A platform trial approach to proof-of-concept (POC) studies in autism spectrum disorder: Autism spectrum POC initiative (ASPI).

Background: Over the past decade, autism spectrum disorder (ASD) research has blossomed, and multiple clinical trials have tested potential interventions, with varying results and no clear demonstration of efficacy. Lack of clarity concerning appropriate biological mechanisms to target and lack of sensitive, objective tools to identify subgroups and measure symptom changes have hampered the efforts to develop treatments. A platform trial for proof-of-concept studies in ASD could help address these issues. A major goal of a platform trial is to find the best treatment in the most expeditious manner, by simultaneously investigating multiple treatments, using specialized statistical tools for allocation and analysis. We describe the setup of a platform trial and perform simulations to evaluate the operating characteristics under several scenarios. We use the Autism Behavior Inventory (ABI), a psychometrically validated web-based rating scale to measure the change in ASD core and associated symptoms. Methods: Detailed description of the setup, conduct, and decision-making rules of a platform trial are explained. Simulations of a virtual platform trial for several scenarios are performed to compare operating characteristics. The success and futility criteria for treatments are based on a Bayesian posterior probability model. Results: Overall, simulation results show the potential gain in terms of statistical properties especially for improved decision-making ability, while careful planning is needed due to the complexities of a platform trial. Conclusions: Autism research, shaped particularly by its heterogeneity, may benefit from the platform trial approach for POC clinical studies.

Economic Evaluation of Cost and Time Required for a Platform Trial vs Conventional Trials.

Importance: Platform trial design allows the introduction of new interventions after the trial is initiated and offers efficiencies to clinical research. However, limited guidance exists on the economic resources required to establish and maintain platform trials. Objective: To compare cost (US dollars) and time requirements of conducting a platform trial vs a series of conventional (nonplatform) trials using a real-life example. Design, Setting, and Participants: For this economic evaluation, an online survey was administered to a group of international experts (146 participants) with publication records of platform trials to elicit their opinions on cost and time to set up and conduct platform, multigroup, and 2-group trials. Using the reported entry dates of 10 interventions into Systemic Therapy in Advancing Metastatic Prostate Cancer: Evaluation of Drug Efficacy, the longest ongoing platform trial, 3 scenarios were designed involving a single platform trial (scenario 1), 1 multigroup followed by 5 2-group trials (scenario 2), and a series of 10 2-group trials (scenario 3). All scenarios started with 5 interventions, then 5 more interventions were either added to the platform or evaluated independently. Simulations with the survey results as inputs were used to compare the platform vs conventional trial designs. Data were analyzed from July to September 2021. Exposure: Platform trial design. Main Outcomes and Measures: Total trial setup and conduct cost and cumulative duration. Results: Although setup time and cost requirements of a single trial were highest for the platform trial, cumulative requirements of setting up a series of multiple trials in scenarios 2 and 3 were larger. Compared with the platform trial, there was a median (IQR) increase of 216.7% (202.2%-242.5%) in cumulative setup costs for scenario 2 and 391.1% (365.3%-437.9%) for scenario 3. In terms of total cost, there was a median (IQR) increase of 17.4% (12.1%-22.5%) for scenario 2 and 57.5% (43.1%-69.9%) for scenario 3. There was a median (IQR) increase in cumulative trial duration of 171.1% (158.3%-184.3%) for scenario 2 and 311.9% (282.0%-349.1%) for scenario 3. Cost and time reductions in the platform trial were observed in both the initial and subsequently evaluated interventions. Conclusions and Relevance: Although setting up platform trials can take longer and be costly, the findings of this study suggest that having a single infrastructure can improve efficiencies with respect to costs and efforts.

A simulation study of outcome adaptive randomization in multi-arm clinical trials.

Randomizing patients among treatments with equal probabilities in clinical trials is the established method to obtain unbiased comparisons. In recent years, motivated by ethical considerations, many authors have proposed outcome adaptive randomization, wherein the randomization probabilities are unbalanced, based on interim data, to favor treatment arms having more favorable outcomes. While there has been substantial controversy regarding the merits and flaws of adaptive versus equal randomization, there has not yet been a systematic simulation study in the multi-arm setting. A simulation study was conducted to evaluate four different Bayesian adaptive randomization methods and compare them to equal randomization in five-arm clinical trials. All adaptive randomization methods included an initial burn-in with equal randomization and some combination of other modifications to avoid extreme randomization probabilities. Trials either with or without a control arm were evaluated, using designs that may terminate arms early for futility and select one or more experimental treatments at the end. The designs were evaluated under a range of scenarios and sample sizes. For trials with a control arm and maximum same size 250 or 500, several commonly used adaptive randomization methods have very low probabilities of correctly selecting a truly superior treatment. Of those studied, the only adaptive randomization method with desirable properties has a burn-in with equal randomization and thereafter randomization probabilities restricted to the interval 0.10-0.90. Compared to equal randomization, this method has a favorable sample size imbalance but lower probability of correctly selecting a superior treatment. In multi-arm trials, compared to equal randomization, several commonly used adaptive randomization methods give much lower probabilities of selecting superior treatments. Aside from randomization method, conducting a multi-arm trial without a control arm may lead to very low probabilities of selecting any superior treatments if differences between the treatment success probabilities are small.

SARC009: Phase 2 study of dasatinib in patients with previously treated, high-grade, advanced sarcoma.

BACKGROUND: Dasatinib exhibited activity in preclinical models of sarcoma. The Sarcoma Alliance for Research through Collaboration (SARC) conducted a multicenter, phase 2 trial of dasatinib in patients with advanced sarcoma. METHODS: Patients received dasatinib twice daily. The primary objective was to estimate the clinical benefit rate (CBR) (complete response or partial response within 6 months or stable disease duration of ≥6 months) with a target of ≥25%. Patients were enrolled into 1 of 7 different cohorts and assessed by imaging every 8 weeks using Choi criteria tumor response and a Bayesian hierarchical design. For each subtype, enrollment was stopped after a minimum of 9 patients were treated if there was a <1% chance the CBR was ≥25%. RESULTS: A total of 200 patients were enrolled. Accrual was stopped early in 5 cohorts because of low CBR. The leiomyosarcoma (LMS) and undifferentiated pleomorphic sarcoma (UPS) cohorts fully accrued and 6 of 47 and 8 of 42 evaluable patients, respectively, exhibited clinical benefit. The probability that the CBR was ≥25% in the LMS and UPS cohorts was 0.008 and 0.10, respectively. The median progression-free survival ranged from 0.9 months in patients with rhabdomyosarcoma to 2.2 months in patients with LMS. The median overall survival was 8.6 months. The most frequent adverse events were constitutional, gastrointestinal, and respiratory, and 36% of patients required dose reduction for toxicity. Serious adverse events attributed to therapy occurred in 11% of patients. CONCLUSIONS: Dasatinib may have activity in patients with UPS but is inactive as a single agent in the other sarcoma subtypes included herein. The Bayesian design allowed for the early termination of accrual in 5 subtypes because of lack of drug activity.

Adaptive Design: Results of 2012 Survey on Perception and Use.

Adaptive designs are increasingly used in clinical trials. The Drug Information Association's Adaptive Design Scientific Working Group (ADSWG) works to foster collaboration among regulatory agencies, academia, and pharmaceutical and biotech companies to further the science of adaptive clinical development. The ADSWG Survey Subteam has collected data on the usage of adaptive designs in clinical research from multiple sources, including a recent ADSWG survey regarding the perception and usage of adaptive designs in academia and industry for studies between 2008 and 2011, as well as barriers to usage; a literature review examining publications of adaptive design methodology and usage between 2000 and 2011; and a trial registry review of adaptive design references from 1996 to 2011. The comprehensive results of the ADSWG 2012 survey are provided in this article with comparisons to our previous 2008 survey, the literature and registry reviews, and recent surveys carried out by the US Food and Drug Administration (FDA) and the European Medicines Agency. Results of the ADSWG 2012 survey illustrate that industry and academia are showing more enthusiasm for adaptive trials, accompanied by an increase in the number of trials using designs described as less well understood in the FDA draft guidance on adaptive designs, published in 2010. The increased use of these methods in exploratory trials is consistent with the FDA draft guidance. The survey also identified several examples of successful marketing applications supported by confirmatory trials utilizing adaptive designs that were considered, at least at the time of the draft guidance, as less well understood. While some of the technological barriers to adaptive design usage identified in the 2008 survey are now less common, there are several important persistent barriers to usage. Organizations can help overcome these barriers through education, preplanning, and early engagement in discussions with the regulators.

Adjuvant therapy for high-grade, uterus-limited leiomyosarcoma: results of a phase 2 trial (SARC 005).

BACKGROUND: Between 30% and 50% of women who have high-grade uterine leiomyosarcoma (uLMS) limited to the uterus at diagnosis remain progression-free at 2 years. Adjuvant pelvic radiation does not improve outcome. The objective of the current study was to determine the 2-year and 3-year progression-free survival (PFS) among a prospective cohort of women who received adjuvant gemcitabine plus docetaxel followed by doxorubicin. METHODS: Women with uterus-limited, high-grade uLMS and adequate organ function were eligible. Within 12 weeks of complete resection and after confirmation that they had no evidence of disease on computed tomography (CT) images, the patients received 4 cycles of fixed-dose-rate gemcitabine plus docetaxel. Those who were confirmed disease-free on CT scans after cycle 4 received 4 cycles of doxorubicin. CT imaging for recurrence was performed every 3 months for 2 years, then every 6 months for 3 years. RESULTS: In total, 47 women were enrolled (46 evaluable) in 3 years. Characteristics included a median age of 53 years; 1988 International Federation of Gynecology and Obstetrics stage I disease in 81% of patients, stage II disease in 15%, and serosa-only stage IIIA disease in 4%; American Joint Committee on Cancer stage II disease in 13% of patients and stage III disease in 87%; a median tumor size of 8 cm (range, 2.5-30 cm); and a median mitotic rate of 18 mitoses per 10 high-power fields (range, 5-83 mitoses per 10 high-power fields). At a median follow-up of 39.8 months, 21 of 46 patients developed recurrent disease (45.7%). The median time to recurrence was 27.4 months (range, 3-40 months). Seventy-eight percent of patients (95% confidence interval, 67%-91%) were progression-free at 2 years, and 57% (95% confidence interval, 44%-74%) were progression-free at 3 years. The median PFS was not reached and exceeded 36 months. CONCLUSIONS: Among women with high-grade, uterus-limited uLMS who received treatment with adjuvant gemcitabine plus docetaxel followed by doxorubicin, 78% remained progression-free at 2 years, and 57% remained progression-free at 3 years. A randomized trial of adjuvant chemotherapy versus observation to determine whether adjuvant chemotherapy can improve survival in women with uterus-limited uLMS is underway.

Phase II study of sequential gemcitabine followed by docetaxel for recurrent Ewing sarcoma, osteosarcoma, or unresectable or locally recurrent chondrosarcoma: results of Sarcoma Alliance for Research Through Collaboration Study 003.

BACKGROUND: Gemcitabine and docetaxel have a broad spectrum of clinical activity in patients with carcinoma. The Sarcoma Alliance for Research Through Collaboration conducted a phase II trial of gemcitabine in combination with docetaxel in children and adults with recurrent Ewing sarcoma (EWS), osteosarcoma (OS), or unresectable or recurrent chondrosarcoma. The primary objective was to determine the objective response rate using response evaluation criteria in solid tumors (RECIST). METHODS: Gemcitabine (675 mg/m2 i.v. over 90 minutes on days 1 and 8) was administered in combination with docetaxel (75 mg/m2 i.v. over 1 hour on day 8) every 21 days. All patients received filgrastim or pegfilgrastim. A Bayesian formulation was used to determine the probability of achieving the target response rate for each subtype-0.35 for EWS and OS or 0.20 for chondrosarcoma. If the probability of achieving the target response rate was <0.05, the combination was considered inactive. Toxicity was graded according to common terminology criteria for adverse events (CTCAE), version 3.0. RESULTS: Fifty-three eligible patients were enrolled in the three subtype groups-OS (n = 14), EWS (n = 14), and chondrosarcoma (n = 25). Toxicities included neutropenia, thrombocytopenia, fatigue, dyspnea, bronchospasm, edema, neuropathy, and liver function abnormalities. Dose modification for toxicity was required for eight patients during cycle 1 and 16 patients in subsequent cycles. Seven patients withdrew from therapy as a result of toxicity. No complete responses were observed. Partial responses were observed in OS (n = 1), EWS (n = 2), and chondrosarcoma (n = 2) patients. CONCLUSION: Gemcitabine in combination with docetaxel was associated with a probability of reaching the target 35% response rate of <5% in OS patients and 5.6% in EWS patients; the probability of reaching a 20% response rate in chondrosarcoma patients was 14%. DISCUSSION: The Bayesian formulation permitted estimation of the probability of achieving the target response rate for each subtype after each response evaluation. By allowing multiple looks at the data, this design stopped the trial after considering the probability of achieving the target response rate and accrual rate. Because this design did not specify a rule for declaring the treatment as "active", a direct comparison with a standard two-stage phase II design is not appropriate. The decision to close the EWS and chondrosarcoma subtype arms was based, in part, on slow accrual and was supported by the low probability of achieving the target response rate. The rate of enrollment, rather than the statistical design, had a significant effect on the trial duration.

Prophylaxis of graft-versus-host disease in unrelated donor transplantation with pentostatin, tacrolimus, and mini-methotrexate: a phase I/II controlled, adaptively randomized study.

PURPOSE: Acute graft-versus-host disease (aGVHD) is a major cause of morbidity and mortality after matched unrelated, related, or mismatched related donor hematopoietic stem-cell transplantation (HSCT). Improved GVHD prevention methods are needed. Pentostatin, an adenosine deaminase inhibitor, leads to lymphocyte depletion with low risk of myelosuppression. We hypothesized that addition of pentostatin to GVHD prophylaxis with tacrolimus and mini-methotrexate may improve outcomes, and we conducted a Bayesian adaptively randomized, controlled, dose-finding study, taking into account toxicity and efficacy. PATIENTS AND METHODS: Success was defined as the patient being alive, engrafted, in remission, without GVHD 100 days post-HSCT and no grade ≥ 3 GVHD at any time. Patients were randomly assigned to pentostatin doses of 0, 0.5, 1.0, 1.5, and 2.0 mg/m(2) with drug administered on HSCT days 8, 15, 22, and 30. Eligible patients were recipients of mismatched related (n = 10) or unrelated (n = 137) donor HSCT. RESULTS: Median age was 47 years. Thirty-seven, 10, 29, 61, and 10 patients were assigned to the control and four treatment groups, respectively, with comparable baseline characteristics. Pentostatin doses of 1.0 and 1.5 mg/m(2) had the highest success rates (69.0% and 70.5%) versus control (54.1%). The posterior probabilities that the success rates were greater with 1.5 mg/m(2) or 1.0 mg/m(2) versus control are 0.944 and 0.821, respectively. Hepatic aGVHD rates were 0%, 17.2%, and 11.1%, respectively, for 1.5 mg/m(2), 1.0 mg/m(2), and control groups. No grades 3 to 4 aGVHD occurred in 11 HLA-mismatched recipients in the 1.5 mg/m(2) group. CONCLUSION: Pentostatin increased the likelihood of success as defined here, and should be further investigated in larger randomized, confirmatory studies.

Efficacy of imatinib in aggressive fibromatosis: Results of a phase II multicenter Sarcoma Alliance for Research through Collaboration (SARC) trial.

PURPOSE: Aggressive fibromatoses (AF; desmoid tumors) are rare clonal neoplastic proliferations of connective tissues that can be locally aggressive despite wide surgical resection and/or radiation therapy. The Sarcoma Alliance for Research through Collaboration (SARC) initiated a prospective phase II trial to investigate the outcome of patients treated with imatinib, a multiple tyrosine kinase inhibitor, in patients with AF, or 1 of 10 sarcoma subtypes. Here, we report specifically on the outcome of patients with AF as well as evaluations undertaken to examine the mechanism of imatinib. EXPERIMENTAL DESIGN: Patients ≥10 years old with desmoid tumors that were not curable by surgical management or in whom curative surgery would lead to undesirable functional impairment were eligible. Imatinib was prescribed at 300 mg twice daily [body surface area (BSA) ≥ 1.5 m(2)], 200 mg twice daily (BSA = 1.0-1.49 m(2)), or 100 mg twice daily (BSA < 1.0 m(2)). Response outcomes at 2 and 4 months were assessed. Tissue specimens were analyzed by immunohistochemistry for expression of cKIT, platelet-derived growth factor receptor α (PDGFRα), PDGFRß, AKT, PTEN, FKHR, and ß-catenin. Tumor DNA was analyzed for PDGFRα exon 18 and APC mutations by allelic discrimination PCR. RESULTS: Fifty-one patients were enrolled. The median number of prior regimens was 1. Kaplan-Meier estimates of 2- and 4-month progression-free survival rates were 94% and 88%, respectively, and 1-year progression-free survival was 66%. Objective response rate was 6% (3 of 51). Expression and polymorphisms of target proteins were identified in tissue samples, but no significant correlation with outcome was observed using the samples available. CONCLUSION: Imatinib may have a role in the management of unresectable or difficult to resect desmoid tumors.

Phase II multicenter trial of imatinib in 10 histologic subtypes of sarcoma using a bayesian hierarchical statistical model.

PURPOSE The purpose of this trial was to assess the efficacy of imatinib in patients with one of 10 different subtypes of advanced sarcoma. PATIENTS AND METHODS Eligible patients were treated daily with imatinib dosed at 300 mg twice a day (for body-surface area > or = 1.5 m(2)). The primary end point was response (clinical benefit response [CBR]), defined as complete (CR) or partial response (PR) at 2 months, or stable disease, CR, or PR at 4 months. Rules for early termination within each disease type were based on a Bayesian hierarchical probability model (BHM) accounting for correlation of the responses of the 10 subtypes. Available tissue samples were analyzed for molecules within the KIT/platelet-derived growth factor receptor (PDGFR) signal transduction pathway. Results One hundred eighty-five assessable patients with one of 10 subtypes of sarcoma were treated. One CR and three PRs were achieved. A CBR was achieved in 28 patients treated overall and by subtype: two angiosarcomas (n = 16), 0 Ewing (n = 13), one fibrosarcoma (n = 12), six leiomyosarcomas (n = 29), seven liposarcomas (n = 31), three malignant fibrous histiocytomas (n = 30), five osteosarcomas (n = 27), one malignant peripheral-nerve sheath tumor (n = 7), 0 rhabdomyosarcoma (n = 2), and three synovial sarcomas (n = 22). Variable expression and mutations within the KIT/PDGFR pathway were observed. CONCLUSION This is the first phase II study of a new agent in sarcoma to include sufficient patients with each of the common histologic subtypes to permit generalizable conclusions. The BHM is an effective method for studying rare diseases and their subtypes, when it is reasonable to assume that their response rates are exchangeable. Although rare dramatic responses were seen, imatinib is not an active agent in advanced sarcoma in these subtypes.

Bayesian model averaging in meta-analysis: vitamin E supplementation and mortality.

CONTEXT: The strength and relevance of a meta-analysis depends on the validity of the statistical methods used. Of special importance is appropriately assessing different sources of variability. Many studies including meta-analyses have evaluated the efficacy and safety of vitamin E and have yielded varying results. Illuminating and resolving these disparities requires addressing study variability and model uncertainty. OBJECTIVE: To describe Bayesian meta-analysis methods for combining data from clinical trials, using recent studies that analyzed the relationship between vitamin E dose and all-cause mortality. DATA SOURCES: Studies used in a previously published meta-analysis appended by studies identified by a search of MEDLINE from August 2004 to December 2005 using the MeSH terms vitamin e and alpha tocopherol. INCLUSION CRITERIA: men and nonpregnant women; use of vitamin E alone or in combination with other vitamins or minerals; random allocation of participants to either vitamin E or a placebo or other control group; intervention and follow-up duration greater than 1 year; 10 or more deaths. DATA EXTRACTION: Independent data extraction by one author was reviewed and confirmed by a second author. Corresponding authors of the original publications were contacted when questions arose. DATA SYNTHESIS: Data collection included the number of patients and deaths, percent men, use of other vitamins or minerals, mean age, and length of follow-up. We combined study results using Bayesian hierarchical model averaging. Analyses used Markov chain Monte Carlo computational techniques. CONCLUSIONS: Vitamin E intake is unlikely to affect mortality regardless of dose. The Bayesian meta-analyses presented here are ideal for incorporating disparate sources of variability, including trial effect and model uncertainty.

Bayesian adaptive model selection for optimizing group sequential clinical trials.

This article presents a new approach to the problem of deriving an optimal design for a randomized group sequential clinical trial based on right-censored event times. We are motivated by the fact that, if the proportional hazards assumption is not met, then a conventional design's actual power can differ substantially from its nominal value. We combine Bayesian decision theory, Bayesian model selection and forward simulation (FS) to obtain a group sequential procedure that maintains targeted false-positive rate and power, under a wide range of true event time distributions. At each interim analysis, the method adaptively chooses the most likely model and then applies the decision bounds that are optimal under the chosen model. A simulation study comparing this design with three conventional designs shows that, over a wide range of distributions, our proposed method performs at least as well as each conventional design, and in many cases it provides a much smaller trial.

Bayesian designs to account for patient heterogeneity in phase II clinical trials.

PURPOSE OF REVIEW: Between-patient heterogeneity is very common in clinical trials. This complicates treatment evaluation, due to known prognostic subgroup effects or potential treatment-subgroup interactions. We review two Bayesian phase II clinical trial designs that account explicitly for patient heterogeneity. The first design uses analysis of covariance to assess treatment effects in subgroups known to have different prognoses. The second design uses a hierarchical model for settings where, a priori, the experimental treatment effects in the subgroups are assumed to be exchangeable. RECENT FINDINGS: Compared with simpler designs that ignore patient heterogeneity, each design provides substantial improvements by reducing both false positive and false negative rates and focusing resources on subgroups where an experimental treatment is more likely to provide an advance over standard therapy. In either case, accounting for potential treatment-subgroup interactions is extremely important. SUMMARY: Due to the rapidly increasing number of potential new treatments to be evaluated clinically, increasing costs and limited resources, it is critically important to perform early phase clinical trials efficiently. The new Bayesian methods described here and other related methods provide efficient, broadly applicable tools to address these problems.

Accounting for patient heterogeneity in phase II clinical trials.

Phase II clinical trials typically are single-arm studies conducted to decide whether an experimental treatment is sufficiently promising, relative to standard treatment, to warrant further investigation. Many methods exist for conducting phase II trials under the assumption that patients are homogeneous. In the presence of patient heterogeneity, however, these designs are likely to draw incorrect conclusions. We propose a class of model-based Bayesian designs for single-arm phase II trials with a binary or time-to-event outcome and two or more prognostic subgroups. The designs' early stopping rules are subgroup specific and allow the possibility of terminating some subgroups while continuing others, thus providing superior results when compared with designs that ignore treatment-subgroup interactions. Because our formulation requires informative priors on standard treatment parameters and subgroup main effects, and non-informative priors on experimental treatment parameters and treatment-subgroup interactions, we provide an algorithm for computing prior hyperparameter values. A simulation study is presented and the method is illustrated by a chemotherapy trial in acute leukemia.

Randomized phase II study of gemcitabine and docetaxel compared with gemcitabine alone in patients with metastatic soft tissue sarcomas: results of sarcoma alliance for research through collaboration study 002 [corrected].

PURPOSE: Gemcitabine as a single agent and the combination of gemcitabine and docetaxel have activity in patients with metastatic soft tissue sarcoma. To determine if the addition of docetaxel to gemcitabine improved clinical outcome of patients with metastatic soft tissue sarcomas, we compared a fixed dose rate infusion of gemcitabine versus a lower dose of gemcitabine with docetaxel. PATIENTS AND METHODS: In this open-label phase II clinical trial, the primary end point was tumor response, defined as complete or partial response or stable disease lasting at least 24 weeks. A Bayesian adaptive randomization procedure was used to produce an imbalance in the randomization in favor of the superior treatment, accounting for treatment-subgroup interactions. RESULTS: One hundred nineteen of 122 randomly assigned patients had assessable outcomes. The adaptive randomization assigned 73 patients (60%) to gemcitabine-docetaxel and 49 patients (40%) to gemcitabine alone, indicating gemcitabine-docetaxel was superior. The objective Response Evaluation Criteria in Solid Tumors response rates were 16% (gemcitabine-docetaxel) and 8% (gemcitabine). Given the data, the posterior probabilities that gemcitabine-docetaxel was superior for progression-free and overall survival were 0.98 and 0.97, respectively. Median progression-free survival was 6.2 months for gemcitabine-docetaxel and 3.0 months for gemcitabine alone; median overall survival was 17.9 months for gemcitabine-docetaxel and 11.5 months for gemcitabine. The posterior probability that patients receiving gemcitabine-docetaxel had a shorter time to discontinuation for toxicity compared with gemcitabine alone was .999. CONCLUSION: Gemcitabine-docetaxel yielded superior progression-free and overall survival to gemcitabine alone, but with increased toxicity. Adaptive randomization is an effective method to reduce the number of patients receiving inferior therapy.

Practical Bayesian adaptive randomisation in clinical trials.

While randomisation is the established method for obtaining scientifically valid treatment comparisons in clinical trials, it sometimes is at odds with what physicians feel is good medical practice. If a physician favours one treatment over another based on personal experience or published data, it may be more appropriate ethically for that physician to use the favoured treatment, rather than enrolling patients on a randomised trial. Still, the randomised trial may later show the physician's favoured treatment to be inferior. This paper reviews a statistical method, Bayesian adaptive randomisation, that provides a practical compromise between the scientific ideal of conventional randomisation and choosing each patient's treatment based on a personal preference that may prove to be incorrect. The method will first be illustrated by a simple hypothetical example, then by a recent trial in which patients with unresectable soft tissue sarcoma were adaptively randomised between two chemotherapy regimens.

Continuous Bayesian adaptive randomization based on event times with covariates.

In comparative clinical trials, the randomization probabilities may be unbalanced adaptively by utilizing the interim data available at each patient's entry time to favour the treatment or treatments having comparatively superior outcomes. This is ethically appealing because, on average, more patients are assigned to the more successful treatments. Consequently, physicians are more likely to enroll patients onto trials where the randomization is outcome-adaptive rather than balanced in the conventional manner. Outcome-adaptive methods based on a binary variable may be applied by reducing an event time to the indicator of the event's occurrence within a predetermined time interval. This results in a loss of information, however, since it ignores the censoring times of patients who have not experienced the event but whose evaluation interval is not complete. This paper proposes and compares exact and approximate Bayesian outcome-adaptive randomization procedures based on time-to-event outcomes. The procedures account for baseline prognostic covariates, and they may be applied continuously over the course of the trial. We illustrate these methods by application to a phase II selection trial in acute leukaemia. A simulation study in the context of this trial is presented.

Covariate-adjusted adaptive randomization in a sarcoma trial with multi-stage treatments.

We present a Bayesian design for a multi-centre, randomized clinical trial of two chemotherapy regimens for advanced or metastatic unresectable soft tissue sarcoma. After randomization, each patient receives up to four stages of chemotherapy, with the patient's disease evaluated after each stage and categorized on a trinary scale of severity. Therapy is continued to the next stage if the patient's disease is stable, and is discontinued if either tumour response or treatment failure is observed. We assume a probability model that accounts for baseline covariates and the multi-stage treatment and disease evaluation structure. The design uses covariate-adjusted adaptive randomization based on a score that combines the patient's probabilities of overall treatment success or failure. The adaptive randomization procedure generalizes the method proposed by Thompson (1933) for two binomial distributions with beta priors. A simulation study of the design in the context of the sarcoma trial is presented.

Hierarchical Bayesian approaches to phase II trials in diseases with multiple subtypes.

We propose a methodology for conducting phase II clinical trials in settings where the disease is categorized into multiple subtypes. A hierarchical Bayesian model is assumed for treatment effects within the subtypes. The hierarchical model, which is tailored to each particular application, allows treatment effects to differ across subtypes while assuming a priori that the effects are exchangeable and correlated. Two applications are described. The first is a trial of imatinib for sarcoma in which treatment activity is characterized by a binary indicator of tumour response. The second is a phase II trial of a new preparative regimen for allogeneic bone marrow transplantation in patients with haematologic malignancies, with treatment effect characterized by the mean time from transplant to disease progression or death. The applications illustrate how the hierarchical Bayesian model borrows strength across subtypes.