Candidate Predictor (candidate + predictor)

Distribution by Scientific Domains
Medical Sciences40%
Mathematics and Statistics40%

Selected Abstracts

The presence of prostate cancer on saturation biopsy can be accurately predicted

BJU INTERNATIONAL, Issue 5 2010
Sascha A. Ahyai
Study Type , Diagnostic (non-consecutive) Level of Evidence 3b OBJECTIVE To improve the ability of our previously reported saturation biopsy nomogram quantifying the risk of prostate cancer, as the use of office-based saturation biopsy has increased. PATIENTS AND METHODS Saturation biopsies of 540 men with one or more previously negative 6,12 core biopsies were used to develop a multivariable logistic regression model-based nomogram, predicting the probability of prostate cancer. Candidate predictors were used in their original or stratified format, and consisted of age, total prostate-specific antigen (PSA) level, percentage free PSA (%fPSA), gland volume, findings on a digital rectal examination, cumulative number of previous biopsy sessions, presence of high-grade prostatic intraepithelial neoplasia on any previous biopsy, and presence of atypical small acinar proliferation (ASAP) on any previous biopsy. Two hundred bootstraps re-samples were used to adjust for overfit bias. RESULTS Prostate cancer was diagnosed in 39.4% of saturation biopsies. Age, total PSA, %fPSA, gland volume, number of previous biopsies, and presence of ASAP at any previous biopsy were independent predictors for prostate cancer (all P < 0.05). The nomogram was 77.2% accurate and had a virtually perfect correlation between predicted and observed rates of prostate cancer. CONCLUSIONS We improved the accuracy of the saturation biopsy nomogram from 72% to 77%; it relies on three previously included variables, i.e. age, %fPSA and prostate volume, and on three previously excluded variables, i.e. PSA, the number of previous biopsy sessions, and evidence of ASAP on previous biopsy. Our study represents the largest series of saturation biopsies to date. [source]

Upregulation of glycolytic enzymes in proteins secreted from human colon cancer cells with 5-fluorouracil resistance

ELECTROPHORESIS, Issue 12 2009
Young-Kyoung Shin
Abstract 5-Fluorouracil (5-FU) is the most commonly used chemotherapeutic agent for colorectal cancer (CRC). However, resistance to this drug is a major obstacle in CRC chemotherapy. Accurate prediction of response to 5-FU would avoid unnecessary chemotherapy and allow the selection of other effective drugs. To identify a candidate predictor of 5-FU resistance, we isolated secreted proteins that were up- or downregulated in a 5-FU-resistant cancer cell line, compared with the parent cell line (SNU-C4), using a stable isotope-coded labeling protocol. For validating the clinical applicability of this method, levels of the identified proteins were determined in the sera of 46 patients treated with 5-FU. In total, 238 proteins with molecular weights ranging from 50 to 75,kDa were identified. Among these, 45 and 35 secreted proteins were up- and downregulated in the 5-FU-resistant cell line, respectively. We observed significant upregulation of glycolytic enzymes, including glyceraldehyde-3-phosphate dehydrogenase, pyruvate kinase M2 (PK-M2), transketolase, and NADP(+)-dependent malic enzyme 1. In particular, the level of PK-M2, a key enzyme in the glycolytic pathway, showed an increasing tendency in both sera and tissues from CRC patients displaying no response to 5-FU-based chemotherapy (progressive and stable disease cases), compared with that in complete or partial responders to 5-FU-based chemotherapy; however, it did not reach the statistical significance. In conclusion, increasing pattern of PK-M2 observed with 5-FU resistance induced in vitro and in sera and tissues from CRC patients displaying poor response to 5-FU-based chemotherapy suggest the relevance of dysregulated glycolysis and 5-FU-resistant CRC. [source]

Nonlinear Indices of Heart Rate Variability in Chronic Heart Failure Patients: Redundancy and Comparative Clinical Value

JOURNAL OF CARDIOVASCULAR ELECTROPHYSIOLOGY, Issue 4 2007
ROBERTO MAESTRI M.S.
Aims: We aimed to assess the mutual interrelationships and to compare the prognostic value of a comprehensive set of nonlinear indices of heart rate variability (HRV) in a population of chronic heart failure (CHF) patients. Methods and Results: Twenty nonlinear HRV indices, representative of symbolic dynamics, entropy, fractality-multifractality, predictability, empirical mode decomposition, and Poincaré plot families, were computed from 24-hour Holter recordings in 200 stable CHF patients in sinus rhythm (median age [interquartile range]: 54 [47,58] years, LVEF: 23 [19,28]%, NYHA class II,III: 88%). End point for survival analysis (Cox model) was cardiac death or urgent transplantation. Homogeneous variables were grouped by cluster analysis, and in each cluster redundant variables were discarded. A prognostic model including only known clinical and functional risk factors was built and the ability of each selected HRV variable to add prognostic information to this model assessed. Bootstrap resampling was used to test the models stability. Four nonlinear variables showed a correlation >0.90 with classical linear ones and were discarded. Correlations >0.80 were found between several nonlinear variables. Twelve clusters were obtained and from each cluster a candidate predictor was selected. Only two variables (from empirical mode decomposition and symbolic dynamics families) added prognostic information to the clinical model. Conclusion: This exploratory study provides evidence that, despite some redundancies in the informative content of nonlinear indices and strong differences in their prognostic power, quantification of nonlinear properties of HRV provides independent information in risk stratification of CHF patients. [source]

Neighborhood search heuristics for selecting hierarchically well-formulated subsets in polynomial regression

NAVAL RESEARCH LOGISTICS: AN INTERNATIONAL JOURNAL, Issue 1 2010
Michael J. Brusco
Abstract The importance of subset selection in multiple regression has been recognized for more than 40 years and, not surprisingly, a variety of exact and heuristic procedures have been proposed for choosing subsets of variables. In the case of polynomial regression, the subset selection problem is complicated by two issues: (1) the substantial growth in the number of candidate predictors, and (2) the desire to obtain hierarchically well-formulated subsets that facilitate proper interpretation of the regression parameter estimates. The first of these issues creates the need for heuristic methods that can provide solutions in reasonable computation time; whereas the second requires innovative neighborhood search approaches that accommodate the hierarchical constraints. We developed tabu search and variable neighborhood search heuristics for subset selection in polynomial regression. These heuristics are applied to a classic data set from the literature and, subsequently, evaluated in a simulation study using synthetic data sets. © 2009 Wiley Periodicals, Inc. Naval Research Logistics, 2010 [source]

A Partially Linear Tree-based Regression Model for Multivariate Outcomes

BIOMETRICS, Issue 1 2010
Kai Yu
Summary In the genetic study of complex traits, especially behavior related ones, such as smoking and alcoholism, usually several phenotypic measurements are obtained for the description of the complex trait, but no single measurement can quantify fully the complicated characteristics of the symptom because of our lack of understanding of the underlying etiology. If those phenotypes share a common genetic mechanism, rather than studying each individual phenotype separately, it is more advantageous to analyze them jointly as a multivariate trait to enhance the power to identify associated genes. We propose a multilocus association test for the study of multivariate traits. The test is derived from a partially linear tree-based regression model for multiple outcomes. This novel tree-based model provides a formal statistical testing framework for the evaluation of the association between a multivariate outcome and a set of candidate predictors, such as markers within a gene or pathway, while accommodating adjustment for other covariates. Through simulation studies we show that the proposed method has an acceptable type I error rate and improved power over the univariate outcome analysis, which studies each component of the complex trait separately with multiple-comparison adjustment. A candidate gene association study of multiple smoking-related phenotypes is used to demonstrate the application and advantages of this new method. The proposed method is general enough to be used for the assessment of the joint effect of a set of multiple risk factors on a multivariate outcome in other biomedical research settings. [source]