classifier. Outcomes of the present work clearly illustrate that by combining the in vivo fluorescence spectroscopy with PCA and PLS-DA, a fast, sensitive, and non-invasive procedure could be obtained for the diagnosis of OSF affected subjects. The developed diagnosis procedure could be of significant importance in rural areas where access to clinical diagnosis is relatively difficult and costly. � 2013 The Royal Society of Chemistry.

Background Preventing diabetic foot problems (DFP) and their associated consequences is a critical in rural regions. The objective is to present an association of non-invasive DFP assessment tools and physiological indicators for early detection among rural cases of diabetes in Taiwan. Methods Secondary data analysis of 387 participants previously diagnosed with type 2 diabetes was used. The Michigan Neuropathy Screening Instrument (MNSI), Ankle Brachial Index (ABI), optimal scaling combination (OSC) of MNSI, and age were used to examine peripheral neurovascular function. The King’s College classification (KC) and Texas risk classification (TRC) were used to understand diabetic foot complications. Results The findings indicated that MNSI was negatively correlated with ABI, but positively with diabetes duration, age, KC, TRC, fasting blood glucose, low density of lipoprotein cholesterol, body mass index and waist circumference. The area under the receiver operating characteristic curves for assessing the risk of ABI based on OSC was larger than for MNSI, KC, and TRC. Conclusion It is shown that using OSC, MNSI, and ABI as community screening tools is useful in detecting early neurovasculopathy. In addition, where an ABI machine is unavailable, primary healthcare providers that perform MNSI or OSC may be cost-effective. The study was approved by the institutional review board of the ethical committee (No 98-2224-B).