Search

To evaluate the clinical utility of new risk markers, a crucial step is to measure their predictive accuracy with prospective studies. However, it is often infeasible to obtain marker values for all study participants. The nested case-control (NCC) design is a useful cost-effective strategy for such settings. Under the NCC design, markers are only ascertained for cases and a fraction of controls sampled randomly from the risk sets. The outcome dependent sampling generates a complex data structure and therefore a challenge for analysis. Existing methods for analyzing NCC studies focus primarily on association measures. Here, we propose a class of nonparametric estimators for commonly used accuracy measures. We derived asymptotic expansions for accuracy estimators based on both finite population and Bernoulli sampling and established asymptotic equivalence between the two. Simulation results suggest that the proposed procedures perform well in finite samples. The new procedures were illustrated with data from the Framingham Offspring study. [ABSTRACT FROM AUTHOR]

In a prospective cohort study, information on clinical parameters, tests, and molecular markers often is collected. Such information is useful to predict patient prognosis and to select patients for targeted therapy. We propose a new graphical approach, the positive predictive value (PPV) curve, to quantify the predictive accuracy of prognostic markers measured on a continuous scale with censored failure time outcome. The proposed method highlights the need to consider both predictive values and the marker distribution in the population when evaluating a marker, and it provides a common scale for comparing different markers. We consider both semiparametric- and nonparametric- based estimating procedures. In addition, we provide asymptotic distribution theory and resampling-based procedures for making statistical inference. We illustrate our approach with numerical studies and data sets from the Seattle Heart Failure Study. [ABSTRACT FROM AUTHOR]

Alteration of gene methylation patterns has been reported to be involved in the early Onsets of many human malignancies. Many exogenous risk factors, such as cigarette smoke, dietary additives, chemical exposures, radiation, and biologic agents including viral infection, are involved in the methylation pathways of cancers. We propose a multidimensional selective item response regression model to describe and test how a risk factor may alter molecular pathways involving aberrant methylation of multiple genes in oncogenesis. Our modeling framework is built on an item response model for multivariate dichotomous responses of high dimension, such as aberrant methylation of multiple tumor-suppressor genes, but we allow risk factors such as SV40 viral infection to alter the distribution of the latent factors that subsequently affect the outcome of cancer. We postulate empirical identification conditions under our model formulation. Moreover, we do not prespecify the links between the multiple dichotomous methylation responses and the latent factors, but rather conduct specification searches with a genetic algorithm to discover the links. Parameter estimation through maximum likelihood and specification searches in models with multidimensional latent factors for multivariate binary responses have become practical only recently, due to modern statistical computing development. We illustrate our proposal with the biological finding that simultaneous methylation of multiple tumor-suppressor genes is associated with the presence of SV40 viral sequences and with the cancer status of lymphoma/leukemia. We are able to test whether the data are consistent with the causal hypothesis that SV40 induces aberrant methylation of multiple genes in its oncogenic pathways. At the same time, we are able to evaluate the role of SV40 in the methylation pathway and to determine whether the methylation pathway is responsible for the development of leukemia/lymphoma. [ABSTRACT FROM AUTHOR]

This article discusses the performance of Bayesian model averaging (BMA) relative to other methods for real datasets. The performance of Bayesian model selection and BMA has been studied. There are three main strands of results: (1) General theoretical results from Jeffreys in 1939, (2) Simulation studies and (3) Results on out-of-sample predictive performance. There are several realistic simulation studies of the performance of BMA relative to other methods in a variety of situations, including linear regression (George and McCulloch 1993 and Raftery, Madigan and Hoeting 1997), log-linear models (Clyde 1999), logistic regression (Viallefont, Raftery and Richardson 2001) and wavelets (Clyde and George 2000). The BMA distribution of a quantity of interest can be viewed as the posterior distribution from the full model, with a mixed discrete-continuous prior distribution that assigns weight to the events that the individual components are zero. A simple normal example has been presented to assess numerical differences in performance and also of the extent to which results hold event if the prior distribution is not the same as the practical distribution. Posterior model probabilities can be computed in four ways.