Predictive models for diagnosis of pleural effusions secondary to tuberculosis or cancer

Background and objective: Tuberculosis (TB) and cancer are two of the main causes of pleural effusions which frequently share similar clinical features and pleural fluid profiles. This study aimed to identify diagnostic models based on clinical and laboratory variables to differentiate tuberculous from malignant pleural effusions. Methods: A retrospective study of 403 patients (200 with TB; 203 with cancer) was undertaken. Univariate analysis was used to select the clinical variables relevant to the models composition. Variables beta coefficients were used to define a numerical score which presented a practical use. The performances of the most efficient models were tested in a sample of pleural exudates (64 new cases). Results: Two models are proposed for the diagnosis of effusions associated with each disease. For TB: (i) adenosine deaminase (ADA), globulins and the absence of malignant cells in the pleural fluid; and (ii) ADA, globulins and fluid appearance. For cancer: (i) patient age, fluid appearance, macrophage percentage and presence of atypical cells in the pleural fluid; and (ii) as for (i) excluding atypical cells. Application of the models to the 64 pleural effusions showed accuracy higher than 85% for all models. Conclusions: The proposed models were effective in suggesting pleural tuberculosis or cancer.

Prognostic modelling of breast cancer patients: a benchmark of predictive models with external validation

Dissertação apresentada para obtenção do Grau de Doutor em Engenharia Electrotécnica e de Computadores – Sistemas Digitais e Percepcionais pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia

Development and Validation of Predictive Models of Cardiac Mortality and Transplantation in Resynchronization Therapy

AbstractBackground:30-40% of cardiac resynchronization therapy cases do not achieve favorable outcomes.Objective:This study aimed to develop predictive models for the combined endpoint of cardiac death and transplantation (Tx) at different stages of cardiac resynchronization therapy (CRT).Methods:Prospective observational study of 116 patients aged 64.8 ± 11.1 years, 68.1% of whom had functional class (FC) III and 31.9% had ambulatory class IV. Clinical, electrocardiographic and echocardiographic variables were assessed by using Cox regression and Kaplan-Meier curves.Results:The cardiac mortality/Tx rate was 16.3% during the follow-up period of 34.0 ± 17.9 months. Prior to implantation, right ventricular dysfunction (RVD), ejection fraction < 25% and use of high doses of diuretics (HDD) increased the risk of cardiac death and Tx by 3.9-, 4.8-, and 5.9-fold, respectively. In the first year after CRT, RVD, HDD and hospitalization due to congestive heart failure increased the risk of death at hazard ratios of 3.5, 5.3, and 12.5, respectively. In the second year after CRT, RVD and FC III/IV were significant risk factors of mortality in the multivariate Cox model. The accuracy rates of the models were 84.6% at preimplantation, 93% in the first year after CRT, and 90.5% in the second year after CRT. The models were validated by bootstrapping.Conclusion:We developed predictive models of cardiac death and Tx at different stages of CRT based on the analysis of simple and easily obtainable clinical and echocardiographic variables. The models showed good accuracy and adjustment, were validated internally, and are useful in the selection, monitoring and counseling of patients indicated for CRT.

Predictive models of syncope causes in an outpatient clinic

The investigation of unexplained syncope remains a challenging clinical problem. In the present study we sought to evaluate the diagnostic value of a standardized work-up focusing on non invasive tests in patients with unexplained syncope referred to a syncope clinic, and whether certain combinations of clinical parameters are characteristic of rhythmic and reflex causes of syncope. METHODS AND RESULTS: 317 consecutive patients underwent a standardized work-up including a 12-lead ECG, physical examination, detailed history with screening for syncope-related symptoms using a structured questionnaire followed by carotid sinus massage (CSM), and head-up tilt test. Invasive testings including an electrophysiological study and implantation of a loop recorder were only performed in those with structural heart disease or traumatic syncope. Our work-up identified an etiology in 81% of the patients. Importantly, three quarters of the causes were established non invasively combining head-up tilt test, CSM and hyperventilation testing. Invasive tests yielded an additional 7% of diagnoses. Logistic analysis identified age and number of significant prodromes as the only predictive factors of rhythmic syncope. The same two factors, in addition to the duration of the ECG P-wave, were also predictive of vasovagal and psychogenic syncope. These factors, optimally combined in predictive models, showed a high negative and a modest positive predictive value. CONCLUSION: A standardized work-up focusing on non invasive tests allows to establish more than three quarters of syncope causes. Predictive models based on simple clinical parameters may help to distinguish between rhythmic and other causes of syncope

The MicroArray Quality Control (MAQC)-II study of common practices for the development and validation of microarray-based predictive models.

Gene expression data from microarrays are being applied to predict preclinical and clinical endpoints, but the reliability of these predictions has not been established. In the MAQC-II project, 36 independent teams analyzed six microarray data sets to generate predictive models for classifying a sample with respect to one of 13 endpoints indicative of lung or liver toxicity in rodents, or of breast cancer, multiple myeloma or neuroblastoma in humans. In total, >30,000 models were built using many combinations of analytical methods. The teams generated predictive models without knowing the biological meaning of some of the endpoints and, to mimic clinical reality, tested the models on data that had not been used for training. We found that model performance depended largely on the endpoint and team proficiency and that different approaches generated models of similar performance. The conclusions and recommendations from MAQC-II should be useful for regulatory agencies, study committees and independent investigators that evaluate methods for global gene expression analysis.

Building predictive models of soil particle-size distribution

Is it possible to build predictive models (PMs) of soil particle-size distribution (psd) in a region with complex geology and a young and unstable land-surface? The main objective of this study was to answer this question. A set of 339 soil samples from a small slope catchment in Southern Brazil was used to build PMs of psd in the surface soil layer. Multiple linear regression models were constructed using terrain attributes (elevation, slope, catchment area, convergence index, and topographic wetness index). The PMs explained more than half of the data variance. This performance is similar to (or even better than) that of the conventional soil mapping approach. For some size fractions, the PM performance can reach 70 %. Largest uncertainties were observed in geologically more complex areas. Therefore, significant improvements in the predictions can only be achieved if accurate geological data is made available. Meanwhile, PMs built on terrain attributes are efficient in predicting the particle-size distribution (psd) of soils in regions of complex geology.

Introduction of snow and geomorphic disturbance variables into predictive models of alpine plant distribution in the Western Swiss Alps

Indirect topographic variables have been used successfully as surrogates for disturbance processes in plant species distribution models (SDM) in mountain environments. However, no SDM studies have directly tested the performance of disturbance variables. In this study, we developed two disturbance variables: a geomorphic index (GEO) and an index of snow redistribution by wind (SNOW). These were developed in order to assess how they improved both the fit and predictive power of presenceabsence SDM based on commonly used topoclimatic (TC) variables for 91 plants in the Western Swiss Alps. The individual contribution of the disturbance variables was compared to TC variables. Maps of models were prepared to spatially test the effect of disturbance variables. On average, disturbance variables significantly improved the fit but not the predictive power of the TC models and their individual contribution was weak (5.6% for GEO and 3.3% for SNOW). However their maximum individual contribution was important (24.7% and 20.7%). Finally, maps including disturbance variables (i) were significantly divergent from TC models in terms of predicted suitable surfaces and connectivity between potential habitats, and (ii) were interpreted as more ecologically relevant. Disturbance variables did not improve the transferability of models at the local scale in a complex mountain system, and the performance and contribution of these variables were highly species-specific. However, improved spatial projections and change in connectivity are important issues when preparing projections under climate change because the future range size of the species will determine the sensitivity to changing conditions.

Minimal requirements for parameters to be used for the development of predictive models for microbial source tracking. The example of the pair somatic coliphages and phages infecting Bacteroides

Several methods and approaches for measuring parameters to determine fecal sources of pollution in water have been developed in recent years. No single microbial or chemical parameter has proved sufficient to determine the source of fecal pollution. Combinations of parameters involving at least one discriminating indicator and one universal fecal indicator offer the most promising solutions for qualitative and quantitative analyses. The universal (nondiscriminating) fecal indicator provides quantitative information regarding the fecal load. The discriminating indicator contributes to the identification of a specific source. The relative values of the parameters derived from both kinds of indicators could provide information regarding the contribution to the total fecal load from each origin. It is also essential that both parameters characteristically persist in the environment for similar periods. Numerical analysis, such as inductive learning methods, could be used to select the most suitable and the lowest number of parameters to develop predictive models. These combinations of parameters provide information on factors affecting the models, such as dilution, specific types of animal source, persistence of microbial tracers, and complex mixtures from different sources. The combined use of the enumeration of somatic coliphages and the enumeration of Bacteroides-phages using different host specific strains (one from humans and another from pigs), both selected using the suggested approach, provides a feasible model for quantitative and qualitative analyses of fecal source identification.

Predictive models for nanotoxicology: current challenges and future opportunities.

Characterizing the risks posed by nanomaterials is extraordinarily complex because these materials can have a wide range of sizes, shapes, chemical compositions and surface modifications, all of which may affect toxicity. There is an urgent need for a testing strategy that can rapidly and efficiently provide a screening approach for evaluating the potential hazard of nanomaterials and inform the prioritization of additional toxicological testing where necessary. Predictive toxicity models could form an integral component of such an approach by predicting which nanomaterials, as a result of their physico-chemical characteristics, have potentially hazardous properties. Strategies for directing research towards predictive models and the ancillary benefits of such research are presented here.

Minimal requirements for parameters to be used for the development of predictive models for microbial source tracking. The example of the pair somatic coliphages and phages infecting Bacteroides

Several methods and approaches for measuring parameters to determine fecal sources of pollution in water have been developed in recent years. No single microbial or chemical parameter has proved sufficient to determine the source of fecal pollution. Combinations of parameters involving at least one discriminating indicator and one universal fecal indicator offer the most promising solutions for qualitative and quantitative analyses. The universal (nondiscriminating) fecal indicator provides quantitative information regarding the fecal load. The discriminating indicator contributes to the identification of a specific source. The relative values of the parameters derived from both kinds of indicators could provide information regarding the contribution to the total fecal load from each origin. It is also essential that both parameters characteristically persist in the environment for similar periods. Numerical analysis, such as inductive learning methods, could be used to select the most suitable and the lowest number of parameters to develop predictive models. These combinations of parameters provide information on factors affecting the models, such as dilution, specific types of animal source, persistence of microbial tracers, and complex mixtures from different sources. The combined use of the enumeration of somatic coliphages and the enumeration of Bacteroides-phages using different host specific strains (one from humans and another from pigs), both selected using the suggested approach, provides a feasible model for quantitative and qualitative analyses of fecal source identification.

Predictive models for chronic renal disease using decision trees, naïve bayes and case-based methods

Data mining can be used in healthcare industry to “mine” clinical data to discover hidden information for intelligent and affective decision making. Discovery of hidden patterns and relationships often goes intact, yet advanced data mining techniques can be helpful as remedy to this scenario. This thesis mainly deals with Intelligent Prediction of Chronic Renal Disease (IPCRD). Data covers blood, urine test, and external symptoms applied to predict chronic renal disease. Data from the database is initially transformed to Weka (3.6) and Chi-Square method is used for features section. After normalizing data, three classifiers were applied and efficiency of output is evaluated. Mainly, three classifiers are analyzed: Decision Tree, Naïve Bayes, K-Nearest Neighbour algorithm. Results show that each technique has its unique strength in realizing the objectives of the defined mining goals. Efficiency of Decision Tree and KNN was almost same but Naïve Bayes proved a comparative edge over others. Further sensitivity and specificity tests are used as statistical measures to examine the performance of a binary classification. Sensitivity (also called recall rate in some fields) measures the proportion of actual positives which are correctly identified while Specificity measures the proportion of negatives which are correctly identified. CRISP-DM methodology is applied to build the mining models. It consists of six major phases: business understanding, data understanding, data preparation, modeling, evaluation, and deployment.

Predictive models for mutations in mismatch repair genes: implication for genetic counseling in developing countries

Background: Lynch syndrome (LS) is the most common form of inherited predisposition to colorectal cancer (CRC), accounting for 2-5% of all CRC. LS is an autosomal dominant disease characterized by mutations in the mismatch repair genes mutL homolog 1 (MLH1), mutS homolog 2 (MSH2), postmeiotic segregation increased 1 (PMS1), post-meiotic segregation increased 2 (PMS2) and mutS homolog 6 (MSH6). Mutation risk prediction models can be incorporated into clinical practice, facilitating the decision-making process and identifying individuals for molecular investigation. This is extremely important in countries with limited economic resources. This study aims to evaluate sensitivity and specificity of five predictive models for germline mutations in repair genes in a sample of individuals with suspected Lynch syndrome. Methods: Blood samples from 88 patients were analyzed through sequencing MLH1, MSH2 and MSH6 genes. The probability of detecting a mutation was calculated using the PREMM, Barnetson, MMRpro, Wijnen and Myriad models. To evaluate the sensitivity and specificity of the models, receiver operating characteristic curves were constructed. Results: Of the 88 patients included in this analysis, 31 mutations were identified: 16 were found in the MSH2 gene, 15 in the MLH1 gene and no pathogenic mutations were identified in the MSH6 gene. It was observed that the AUC for the PREMM (0.846), Barnetson (0.850), MMRpro (0.821) and Wijnen (0.807) models did not present significant statistical difference. The Myriad model presented lower AUC (0.704) than the four other models evaluated. Considering thresholds of >= 5%, the models sensitivity varied between 1 (Myriad) and 0.87 (Wijnen) and specificity ranged from 0 (Myriad) to 0.38 (Barnetson). Conclusions: The Barnetson, PREMM, MMRpro and Wijnen models present similar AUC. The AUC of the Myriad model is statistically inferior to the four other models.

Secondary production of sandy beach macrofauna: An evaluation of predictive models

During the last three decades, several predictive models have been developed to estimate the somatic production of macroinvertebrates. Although the models have been evaluated for their ability to assess the production of macrobenthos in different marine ecosystems, these approaches have not been applied specifically to sandy beach macrofauna and may not be directly applicable to this transitional environment. Hence, in this study, a broad literature review of sandy beach macrofauna production was conducted and estimates obtained with cohort-based and size-based methods were collected. The performance of nine models in estimating the production of individual populations from the sandy beach environment, evaluated for all taxonomic groups combined and for individual groups separately, was assessed, comparing the production predicted by the models to the estimates obtained from the literature (observed production). Most of the models overestimated population production compared to observed production estimates, whether for all populations combined or more specific taxonomic groups. However, estimates by two models developed by Cusson and Bourget provided best fits to measured production, and thus represent the best alternatives to the cohort-based and size-based methods in this habitat. The consistent performance of one of these Cusson and Bourget models, which was developed for the macrobenthos of sandy substrate habitats (C&B-SS), shows that the performance of a model does not depend on whether it was developed for a specific taxonomic group. Moreover, since some widely used models (e.g., the Robertson model) show very different responses when applied to the macrofauna of different marine environments (e.g., sandy beaches and estuaries), prior evaluation of these models is essential.