Ying Kuen Cheung

Dose finding by the continual reassessment method

Fundamentals Introduction Dose Finding in Clinical Trials The Maximum Tolerated Dose An Overview of Methodology Bibliographic Notes Exercises and Further Results The Continual Reassessment Method Introduction One-Stage Bayesian CRM Two-Stage CRM Simulating CRM Trials Practical Modifications Bibliographic Notes Exercises and Further Results One-Parameter Dose-ToxicityModels Introduction -Equivalent Models Model Assumptions+ Proof of Theorem 4.1+ Exercises and Further Results Theoretical Properties Introduction Coherence Large-Sample Properties Proofs+ Exercises and Further Results Empirical Properties Introduction Operating Characteristics A Nonparametric Optimal Benchmark Exercises and Further Results Design Calibration Specifications of a CRM Design Introduction Specifying the Clinical Parameters A Roadmap for Choosing the Statistical Component The Trial-and-Error Approach: Two Case Studies Initial Guesses of Toxicity Probabilities Introduction Half-width ( ) of Indifferent Interval Calibration of 77 Case Study: The Bortezomib Trial Exercises and Further Results Least Informative Normal Prior Introduction Least Informative Prior Calibration of 93 Optimal Least Informative Model Revisiting the Bortezomib Trial Initial Design Introduction Ordering of Dose Sequences Building Reference Initial Designs Practical Issues Case Study: NeuSTART Exercises and Further Results CRM and Beyond The Time-to-Event CRM Introduction The Basic Approach Numerical Illustration Enrollment Scheduling Theoretical Properties+ Two-Stage Design BibliographicNotes Exercises and Further Results CRM withMultiparameter Models Introduction Curve-Free Methods Rigidity Two-Parameter CRM+ BibliographicNotes Exercise and Further Results When the CRM Fails Introduction Trade-Off Perspective of MTD Bivariate Dose Finding Stochastic Approximation Introduction The Past Literature The Present Relevance The Future Challenge Assumptions on M(x) and Y (x)+ Exercises and Further Results References Index

A novel, efficient, randomized selection trial comparing combinations of drug therapy for ALS

Combining agents with different mechanisms of action may be necessary for meaningful results in treating ALS. The combinations of minocycline-creatine and celecoxib-creatine have additive effects in the murine model. New trial designs are needed to efficiently screen the growing number of potential neuroprotective agents. Our objective was to assess two drug combinations in ALS using a novel phase II trial design. We conducted a randomized, double-blind selection trial in sequential pools of 60 patients. Participants received minocycline (100 mg)-creatine (10 g) twice daily or celecoxib (400 mg)-creatine (10 g) twice daily for six months. The primary objective was treatment selection based on which combination best slowed deterioration in the ALS Functional Rating Scale-Revised (ALSFRS-R); the trial could be stopped after one pool if the difference between the two arms was adequately large. At trial conclusion, each arm was compared to a historical control group in a futility analysis. Safety measures were also examined. After the first patient pool, the mean six-month decline in ALSFRS-R was 5.27 (SD=5.54) in the celecoxib-creatine group and 6.47 (SD=9.14) in the minocycline-creatine group. The corresponding decline was 5.82 (SD=6.77) in the historical controls. The difference between the two sample means exceeded the stopping criterion. The null hypothesis of superiority was not rejected in the futility analysis. Skin rash occurred more frequently in the celecoxib-creatine group. In conclusion, the celecoxib-creatine combination was selected as preferable to the minocycline-creatine combination for further evaluation. This phase II design was efficient, leading to treatment selection after just 60 patients, and can be used in other phase II trials to assess different agents.

Selecting promising ALS therapies in clinical trials

DOPA-RESPONSIVE DYSTONIC CAMPTOCORMIA To the Editor: We read Dr. van Gerpen’s article in which he reports a patient with camptocormia remarkably responsive to levodopa treatment.1 The authors report that having the patient walk backwards improved the position of the patient’s spine, which was equated to a sensory trick. It was concluded that camptocormia was due to dystonia because of the presence of diurnal fluctuation and that the camptocormia was exacerbated by prolonged activity. No electrophysiologic studies were carried out but response to levodopa during a 6-year follow-up led to dramatic and sustained improvement in his condition. There were no side effects reported including dyskinesias and a diagnosis of dopa-responsive dystonia was made. We reported a similar patient with axial dystonia.2 A 19-year-old girl had a long history of kyphoscoliosis and was misdiagnosed as idiopathic. She was offered corrective surgery on several occasions but refused. We were able to examine her and found some mild signs of dystonia on her leg and neck that worsened in the evening but improved with rest. We performed a trial with levodopa. Treatment with low doses of levodopa led to total remission of all symptoms within a month. More than 15 years later, she is asymptomatic as long as she continues levodopa treatment. At that time, we concluded that spinal curvature can be a major symptom of dopa-responsive dystonia and this should warrant its inclusion in the differential diagnoses of idiopathic kyphoscoliosis.2 Dr. van Gerpen’s case illustrates that the spectrum of dopa-responsive dystonia encompasses axial disturbances regardless of the manifestations (e.g., camptocormia, Pisa syndrome, kyphoscoliosis, or another abnormality). Given the marked response to levodopa, all of these possibilities should be considered.

Calibration of prior variance in the Bayesian continual reassessment method

The continual reassessment method (CRM) is an adaptive model-based design used to estimate the maximum tolerated dose in phase I clinical trials. Asymptotically, the method has been shown to select the correct dose given that certain conditions are satisfied. When sample size is small, specifying a reasonable model is important. While an algorithm has been proposed for the calibration of the initial guesses of the probabilities of toxicity, the calibration of the prior distribution of the parameter for the Bayesian CRM has not been addressed. In this paper, we introduce the concept of least informative prior variance for a normal prior distribution. We also propose two systematic approaches to jointly calibrate the prior variance and the initial guesses of the probability of toxicity at each dose. The proposed calibration approaches are compared with existing approaches in the context of two examples via simulations. The new approaches and the previously proposed methods yield very similar results since the latter used appropriate vague priors. However, the new approaches yield a smaller interval of toxicity probabilities in which a neighboring dose may be selected.

Cognitive correlates of white matter lesion load and brain atrophy: The Northern Manhattan Study

Objective: We investigated white matter lesion load and global and regional brain volumes in relation to domain-specific cognitive performance in the stroke-free Northern Manhattan Study (NOMAS) population. Methods: We quantified white matter hyperintensity volume (WMHV), total cerebral volume (TCV), and total lateral ventricular (TLV) volume, as well as hippocampal and cortical gray matter (GM) lobar volumes in a subgroup. We used general linear models to examine MRI markers in relation to domain-specific cognitive performance, adjusting for key covariates. Results: MRI and cognitive data were available for 1,163 participants (mean age 70 ± 9 years; 60% women; 66% Hispanic, 17% black, 15% white). Across the entire sample, those with greater WMHV had worse processing speed. Those with larger TLV volume did worse on episodic memory, processing speed, and semantic memory tasks, and TCV did not explain domain-specific variability in cognitive performance independent of other measures. Age was an effect modifier, and stratified analysis showed that TCV and WMHV explained variability in some domains above age 70. Smaller hippocampal volume was associated with worse performance across domains, even after adjusting for APOE ε4 and vascular risk factors, whereas smaller frontal lobe volumes were only associated with worse executive function. Conclusions: In this racially/ethnically diverse, community-based sample, white matter lesion load was inversely associated with cognitive performance, independent of brain atrophy. Lateral ventricular, hippocampal, and lobar GM volumes explained domain-specific variability in cognitive performance.

Sequential Implementation of Stepwise Procedures for Identifying the Maximum Tolerated Dose

This article considers the problem of finding the maximum tolerated dose (MTD) of a drug in human trials. The MTD is defined as the maximum test dose with toxicity probability less than or equal to a target toxicity rate. We adopt the multiple test framework, with step-down tests used in an escalation stage and step-up tests used in a deescalation stage, to allow sequential dose assignments for ethical purposes. By formulating the estimation problem as a testing problem, the proposed procedures formally control the error probability of selecting an unsafe dose. In addition, we can control the probability of correctly selecting the MTD under a parameter subspace where no toxicity probability lies in an interval bracketed by the target toxicity rate and an unacceptably high toxicity rate, the so-called “indifference zone.” This frequentist property, which is currently lacking in the conduct of dose-finding trials in humans, is appealing from a regulatory standpoint. We give the general expressions of the selection probabilities and apply some common statistical tests to the stepwise procedure. The design parameters are calibrated so that the average number of patients receiving an overdose is kept low. From a practical viewpoint, stepwise tests are simple and easy to understand, and the sequential implementation operates in a manner similar to the traditional algorithm familiar to clinicians. Extensive simulations illustrate that our methods yield good, competitive operating characteristics under a wide range of scenarios with realistic sample size and performs well even in situations in which other existing methods may fail, namely when the dose–toxicity curve is flat up to the targeted MTD.

Lipid profile components and subclinical cerebrovascular disease in the northern Manhattan study.

Background: Subclinical cerebrovascular disease has been associated with multiple adverse events related to aging, including stroke and dementia. The modifiable risk factors for subclinical cerebrovascular disease beyond hypertension have not been well characterized. Our objective was to examine the association between baseline, and changes over time, in lipid profile components and subclinical cerebrovascular disease on magnetic resonance imaging (MRI). Methods: Fasting plasma lipids were collected on participants in the Northern Manhattan Study, a prospective cohort study examining risk factors for cardiovascular disease in a multiethnic elderly urban-dwelling population. A subsample of the cohort underwent brain MRI between 2003 and 2008 (a median of 6.2 years, range = 0-14, after enrollment), when repeat fasting lipids were obtained. We used lipid profile components at the time of initial enrollment (n = 1,256 with lipids available) as categorical variables, as well as change in clinical categories over the two measures (n = 1,029). The main outcome measures were (1) total white matter hyperintensity volume (WMHV) using linear regression and (2) silent brain infarcts (SBI) using logistic regression. Results: None of the plasma lipid profile components at the time of enrollment were associated with WMHV. The association between baseline lipids and WMHV was, however, modified by apolipoprotein E (apoE) status (χ2 with 2 degrees of freedom, p = 0.03), such that among apoE4 carriers those with total cholesterol (TC) ≥200 mg/dl had a trend towards smaller WMHV than those with TC <200 mg/dl (difference in logWMHV -0.19, p = 0.07), while there was no difference among apoE3 carriers. When examining the association between WMHV and change in lipid profile components we noted an association with change in high-density lipoprotein cholesterol (HDL-C, >50 mg/dl for women, >40 mg/dl for men) and TC. A transition from low-risk HDL-C (>50 mg/dl for women, >40 mg/dl for men) at baseline to high-risk HDL-C at the time of MRI (vs. starting and remaining low risk) was associated with greater WMHV (difference in logWMHV 0.34, p value 0.03). We noted a similar association with transitioning to a TC ≥200 mg/dl at the time of MRI (difference in logWMHV 0.25, p value 0.006). There were no associations with baseline or change in lipid profile components with SBI. Conclusions: The association of plasma lipid profile components with greater WMHV may depend on apoE genotype and worsening HDL and TC risk levels over time.

Sequential multiple assignment randomized trial (SMART) with adaptive randomization for quality improvement in depression treatment program

Implementation study is an important tool for deploying state-of-the-art treatments from clinical efficacy studies into a treatment program, with the dual goals of learning about effectiveness of the treatments and improving the quality of care for patients enrolled into the program. In this article, we deal with the design of a treatment program of dynamic treatment regimens (DTRs) for patients with depression post-acute coronary syndrome. We introduce a novel adaptive randomization scheme for a sequential multiple assignment randomized trial of DTRs. Our approach adapts the randomization probabilities to favor treatment sequences having comparatively superior Q-functions used in Q-learning. The proposed approach addresses three main concerns of an implementation study: it allows incorporation of historical data or opinions, it includes randomization for learning purposes, and it aims to improve care via adaptation throughout the program. We demonstrate how to apply our method to design a depression treatment program using data from a previous study. By simulation, we illustrate that the inputs from historical data are important for the program performance measured by the expected outcomes of the enrollees, but also show that the adaptive randomization scheme is able to compensate poorly specified historical inputs by improving patient outcomes within a reasonable horizon. The simulation results also confirm that the proposed design allows efficient learning of the treatments by alleviating the curse of dimensionality.

High-density lipoprotein subfractions and carotid plaque: The Northern Manhattan Study

OBJECTIVE The objective of this cross-sectional analysis was to investigate the relation between two major high-density lipoprotein cholesterol (HDL-C) subfractions (HDL2-C and HDL3-C) and carotid plaque in a population based cohort. METHODS We evaluated 988 stroke-free participants (mean age 66 ± 8 years; 40% men; 66% Hispanic and 34% Non-Hispanic) with available data on HDL subfractions using precipitation method and carotid plaque area and thickness assessed by a high-resolution 2D ultrasound. The associations between HDL-C subfractions and plaque measurements were analyzed by quantile regression. RESULTS Plaque was present in 56% of the study population. Among those with plaque, the mean ± SD plaque area was 19.40 ± 20.46 mm² and thickness 2.30 ± 4.45 mm. The mean ± SD total HDL-C was 46 ± 14 mg/dl, HDL2-C 14 ± 8 mg/dl, and HDL3-C 32 ± 8 mg/dl. After adjusting for demographics and vascular risk factors, there was an inverse association between HDL3-C and plaque area (per mg/dl: beta = -0.26 at the 75th percentile, p = 0.001 and beta = -0.32 at the 90th percentile, p = 0.02). A positive association was observed between HDL2-C and plaque thickness (per mg/dl; beta = 0.02 at the 90% percentile, p = 0.003). HDL-C was associated with plaque area (per mg/dl: beta = -0.18 at the 90th percentile, p = 0.01), but only among Hispanics. CONCLUSION In our cohort we observed an inverse association between HDL3-C and plaque area and a positive association between HDL2-C and plaque thickness. HDL-C subfractions may have different contributions to the risk of vascular disease. More studies are needed to fully elucidate HDL-C anti-atherosclerotic functions in order to improve HDL-based treatments in prevention of vascular disease and stroke.

Patterns of leisure-time physical activity using multivariate finite mixture modeling and cardiovascular risk factors in the Northern Manhattan Study.

PURPOSE Physical activity is currently commonly summarized by simple composite scores of total activity, such as total metabolic equivalent score (METS), without further information about the many specific aspects of activities. We sought to identify more comprehensive physical activity patterns, and their association with cardiovascular disease risk factors. METHODS The Northern Manhattan Study is a multiethnic cohort of stroke-free individuals. Questionnaires were used to capture multiple dimensions of leisure-time physical activity. Participants were grouped into METS categories and also into clusters by multivariate mixture modeling of activity frequency, duration, energy expenditure, and number of activity types. Associations between clusters and risk factors were assessed using χ(2) tests. RESULTS Using data available in 3293 participants, we identified six model-based clusters that were differentiated by frequency and diversity of activities, rather than activity duration. High activity clusters had lower prevalence of the risk factors compared with those with lower activity; associations with obesity and hypertension remained significant after adjusting for METS (P = .027, .043). METS and risk factors were not significantly associated after adjusting for the clusters. CONCLUSIONS Data-driven clustering method is a principled, generalizable approach to depict physical activity and form subgroups associated with cardiovascular risk factors independently of METS.