Biomedical subjects

Brent R Logan

Publications and source records attributed to Brent R Logan.

8 recordsLinked to original sources

PubMed2006

Regression models for hazard rates versus cumulative incidence probabilities in hematopoietic cell transplantation data.

In this article, we consider methods of regression modeling in the competing risks setting commonly encountered in analyzing stem cell transplantation data. We clarify the distinction between modeling the cause-specific hazard rate and modeling the cumulative incidence probability or function, and we review regression techniques for both types of quantities. We apply them to 2 examples: 1 comparing engraftment and 1 examining relapse. These examples illustrate that different conclusions may result depending on the type of regression model used for comparing treatments. Finally, we show how these discrepancies occur because 2 different characteristics of the time-to-event distribution are being modeled.

Graft Rejection↗

PubMed2005

Pairwise multiple comparison adjustment in survival analysis.

Many clinical studies have as their endpoint the time until some event (such as death) occurs. Often in such studies researchers are interested in comparing several treatment or prognostic groups with one another in terms of their survival curves. When many such pairwise group comparisons are done, the chance of finding a false significance among all of the comparisons is inflated above the usual desired significance level. This paper investigates methods of adjusting the survival analysis for the number of comparisons being made. These methods are applied to a retrospective study conducted by the International Bone Marrow Transplant Registry and compared in a simulation study in terms of the power to detect actual differences in the survival curves between the groups.

Bone Marrow Transplantation↗

PubMed2005

Optimal two-stage randomized phase II clinical trials.

Randomized phase II clinical trials can be an efficient means of evaluating several potential new treatments prior to a comparative phase III clinical trial. However, selection designs do not allow for an assessment of minimal efficacy, nor do they allow early closing of a study arm for lack of efficacy, and so are only appropriate when all arms have met minimal efficacy requirements. Classic two-stage hypothesis testing based designs such as a Simon two-stage optimal design can satisfy these requirements, when used in parallel for the multiple treatment arms. However, these can be inefficient and impractical in terms of the maximum planned sample size required for such a study. This article proposes a two-stage randomized phase II clinical trial, where the sample size is adaptive at the second stage to the number of treatment arms passing through the first stage. The study design relies on optimality criteria, which are analogous to those used by Simon for single-arm studies. The proposed design maintains similar operating characteristics to a strategy of using multiple single-arm studies, in most situations of primary interest. Furthermore, it generally has smaller actual sample size when most of the treatments are effective than using multiple single-arm studies, leading to substantial efficiencies in terms of maximum planned sample size.

Clinical Trials, Phase II as Topic↗

PubMed2004

Complex fMRI analysis with unrestricted phase is equivalent to a magnitude-only model.

Due to phase imperfections, voxel time course measurements are complex valued. However, most fMRI studies measure activation using magnitude-only time courses. We show that magnitude-only analyses are equivalent to a complex fMRI activation model in which the phase is unrestricted, or allowed to dynamically change over time. This suggests that improvements to the magnitude-only model are possible by modeling the phase in each voxel over time.

Animals↗

PubMed2004

An evaluation of thresholding techniques in fMRI analysis.

This paper reviews and compares individual voxel-wise thresholding methods for identifying active voxels in single-subject fMRI datasets. Different error rates are described which may be used to calibrate activation thresholds. We discuss methods which control each of the error rates at a prespecified level alpha, including simple procedures which ignore spatial correlation among the test statistics as well as more elaborate ones which incorporate this correlation information. The operating characteristics of the methods are shown through a simulation study, indicating that the error rate used has an important impact on the sensitivity of the thresholding method, but that accounting for correlation has little impact. Therefore, the simple procedures described work well for thresholding most single-subject fMRI experiments and are recommended. The methods are illustrated with a real bilateral finger tapping experiment.

Algorithms↗

PubMed2004

A complex way to compute fMRI activation.

In functional magnetic resonance imaging, voxel time courses after Fourier or non-Fourier "image reconstruction" are complex valued as a result of phase imperfections due to magnetic field inhomogeneities. Nearly all fMRI studies derive functional "activation" based on magnitude voxel time courses [Bandettini, P., Jesmanowicz, A., Wong, E., Hyde, J.S., 1993. Processing strategies for time-course data sets in functional MRI of the human brain. Magn. Reson. Med. 30 (2): 161-173 and Cox, R.W., Jesmanowicz, A., Hyde, J.S., 1995. Real-time functional magnetic resonance imaging. Magn. Reson. Med. 33 (2): 230-236]. Here, we propose to directly model the entire complex or bivariate data rather than just the magnitude-only data. A nonlinear multiple regression model is used to model activation of the complex signal, and a likelihood ratio test is derived to determine activation in each voxel. We investigate the performance of the model on a real dataset, then compare the magnitude-only and complex models under varying signal-to-noise ratios in a simulation study with varying activation contrast effects.

Algorithms↗

PubMed2004

Finding the maximum safe dose level for heteroscedastic data.

In this article we extend to the heteroscedastic setting the multiple stepwise test procedures proposed in Tamhane et al. [Tamhane. A. C., Dunnett, C. W., Green, J. W., Wetherington, J. F. (2001). Multiple test procedures for identifying a safe dose. J. Am. Statist. Assoc. 96:835-843] for finding the maximum safe dose. Toxicological data are often heteroscedastic; therefore, the extensions given herein should be highly useful in practice. Simulations are performed to study the Type I familywise error rate and power properties of the procedures. A real data example is given to illustrate the procedures.

Algorithms↗

PubMed2002

Accurate critical constants for the one-sided approximate likelihood ratio test of a normal mean vector when the covariance matrix is estimated.

Tang, Gnecco, and Geller (1989, Biometrika 76, 577-583) proposed an approximate likelihood ratio (ALR) test of the null hypothesis that a normal mean vector equals a null vector against the alternative that all of its components are nonnegative with at least one strictly positive. This test is useful for comparing a treatment group with a control group on multiple endpoints, and the data from the two groups are assumed to follow multivariate normal distributions with different mean vectors and a common covariance matrix (the homoscedastic case). Tang et al. derived the test statistic and its null distribution assuming a known covariance matrix. In practice, when the covariance matrix is estimated, the critical constants tabulated by Tang et al. result in a highly liberal test. To deal with this problem, we derive an accurate small-sample approximation to the null distribution of the ALR test statistic by using the moment matching method. The proposed approximation is then extended to the heteroscedastic case. The accuracy of both the approximations is verified by simulations. A real data example is given to illustrate the use of the approximations.

Biometry↗