BAYESIAN PHASE I DOSE FINDING IN CANCER TRIALS

This dissertation explores phase I dose-finding designs in cancer trials from three perspectives: the alternative Bayesian dose-escalation rules, a design based on a time-to-dose-limiting toxicity (DLT) model, and a design based on a discrete-time multi-state (DTMS) model. We list alternative Bayesian dose-escalation rules and perform a simulation study for the intra-rule and inter-rule comparisons based on two statistical models to identify the most appropriate rule under certain scenarios. We provide evidence that all the Bayesian rules outperform the traditional ``3+3'' design in the allocation of patients and selection of the maximum tolerated dose. The design based on a time-to-DLT model uses patients' DLT information over multiple treatment cycles in estimating the probability of DLT at the end of treatment cycle 1. Dose-escalation decisions are made whenever a cycle-1 DLT occurs, or two months after the previous check point. Compared to the design based on a logistic regression model, the new design shows more safety benefits for trials in which more late-onset toxicities are expected. As a trade-off, the new design requires more patients on average. The design based on a discrete-time multi-state (DTMS) model has three important attributes: (1) Toxicities are categorized over a distribution of severity levels, (2) Early toxicity may inform dose escalation, and (3) No suspension is required between accrual cohorts. The proposed model accounts for the difference in the importance of the toxicity severity levels and for transitions between toxicity levels. We compare the operating characteristics of the proposed design with those from a similar design based on a fully-evaluated model that directly models the maximum observed toxicity level within the patients' entire assessment window. We describe settings in which, under comparable power, the proposed design shortens the trial. The proposed design offers more benefit compared to the alternative design as patient accrual becomes slower.

Clinical, microbiological, and immunological markers of gingival and periodontal disease induction and progression

The potential impact of periodontal disease, a suspected risk factor for systemic diseases, presents challenges for health promotion and disease prevention strategies. This study examined clinical, microbiological, and immunological factors in a disease model to identify potential biomarkers that may be useful in predicting the onset and severity of both inflammatory and destructive periodontal disease. This project used an historical cohort design based on data obtained from 47 adult, female nonhuman primates followed over a 6-year period for 5 unique projects where the ligature-induced model of periodontitis was utilized. Standardization of protocols for sample collection allowed for comparison over time. Bleeding and pocket depth measures were selected as the dependent variables of relevance to humans based upon the literature and historical observations. Exposure variables included supragingival plaque, attachment level, total bacteria, black-pigmented bacteria, Gram-negative and Gram-positive bacteria, total IgG and IgA in crevicular fluid, specific IgG antibody in both crevicular fluid and serum, and IgG antibody to four select pathogenic microorganisms. Three approaches were used to analyze the data from this study. The first approach tested for differences in the means of the response variables within the group and among longitudinal observations within the group at each time point. The second approach examined the relationship among the clinical, microbiological, and immunological variables using correlation coefficients and stratified analyses. Multivariable models using GEE for repeated measures were produced as a predictive description of the induction and progression of gingivitis and periodontal disease. The multivariable models for bleeding (gingivitis) include supragingival plaque, total bacteria and total IgG while the second also contains supragingival plaque, Gram-positive bacteria, and total IgG. Two multivariable models emerged for periodontal disease. One multivariable model contains plaque, total bacteria, total IgG and attachment level. The second model includes black-pigmented bacteria, total bacteria, antibody to Campylobacter rectus, and attachment level. Utilization of the nonhuman primate model to prospectively examine causal hypotheses can provide a focus for human research on the mechanisms of progression from health to gingivitis to periodontitis. Ultimately, causal theories can guide strategies to prevent disease initiation and reduce disease severity. ^

Bayesian adaptive designs for drug combinations in early phase clinical trials

Treating patients with combined agents is a growing trend in cancer clinical trials. Evaluating the synergism of multiple drugs is often the primary motivation for such drug-combination studies. Focusing on the drug combination study in the early phase clinical trials, our research is composed of three parts: (1) We conduct a comprehensive comparison of four dose-finding designs in the two-dimensional toxicity probability space and propose using the Bayesian model averaging method to overcome the arbitrariness of the model specification and enhance the robustness of the design; (2) Motivated by a recent drug-combination trial at MD Anderson Cancer Center with a continuous-dose standard of care agent and a discrete-dose investigational agent, we propose a two-stage Bayesian adaptive dose-finding design based on an extended continual reassessment method; (3) By combining phase I and phase II clinical trials, we propose an extension of a single agent dose-finding design. We model the time-to-event toxicity and efficacy to direct dose finding in two-dimensional drug-combination studies. We conduct extensive simulation studies to examine the operating characteristics of the aforementioned designs and demonstrate the designs' good performances in various practical scenarios.^