Bayesian doubly optimal group sequential designs for randomized clinical trials

When conducting a randomized comparative clinical trial, ethical, scientific or economic considerations often motivate the use of interim decision rules after successive groups of patients have been treated. These decisions may pertain to the comparative efficacy or safety of the treatments under study, cost considerations, the desire to accelerate the drug evaluation process, or the likelihood of therapeutic benefit for future patients. At the time of each interim decision, an important question is whether patient enrollment should continue or be terminated; either due to a high probability that one treatment is superior to the other, or a low probability that the experimental treatment will ultimately prove to be superior. The use of frequentist group sequential decision rules has become routine in the conduct of phase III clinical trials. In this dissertation, we will present a new Bayesian decision-theoretic approach to the problem of designing a randomized group sequential clinical trial, focusing on two-arm trials with time-to-failure outcomes. Forward simulation is used to obtain optimal decision boundaries for each of a set of possible models. At each interim analysis, we use Bayesian model selection to adaptively choose the model having the largest posterior probability of being correct, and we then make the interim decision based on the boundaries that are optimal under the chosen model. We provide a simulation study to compare this method, which we call Bayesian Doubly Optimal Group Sequential (BDOGS), to corresponding frequentist designs using either O'Brien-Fleming (OF) or Pocock boundaries, as obtained from EaSt 2000. Our simulation results show that, over a wide variety of different cases, BDOGS either performs at least as well as both OF and Pocock, or on average provides a much smaller trial. ^

Genetic analysis of factors regulating mesenchymal cell fates

Formation of cartilage and bone involves sequential processes in which undifferentiated mesenchyme aggregates into primordial condensations which subsequently grow and differentiate, resulting in morphogenesis of the adult skeleton. While much has been learned about the structural molecules which comprise cartilage and bone, little is known about the nuclear factors which regulate chondrogenesis and osteogenesis. MHox is a homeobox-containing gene which is expressed in the mesenchyme of facial, limb, and vertebral skeletal precursors during mouse embryogenesis. MHox expression has been shown to require epithelial-derived signals, suggesting that MHox may regulate the epithelial-mesenchymal interactions required for skeletal organogenesis. To determine the functions of MHox, we generated a loss-of-function mutation in the MHox gene. Mice homozygous for a mutant MHox allele exhibit defects of skeletogenesis, involving the loss or malformation of craniofacial, limb and vertebral skeletal structures. The affected skeletal elements are derived from the cranial neural crest, as well as somitic and lateral mesoderm. Analysis of the mutant phenotype during ontogeny demonstrated a defect in the formation or growth of chondrogenic and osteogenic precursors. These findings provide evidence that MHox regulates the formation of preskeletal condensations from undifferentiated mesenchyme. In addition, generation of mice doubly mutant for the MHox and S8 homeobox genes reveal that these two genes interact to control formation of the limb and craniofacial skeleton. Mice carrying mutant alleles for S8 and MHox exhibit an exaggeration of the craniofacial and limb phenotypes observed in the MHox mutant mouse. Thus, MHox and S8 are components of a combinatorial genetic code controlling generation of the skeleton of the skull and limbs. ^

A study of the relationship between physical activity and cardiovascular fitness and coronary heart disease risk factors in female adolescents

The association of measures of physical activity with coronary heart disease (CHD) risk factors in children, especially those for atherosclerosis, is unknown. The purpose of this study was to determine the association of physical activity and cardiovascular fitness with blood lipids and lipoproteins in pre-adolescent and adolescent girls.^ The study population was comprised of 131 girls aged 9 to 16 years who participated in the Children's Nutrition Research Center's Adolescent Study. The dependent variables, blood lipids and lipoproteins, were measured by standard techniques. The independent variables were physical activity measured as the difference between total energy expenditure (TEE) and basal metabolic rate (BMR), and cardiovascular fitness, VO$\rm\sb{2max}$(ml/min/kg). TEE was measured by the doubly-labeled water (DLW) method, and BMR by whole-room calorimetry. Cardiovascular fitness, VO$\rm\sb{2max}$(ml/min/kg), was measured on a motorized treadmill. The potential confounding variables were sexual maturation (Tanner breast stage), ethnic group, body fat percent, and dietary variables. A systematic strategy for data analysis was used to isolate the effects of physical activity and cardiovascular fitness on blood lipids, beginning with assessment of confounding and interaction. Next, from regression models predicting each blood lipid and controlling for covariables, hypotheses were evaluated by the direction and value of the coefficients for physical activity and cardiovascular fitness.^ The main result was that cardiovascular fitness appeared to be more strongly associated with blood lipids than physical activity. An interaction between cardiovascular fitness and sexual maturation indicated that the effect of cardiovascular fitness on most blood lipids was dependent on the stage of sexual maturation.^ A difference of 760 kcal/d physical activity (which represents the difference between the 25th and 75th percentile of physical activity) was associated with negligible differences in blood lipids. In contrast, a difference in 10 ml/min/kg of VO$\rm\sb{2max}$ or cardiovascular fitness (which represents the difference between the 25th and 75th percentile in cardiovascular fitness) in the early stages of sexual maturation was associated with an average positive difference of 15 mg/100 ml ApoA-1 and 10 mg/100 ml HDL-C. ^

Outcome-based adaptive randomization for binary data in clinical trials---Compare and contrast

Monte Carlo simulation has been conducted to investigate parameter estimation and hypothesis testing in some well known adaptive randomization procedures. The four urn models studied are Randomized Play-the-Winner (RPW), Randomized Pôlya Urn (RPU), Birth and Death Urn with Immigration (BDUI), and Drop-the-Loses Urn (DL). Two sequential estimation methods, the sequential maximum likelihood estimation (SMLE) and the doubly adaptive biased coin design (DABC), are simulated at three optimal allocation targets that minimize the expected number of failures under the assumption of constant variance of simple difference (RSIHR), relative risk (ORR), and odds ratio (OOR) respectively. Log likelihood ratio test and three Wald-type tests (simple difference, log of relative risk, log of odds ratio) are compared in different adaptive procedures. ^ Simulation results indicates that although RPW is slightly better in assigning more patients to the superior treatment, the DL method is considerably less variable and the test statistics have better normality. When compared with SMLE, DABC has slightly higher overall response rate with lower variance, but has larger bias and variance in parameter estimation. Additionally, the test statistics in SMLE have better normality and lower type I error rate, and the power of hypothesis testing is more comparable with the equal randomization. Usually, RSIHR has the highest power among the 3 optimal allocation ratios. However, the ORR allocation has better power and lower type I error rate when the log of relative risk is the test statistics. The number of expected failures in ORR is smaller than RSIHR. It is also shown that the simple difference of response rates has the worst normality among all 4 test statistics. The power of hypothesis test is always inflated when simple difference is used. On the other hand, the normality of the log likelihood ratio test statistics is robust against the change of adaptive randomization procedures. ^

Functions of GCN5 and P /CAF acetyltransferases in mammalian development

Histone acetyltransferases are important chromatin modifiers that function as transcriptional co-activators. The identification of the transcriptional regulator GCN5 as the first nuclear histone acetyltransferase in yeast directly linked chromatin remodeling to transcriptional regulation. Although emerging evidence suggests that acetyltransferases participate in multiple cellular processes, their roles in mammalian development remain undefined. In this study, I have cloned and characterized the mouse homolog of GCN5 and a closely related protein P/CAF that interacts with p300/CBP. In contrast to yeast GCN5, but similar to P/CAF, mouse GCN5 possesses an additional N-terminal domain that confers the ability to acetylate nucleosomal histones. GCN5 and P/CAF exhibit identical substrate specificity and both interact with p300/CBP. Interestingly, expression levels of GCN5 and P/CAF display a complementary pattern in mouse embryos and in adult tissues, suggesting that they have distinct tissue or developmental stage specific roles. To define the in vivo function of GCN5 and P/CAF, I have generated mice that are nullizygous for GCN5 or P/CAF. P/CAF null mice are viable and fertile with no gross morphological defects, indicating that P/CAF is dispensable for development and p300/CBP function in vivo. In contrast, mice lacking GCN5 die between 10.5–11 days of gestation. GCN5 null mice are severely retarded but have anterior ectopic outgrowth. Molecular marker analyses reveal that early mesoderm is formed in GCN5 null mice but further differentiation into distinct mesodermal lineages is perturbed. While presomitic mesoderm and chodamesoderm are missing in GCN5 mutant mice, extraembryonic tissues and lateral mesoderm are unaffected. This is consistent with our finding that GCN5 expression is absent in the heart and extraembryonic tissues but is uniform throughout the rest of the embryo. Remarkably, GCN5 mutant mice exhibit an unusually high incidence of apoptosis in the embryonic ectoderm and mesoderm. Finally, mice doubly null for GCN5 and P/CAF die much earlier than mice harboring the GCN5 mutation alone, suggesting that P/CAF and GCN5 share some overlapping function during embryogenesis. This work is the first study to show that specific acetyltransferase is important for cell survival as well as mesoderm differentiation or maintenance during early mammalian development. ^