Rebecca A Betensky
Rebecca Betensky
Scroll
Chair of the Department of Biostatistics
Professor of Biostatistics
-
Professional overview
-
Prior to NYU, Dr. Betensky was Professor of Biostatistics at the Harvard T.H. Chan School of Public Health. She was director of the Harvard Catalyst (Clinical and Translational Science Award) Biostatistics Program; director of the Data and Statistics Core for the Massachusetts Alzheimer’s Disease Research Center; and director of the Biostatistics Neurology Core at Massachusetts General Hospital. Previously, she was the Biostatistics Program Leader for the Dana-Farber/Harvard Cancer Center.
Dr. Betensky’s research focuses on methods for the analysis of censored and truncated outcomes and covariates, which frequently arise from the subsampling of cohort studies. She has a long-time interest in clinical trials, and has written on the evaluation of biomarkers and the use and interpretation of p-values. She has collaborated extensively in studies in neurologic diseases, and serves as statistical editor for Annals of Neurology.
Dr. Betensky was awarded, and directed for 15 years, an NIH T32 training program in neurostatistics and neuroepidemiology for pre- and post-doctoral students in biostatistics and epidemiology and for clinician-scientists. She previously directed Harvard’s Biostatistics programs to promote and support diversity at all levels in the field of quantitative public health. She was also a member of the BMRD Study Section for review of NIH statistical methodology grants; on committees for the Institute of Medicine; and a co-chair of the technical advisory committee for the scientific registry of transplant recipients.
Dr. Betensky an elected Fellow of the American Statistical Association and of the International Statistical Institute, and is a past recipient of the Spiegelman Award from the American Public Health Association. She currently serves as a member of the Board of Scientific Counselors for Clinical Science and Epidemiology at the National Cancer Institute.
-
Education
-
AB, Mathematics, Harvard University, Cambridge, MAPhD, Statistics, Stanford University, Stanford, CA
-
Areas of research and study
-
BiologyBiostatisticsNeuroepidemiologyNeurologyNeurostatisticsTranslational science
-
Publications
Publications
Neurolymphomatosis
AbstractBaehring, J. M., Damek, D., Martin, E. C., Betensky, R. A., & Hochberg, F. H. (n.d.).Publication year
2003Journal title
Neuro-OncologyVolume
5Issue
2Page(s)
104-115AbstractThe term "neurolymphomatosis" (NL) has included infiltration of the peripheral nervous system by lymphoma and nontumor lymphocytes. We describe NL as a lymphoma entity that affects cranial and peripheral nerves and roots. We reviewed the medical records of patients at the Massachusetts General Hospital (MGH) who registered between 1972 and 2000, as well as cases published in the English-language literature. Inclusion criteria were (A) histopathologic demonstration of lymphoma within peripheral nerve, nerve root/plexus, or cranial nerve or (B) CT/MRI or intraoperative evidence of nerve enlargement and/or enhancement beyond the dural sleeve in the setting of prior or concurrent lymphoma in systemic or CNS sites. We identified 25 patients with NL in addition to 47 reported by others. Four clinical presentations were (1) painful involvement of nerves or roots, (2) cranial neuropathy with or without pain, (3) painless involvement of peripheral nerves, (4) painful or painless involvement of a single peripheral nerve. Twenty of our patients and 44 of those reported had histopathologic confirmation of lymphoma infiltrating root or nerve. In the remainder, diagnosis was based upon clinical presentation, nodular nerve enlargement or enhancement, and lymphoma cells in spinal fluid or extraneural sites. For antemortem diagnosis, imaging studies were of greatest utility, followed by biopsy. Thirty-three patients of the combined series were not correctly diagnosed until postmortem examination. Systemic chemotherapy was used to address the multiple potential sites of involvement. When properly treated, NL carries a prognosis similar to primary CNS lymphoma in the modern era.The Leapfrog Volume Criteria May Fall Short in Identifying High-Quality Surgical Centers
AbstractChristian, C. K., Gustafson, M. L., Betensky, R. A., Daley, J., Zinner, M. J., Greenfield, L. J., & Rhodes, R. S. (n.d.).Publication year
2003Journal title
Annals of SurgeryVolume
238Issue
4Page(s)
447-457AbstractObjective: The original Leapfrog Initiative recommends selective referral based on procedural volume thresholds (500 coronary artery bypass graft [CABG] surgeries, 30 abdominal aortic aneurysm [AAA] repairs, 100 carotid endarterectomies [CEA], and 7 esophagectomies annually). We tested the volume-mortality relationship for these procedures in the University HealthSystem Consortium (UHC) Clinical DatabaseSM, a database of all payor discharge abstracts from UHC academic medical center members and affiliates. We determined whether the Leapfrog thresholds represent the optimal cutoffs to discriminate between high- and low-mortality hospitals. Methods: Logistic regression was used to test whether volume was a significant predictor of mortality. Volume was analyzed in 3 different ways: as a continuous variable, a dichotomous variable (above and below the Leapfrog threshold), and a categorical variable. We examined all possible thresholds for volume and observed the optimal thresholds at which the odds ratio is the highest, representing the greatest difference in odds of death between the 2 groups of hospitals. Results: In multivariate analysis, a relationship between volume and mortality exists for AAA in all 3 models. For CABG, there is a strong relationship when volume is tested as a dichotomous or categorical variable. For CEA and esophagectomy, we were unable to identify a consistent relationship between volume and outcome. We identified empirical thresholds of 250 CABG, 15 AAA, and 22 esophagectomies, but were unable to find a meaningful threshold for CEA. Conclusions: In this group of academic medical centers and their affiliated hospitals, we demonstrated a significant relationship between volume and mortality for CABG and AAA but not for CEA and esophagectomy, based on the Leapfrog thresholds. We described a new methodology to identify optimal data-based volume thresholds that may serve as a more rational basis for selective referral.A local likelihood proportional hazards model for interval censored data
AbstractBetensky, R. A., Lindsey, J. C., Ryan, L. M., & Wand, M. P. (n.d.).Publication year
2002Journal title
Statistics in MedicineVolume
21Issue
2Page(s)
263-275AbstractWe discuss the use of local likelihood methods to fit proportional hazards regression models to right and interval censored data. The assumed model allows for an arbitrary, smoothed baseline hazard on which a vector of covariates operates in a proportional manner, and thus produces an interpretable baseline hazard function along with estimates of global covariate effects. For estimation, we extend the modified EM algorithm suggested by Betensky, Lindsey, Ryan and Wand. We illustrate the method with data on times to deterioration of breast cosmeses and HIV-1 infection rates among haemophiliacs.C455R notch3 mutation in a Colombian CADASIL kindred with early onset of stroke
AbstractArboleda-Velasquez, J. F., Lopera, F., Lopez, E., Frosch, M. P., Sepulveda-Falla, D., Gutierrez, J. E., Vargas, S., Medina, M., Martinez De Arrieta, C., Lebo, R. V., Slaugenhaupt, S. A., Betensky, R. A., Villegas, A., Arcos-Burgos, M., Rivera, D., Restrepo, J. C., & Kosik, K. S. (n.d.).Publication year
2002Journal title
NeurologyVolume
59Issue
2Page(s)
277-279AbstractCerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is caused by mutations in the notch3 epidermal growth factor-like repeats. A Colombian kindred carries a novel C455R mutation located in the predicted ligand-binding domain. Stroke occurred in the patients at an unusually early age (median age: 31 years) in comparison to the more frequent onset in the fourth decade of life in other CADASIL populations, including a second Colombian kindred with an R1031C mutation.DMBT1 polymorphisms: Relationship to malignant glioma tumorigenesis
AbstractSasaki, H., Betensky, R. A., Cairncross, J. G., & Louis, D. N. (n.d.).Publication year
2002Journal title
Cancer ResearchVolume
62Issue
6Page(s)
1790-1796AbstractThe deleted in malignant brain tumors 1 (DMBT1) gene on 10q25-26 is a candidate tumor suppressor gene in malignant gliomas, but its role is controversial, e.g., some DMBT1 homozygous deletions reflect unmasking of constitutional deletion polymorphisms by 10q loss. To clarify the role of DMBT1 in gliomagenesis, we investigated three reported deletion hot spots. Homozygous deletions at DMBT1 repeat 2-4 to 2-7 were found in 10 of 73 gliomas with 10q loss, but all 10 deletions reflected unmasking of constitutional hemizygous deletions. Alleles bearing deletion 2-4/2-7 were not selected significantly for by 10q loss, with retention of only 10 of 16 deleted alleles. No homozygous deletion was detected at locus 74k in the 5′ upstream region of DMBT1, and four tightly linked polymorphisms were found around this region; chromosome 10q loss randomly affected alleles with or without the variant sequences around locus 74k. Moreover, no significant selection pressure was detected for the haplotype with both deletion 2-4/2-7 and 5′ polymorphisms. There was no segregation of deletion 2-4/2-7 in glioma patients compared with unrelated individuals from reference families but a suggestion of a difference in the distribution of the 5′ polymorphisms between the reference individuals and glioma patients. Constitutional polymorphisms at DMBT1 repeat 2-9/2-10 appeared common in patients with both benign brain tumors and gliomas. A homozygote for both the 2-4/2-7 deletion and the 5′ polymorphisms had a glioma arise at a typical age and without an apparent family cancer predisposition. These data suggest that DMBT1 polymorphisms are not likely primary targets of 10q loss in malignant gliomas and do not support a major role for DMBT1 in gliomagenesis.Histopathological-molecular genetic correlations in referral pathologist-diagnosed low-grade "oligodendroglioma"
AbstractSasaki, H., Zlatescu, M. C., Betensky, R. A., Johnk, L. B., Cutone, A. N., Cairncross, J. G., & Louis, D. N. (n.d.).Publication year
2002Journal title
Journal of Neuropathology and Experimental NeurologyVolume
61Issue
1Page(s)
58-63AbstractAllelic loss of chromosome 1p predicts increased chemosensitivity and better survival in oligodendroglial tumors. Clinical testing for 1p loss in oligodendroglial tumors at our hospital has allowed us to postulate that certain histological appearances are associated with 1p allelic status. Forty-four cases received for genetic testing were diagnosed by referring pathologists as pure low-grade oligodendroglioma. Central neuropathological review divided the series equally into 22 cases with classical oligodendroglioma histology and 22 with more astrocytic features. Molecular genetic analyses demonstrated 1p loss in 19 of 22 classic oligodendrogliomas (86%) and maintenance of both 1p alleles in 16 of 22 gliomas with astrocytic features (73%). No glial fibrillary acidic protein-positive cell type (gliofibrillary oligodendrocyte, minigemistocyte, cellular processes) was associated with 1p allelic status. Fourteen of the 44 cases were treated with chemotherapy at tumor progression: 3 "astrocytic" gliomas with 1p loss responded to PCV chemotherapy and 2 classic oligodendrogliomas that maintained both 1p alleles included a responder and a non-responder. These results suggest that histological appearance correctly predicts genotype in approximately 80% of low-grade gliomas, but that tumor genotype more closely predicts chemosensitivity. As a result, such objective molecular genetic analyses should be incorporated into patient management and into clinical trials of low-grade diffuse gliomas.Influence of unrecognized molecular heterogeneity on randomized clinical trials
AbstractBetensky, R. A., Louis, D. N., & Cairncross, J. G. (n.d.).Publication year
2002Journal title
Journal of Clinical OncologyVolume
20Issue
10Page(s)
2495-2499AbstractPurpose: In solid tumor oncology, decisions regarding treatment and eligibility for trials are governed by histologic diagnosis. Despite this reliance on histology and the assumption that histology defines the disease, underlying molecular heterogeneity likely differentiates among patients' outcomes. Patients and Methods: To illustrate how unrecognized molecular heterogeneity might obscure a truly effective new therapy for cancer, we analyzed the planning assumptions and results of a hypothetical randomized controlled trial of chemoradiotherapy for a cancer found to be drug sensitive in preliminary phase II studies. Results: Randomized controlled trials of effective cancer therapies can be falsely negative if therapeutic benefit is overestimated during study design because of enrichment of phase II trials for treatment-sensitive subtypes, a beneficial effect in responding patients is diluted by large numbers of nonresponding patients, or a beneficial effect in responders is reversed by a negative effect in nonresponders. Conclusion: Molecular heterogeneity, if it confers different risks to patients and is unaccounted for in the design of a randomized study, can result in a clinical trial that is underpowered and fails to detect a truly effective new therapy for cancerLocal likelihood analysis of the latency distribution with interval censored intermediate events
AbstractBebchuk, J. D., & Betensky, R. A. (n.d.).Publication year
2002Journal title
Statistics in MedicineVolume
21Issue
22Page(s)
3475-3491AbstractThe distribution of time between an intermediate event and a terminal event is frequently of interest both in studying the behaviour of populations and in predicting outcomes for individuals. Current methods for estimating this latency distribution have either imposed assumptions on the hazard for the terminal event following the occurrence of the intermediate event or on the parametric form of the hazards. Here, local likelihood estimation is applied to the underlying hazard functions of a three-state process in which the time of the intermediate event may be interval censored and the time of the terminal event is either observed or right censored. Smooth non-parametric estimates of the latency distribution, along with bootstrap confidence intervals, are calculated, and tests for the comparison of two latency distributions are presented. The method is applied to two studies: a cohort of haemophiliacs who were infected with HIV by contaminated blood factor and followed for AIDS onset, and an AIDS clinical trial in which the intermediate event is the time to 50 per cent drop in CD4 count and the terminal event is AIDS or death. Simulations are presented to assess the performance of the estimation procedure.Sample size re-estimation in cluster randomization trials
AbstractLake, S., Kammann, E., Klar, N., & Betensky, R. (n.d.).Publication year
2002Journal title
Statistics in MedicineVolume
21Issue
10Page(s)
1337-1350AbstractCluster randomization trials in which families are the unit of allocation are commonly adopted for the evaluation of disease prevention interventions. Sample size estimation for cluster randomization trials depends on parameters that quantify the variability within and between clusters and the variability in cluster size. Accurate advance estimates of these nuisance parameters may be difficult to obtain and misspecification may lead to an underpowered study. Since families are typically recruited over time, we propose using a portion of the data to estimate the nuisance parameters and to re-estimate sample size based on the estimates. This extends the standard internal pilot study methods to the setting of cluster randomization trials. The effect of this design on the power, significance level and sample size is analysed via simulation and is shown to provide a flexible and practical approach to cluster randomization trials.Testing for dependence between failure time and visit compliance with interval-censored data
AbstractBetensky, R. A., & Finkelstein, D. M. (n.d.).Publication year
2002Journal title
BiometricsVolume
58Issue
1Page(s)
58-63AbstractInterval-censored failure-time data arise when subjects miss prescheduled visits at which the failure is to be assessed. The resulting intervals in which the failure is known to have occurred are overlapping. Most approaches to the analysis of these data assume that the visit-compliance process is ignorable with respect to likelihood analysis of the failure-time distribution. While this assumption offers considerable simplification, it is not always plausible. Here we test for dependence between the failure- and visit-compliance processes, applicable to studies in which data collection continues after the occurrence of the failure. We do not make any of the assumptions made by previous authors about the joint distribution of the visit-compliance process, a covariate process, and the failure time. Instead, we consider conditional models of the true failure history given the current visit compliance at each visit time, allowing for correlation across visit times. Because failure status is not known at some visit times due to missed visits, only models of the observed failure history given current visit compliance are estimable. We describe how the parameters from these models can be used to test for a negative association and how bounds on unestimable parameters provided by the observed data are needed additionally to infer a positive association. We illustrate the method with data from an AIDS study and we investigate the power of the test through a simulation study.The use of frailty hazard models for unrecognized heterogeneity that interacts with treatment: Considerations of efficiency and power
AbstractLi, Y., Betensky, R. A., Louis, D. N., & Cairncross, J. G. (n.d.).Publication year
2002Journal title
BiometricsVolume
58Issue
1Page(s)
232-236AbstractIncreasingly, genetic studies of tumors of the same histologic diagnosis are elucidating subtypes that are distinct with respect to clinical endpoints such as response to treatment and survival. This raises concerns about the efficiency of using the simple log-rank test for analysis of treatment effect on survival in studies of possibly heterogeneous tumors. Furthermore, such studies, designed under the assumption of homogeneity, may be severely underpowered. We derive analytic approximations for the asymptotic relative efficiency of the simple log-rank test relative to the optimally weighted log-rank test and for the power of the simple log-rank test when applied to subjects with unobserved heterogeneity, as reflected in a continuous frailty, that may interact with treatment. Numerical studies demonstrate that the simple log-rank test may be quite inefficient if the frailty interacts with treatment. Further, there may be a substantial loss of power in the presence of the frailty with or without an interaction with treatment.A comparison of models for clustered binary outcomes: Analysis of a designed immunology experiment
AbstractBetensky, R. A., Williams, P. L., & Lederman, H. M. (n.d.).Publication year
2001Journal title
Journal of the Royal Statistical Society. Series C: Applied StatisticsVolume
50Issue
1Page(s)
43-61AbstractThe lymphocyte proliferative assay (LPA) of immune competence was conducted on 52 subjects, with up to 36 processing conditions per subject, to evaluate whether samples could be shipped or stored overnight, rather than being processed on fresh blood as currently required. The LPA study resulted in clustered binary data, with both cluster level and cluster-varying covariates. Two modelling strategies for the analysis of such clustered binary data are through the cluster-specific and population-averaged approaches. Whereas most research in this area has focused on the analysis of matched pairs data, in many situations, such as the LPA study, cluster sizes are naturally larger. Through considerations of interpretation and efficiency of these models when applied to large clusters, the mixed effect cluster-specific model was selected as most appropriate for the analysis of the LPA data. The model confirmed that the LPA response is significantly impaired in individuals infected with the human immunodeficiency virus (HIV). The LPA response was found to be significantly lower for shipped and overnight samples than for fresh samples, and this effect was significantly stronger among HIV-infected individuals. Surprisingly, an anticoagulant effect was not detected.A computationally simple test of homogeneity of odds ratios for twin data
AbstractBetensky, R. A., Hudson, J. I., Jones, C. A., Hu, F., Wang, B., Chen, C., & Xu, X. (n.d.).Publication year
2001Journal title
Genetic EpidemiologyVolume
20Issue
2Page(s)
228-238AbstractIt is of interest to compare measures of association of binary traits among samples of bivariate data. One example is the comparison of association within a sample of monozygotic (MZ) twins to that within a sample of dizygotic (DZ) twins. A larger association in the MZ twins suggests that the trait of interest may have a genetic component. The Bivariate data in this example are binary traits for the twins in each pair. Another example is the comparison of a measure of Hardy-Weinberg disequilibrium across several populations. The bivariate data in this example are the two alleles comprising the genotype of interest. We propose using derived logistic regression equations from the full exponential model for the bivariate outcomes to test for homogeneity. We adjust for correlation among outcomes via generalized estimating equations. This modeling approach allows for adjustment for individual-level and pair-level covariates and thereby allows for testing for gene x environment interactions. Further, we extend the model to allow for simultaneous analysis of two diseases, which allows for testing for a genetic component to the coaggregation of two diseases. In contrast to approaches proposed by previous authors, no special software is required; our approach can be easily implemented in standard software packages. We compare our results to those of other methods proposed in the literature for data from the Vietnam Era Twin Study. We apply our methods also to the Anqing Twin Study and data on major depression and generalized anxiety disorder from the Virginia Twin Register.Computationally simple accelerated failure time regression for interval censored data
AbstractBetensky, R. A., Rabinowitz, D., & Tsiatis, A. A. (n.d.).Publication year
2001Journal title
BiometrikaVolume
88Issue
3Page(s)
703-711AbstractAn approach is presented for fitting the accelerated failure time model to interval censored data that does not involve computing the nonparametric maximum likelihood estimate of the distribution function at the residuals. The approach involves estimating equations computed with the examination times from the same individual treated as if they had actually been obtained from different individuals. The dependence between different measurements obtained from the same individual is then accounted for in the calculation of the standard error of the regression coefficients. The approach is applicable to interval censored data in settings in which examinations continue to occur regardless of whether the failure time has occurred. Simulations are presented to assess the behaviour of the approach, and the methodology is illustrated through an application to data from an AIDS clinical trial.Local likelihood analysis of survival data with censored intermediate events
AbstractBebchuk, J. D., & Betensky, R. A. (n.d.).Publication year
2001Journal title
Journal of the American Statistical AssociationVolume
96Issue
454Page(s)
449-457AbstractAIDS Clinical Trials Group protocol 193A was a randomized trial designed to compare survival and progression-free survival among patients on different treatment regimens. A complicating feature of the analysis of progression-free survival is that different censoring mechanisms operated on progression and survival, which resulted in more complete information on survival. A simple analysis that uses the minimum of the times to progression and survival and the minimum of the corresponding censoring times may sacrifice the extra information available on survival. To address this problem, we have developed a method that exploits the bivariate nature of these data and thereby uses all of the available information. We obtain smooth, nonparametric estimates of the hazard functions for a terminal event, before and after the occurrence of an intermediate event. These hazards can be used to estimate the distribution of progression-free survival. Our method uses local likelihood estimation, which assumes that the underlying true hazard functions can be approximated locally by polynomials. We use an iterative imputation algorithm to perform the estimation when the intermediate events are right censored.Molecular subtypes of anaplastic oligodendroglioma: Implications for patient management at diagnosis
AbstractIno, Y., Betensky, R. A., Zlatescu, M. C., Sasaki, H., Macdonald, D. R., Stemmer-Rachamimov, A. O., Ramsay, D. A., Cairncross, J. G., & Louis, D. N. (n.d.).Publication year
2001Journal title
Clinical Cancer ResearchVolume
7Issue
4Page(s)
839-845AbstractPurpose: In a prior study of anaplastic oligodendrogliomas treated with chemotherapy at diagnosis or at recurrence after radiotherapy, allelic loss of chromosome 1p correlated with better chemotherapeutic response and overall survival. However, in this group of patients in whom therapeutic management was not uniform, loss of 1p did not identify all chemosensitive tumors, nor did all patients whose tumors harbor a 1p loss have long survival. Experimental Design: To clarify the clinical relevance of molecular genetic testing at the time of diagnosis for patients with anaplastic oligodendrogliomas, we studied a larger, more homogeneous group of 50 patients with histologically defined anaplastic oligodendrogliomas treated with a chemotherapeutic regimen as the principal initial therapy. Results: We demonstrate that these tumors can be divided genetically into four therapeutically and prognostically relevant subgroups. Patients whose tumors have combined but isolated losses of 1p and 19q have marked and durable responses to chemotherapy associated with long survival, with or without postoperative radiation therapy. Other tumors with chromosome 1p alterations also respond to chemotherapy, but with shorter duration of response and patient survival. Tumors lacking 1p loss can also be divided into two subgroups: those with TP53 mutations, which may also respond to chemotherapy but recur quickly, and those without TP53 mutations, which are poorly responsive, aggressive tumors that are clinically and genotypically similar to glioblastomas. Conclusions: These data raise the possibility, for the first time, that therapeutic decisions at the time of diagnosis might be tailored to particular genetic subtypes of anaplastic oligodendroglioma.Multivariate logistic regression for familial aggregation of two disorders. I. Development of models and methods
AbstractHudson, J. I., Laird, N. M., & Betensky, R. A. (n.d.).Publication year
2001Journal title
American Journal of EpidemiologyVolume
153Issue
5Page(s)
500-505AbstractThe question of whether two disorders cluster together, or coaggregate, within families often arises. This paper considers how to analyze familial aggregation of two disorders and presents two multivariate logistic regression methods that model both disorder outcomes simultaneously. The first, a proband predictive model, predicts a relative's outcomes (the presence or absence of each of the two disorders) by using the proband's disorder status. The second, a family predictive model derived from the quadratic exponential model, predicts a family member's outcomes by using all of the remaining family members' disorder statuses. The models are more realistic, flexible, and powerful than univariate models. Methods for estimation and testing account for the correlation of outcomes among family members and can be implemented by using commercial software.Multivariate logistic regression for familial aggregation of two disorders. II. Analysis of studies of eating and mood disorders
AbstractHudson, J. I., Laird, N. M., Betensky, R. A., Keck, P. E., & Pope, H. G. (n.d.).Publication year
2001Journal title
American Journal of EpidemiologyVolume
153Issue
5Page(s)
506-514AbstractFamily studies have suggested that eating disorders and mood disorders may coaggregate within families. Previous studies, however, have been limited by use of univariate modeling techniques and failure to account for the correlation of observations within families. To provide a more efficient analysis and to illustrate multivariate logistic regression models for familial aggregation of two disorders, the authors analyzed pooled data from two previously published family studies (conducted in Massachusetts in 1984-1986 and 1986-1987) by using multivariate proband predictive and family predictive models. Both models demonstrated a significant familial aggregation of mood disorders and familial coaggregation of eating and mood disorders. The magnitude of the coaggregation between eating and mood disorders was similar to that of the aggregation of mood disorders. Similar results were obtained with alternative models, including a traditional univariate proband predictive model. In comparison with the univariate model, the multivariate models provided greater flexibility, improved precision, and wider generality for interpreting aggregation effects.Nonparametric estimation in a cure model with random cure times
AbstractBetensky, R. A., & Schoenfeld, D. A. (n.d.).Publication year
2001Journal title
BiometricsVolume
57Issue
1Page(s)
282-286AbstractAcute respiratory distress syndrome (ARDS) is a life-threatening acute condition that sometimes follows pneumonia or surgery. Patients who recover and leave the hospital are considered to have been cured at the time they leave the hospital. These data differ from typical data in which cure is a possibility: death times are not observed for patients who are cured and cure times are observed and vary among patients. Here we apply a competing risks model to these data and show it to be equivalent to a mixture model, the more common approach for cure data. Further, we derive an estimator for the variance of the cumulative incidence function from the competing risks model, and thus for the cure rate, based on elementary calculations. We compare our variance estimator to Gray's (1988, Annals of Statistics 16, 1140-1154) estimator, which is based on counting process theory. We find our estimator to be slightly more accurate in small samples. We apply these results to data from an ARDS clinical trial.Optimally selected chi square statistics for equivalence testing
AbstractBetensky, R. A. (n.d.).Publication year
2001Journal title
Journal of Statistical Planning and InferenceVolume
93Issue
1Page(s)
247-257AbstractIn some experiments, the aim is to establish equivalence between treatment groups, rather than to establish greater efficacy of one group over another. Examples are drug bioavailability studies, in which it is desired to exhibit bioequivalence between a new formulation of a drug and the standard formulation, and studies of experimental conditions for biological assays, in which it is desired to exhibit equivalent assay response between fresh and non-fresh specimens. In many cases, the outcome measure is continuous, but the investigator prefers to dichotomize it on the basis of a threshold. If the threshold is selected to make the test statistic most favorable for equivalence, standard chi square percentile points are incorrect. Here, we derive asymptotically correct percentile points and significance levels for two approaches to equivalence testing.PTEN is a target of chromosome 10q loss in anaplastic oligodendrogliomas and PTEN alterations are associated with poor prognosis
AbstractSasaki, H., Zlatescu, M. C., Betensky, R. A., Ino, Y., Cairncross, J. G., & Louis, D. N. (n.d.).Publication year
2001Journal title
American Journal of PathologyVolume
159Issue
1Page(s)
359-367AbstractAllelic loss of 10q is a common genetic event in malignant gliomas, with three 10q tumor suppressor genes, ERCC6, PTEN, and DMBT1, putatively implicated in the most common type of malignant glioma, glioblastoma. Anaplastic oligodendroglioma, another type of malignant glioma, provides a unique opportunity to study the relevance of particular genetic alterations to chemosensitivity and survival. We therefore analyzed these three genes in 72 anaplastic oligodendrogliomas. Deletion mapping demonstrated 10q loss in 14 of 67 informative cases, with the PTEN and DMBT1 regions involved in all deletions but with the ERCC6 locus spared in two cases. Seven tumors had PTEN gene alterations; two had homozygous DMBT1 deletions, but at least one reflected unmasking of a germline DMBT1 deletion. No mutations were found in ERCC6 exon 2. Chemotherapeutic response occurred in two of the seven tumors with PTEN alterations, but with unexpected short survival times. PTEN gene alterations were not associated with poor therapeutic response in multivariate analysis, but were independently predictive of poor prognosis even after multivariate adjustment for both 10q and 1p loss. In anaplastic oligodendroglioma, therefore, PTEN is a target of 10q loss, and PTEN alterations are associated with poor prognosis, even in chemosensitive cases.Tumor location and growth pattern correlate with genetic signature in oligodendroglial neoplasms
AbstractZlatescu, M. C., TehraniYazdi, A. R., Sasaki, H., Megyesi, J. F., Betensky, R. A., Louis, D. N., & Cairncross, J. G. (n.d.).Publication year
2001Journal title
Cancer ResearchVolume
61Issue
18Page(s)
6713-6715AbstractMolecular genetic subsets of anaplastic oligodendroglioma behave in biologically distinct ways, in both their rates of growth and their responses to standard therapies. In a series of 64 cases, we evaluated whether allelic loss of chromosomal arms 1p and 19q, an early molecular event in the genesis of chemosensitive oligodendrogliomas, is related to tumor location and extent of tumor spread in the brain. We observed that tumor genotype was closely associated with tumor location (P < 0.001). Anaplastic oligodendrogliomas located in the frontal, parietal, and occipital lobes were significantly more likely to harbor allelic loss of chromosomal arms 1p and 19q than histologically indistinguishable tumors arising in the temporal lobe, insula, and diencephalon (P < 0.001). In addition, loss of heterozygosity for 1p and 19q was significantly associated with a bilateral pattern of growth (P = 0.037); all seven bilaterally distributed anaplastic oligodendrogliomas had 1p and 19q allelic loss. We conclude, therefore, that molecular subtypes of oligodendrogliomas may arise preferentially in certain lobes of the brain and have differential patterns of growth, with tumors having allelic loss of chromosomes 1p and 19q occurring most frequently in the frontal lobes and having a tendency for widespread growth across the midline. These findings encourage inquiries into the biological basis of such marked differences and already have implications for the current management of these neoplasms.α-synuclein occurs in lipid-rich high molecular weight complexes, binds fatty acids, and shows homology to the fatty acid-binding proteins
AbstractSharon, R., Goldberg, M. S., Bar-Josef, I., Betensky, R. A., Shen, J., & Selkoe, D. J. (n.d.).Publication year
2001Journal title
Proceedings of the National Academy of Sciences of the United States of AmericaVolume
98Issue
16Page(s)
9110-9115Abstractα-Synuclein (αS) is a 140-residue neuronal protein that forms insoluble cytoplasmic aggregates in Parkinson's disease (PD) and several other neurodegenerative disorders. Two missense mutations (A53T and A30P) are linked to rare forms of familial PD. The normal function of αS is unknown, and cultured cell systems that model its modification from soluble monomers to aggregated forms have not been reported. Through a systematic centrifugal fractionation of mesencephalic neuronal cell lines and transgenic mouse brains expressing wild-type or A53T human αS, we observed unusual, previously unrecognized species of αS that migrate well above the 17-kDa monomeric form in denaturing gels. Incubation at 65°C of high-speed cytosols from cells or brains revealed a modified αS species migrating at ≈36 kDa and an extensive higher molecular mass αS-reactive smear. Extraction of the cytosols with chloroform/methanol or with a resin (Lipidex 1000) that binds fatty acids resulted in a similar pattern of higher molecular mass αS forms. On the basis of this effect of delipidation, we reexamined the primary structure of αS and detected a motif at the N and C termini that is homologous to a fatty acid-binding protein signature. In accord, we found that purified human αS binds oleic acid, with an apparent Kd of 12.5 μM. We also observed an enhanced association of A53T αS with microsomal membranes in both mesencephalic cells and transgenic mouse brains. We conclude that αS has biochemical properties and a structural motif that suggest it is a novel member of the fatty acid-binding protein family and may thus transport fatty acids between the aqueous and membrane phospholipid compartments of the neuronal cytoplasm.Alternative derivations of a rule for early stopping in favor of H0
AbstractBetensky, R. A. (n.d.).Publication year
2000Journal title
American StatisticianVolume
54Issue
1Page(s)
35-39AbstractIt is often desirable to stop a large clinical trial before its planned end if a null result seems inevitable. This early stopping can save considerable resources. It is especially appealing when an experimental treatment is being compared to a standard treatment. Three procedures for early stopping, all with different interpretations and derivations, are described and shown to produce identical rules for normal data and certain parameters. In some cases, this is unexpected and informative. The procedures differ in which of their parameters are adjusted from the fixed sample values to maintain the desired Type I error in this setting of multiple looks at the data.Approximating the distribution of maximally selected McNemar's statistics
AbstractRabinowitz, D., & Betensky, R. A. (n.d.).Publication year
2000Journal title
BiometricsVolume
56Issue
3Page(s)
897-902AbstractIt is common in epidemiologic analyses to summarize continuous outcomes as falling above or below a threshold. With paired data and with a threshold chosen without reference to the outcomes, McNemar's test of marginal homogeneity may be applied to the resulting dichotomous pairs when testing for equality of the marginal distributions of the underlying continuous outcomes. If the threshold is chosen to maximize the test statistic, however, referring the resulting test statistic to the nominal χ2 distribution is incorrect; Instead, the p-value must be adjusted for the multiple comparisons. Here the distribution of a maximally selected McNemar's statistic is derived, and it is shown that an approximation due to Durbin (1985, Journal of Applied Probability 22, 99-122) may be used to estimate approximate p-values. The methodology is illustrated by an application to measurements of insulin-like growth factor-I (IGF-I)in matched prostate cancer cases and controls from the Physicians' Health Study. The results of Simulation experiments that assess the accuracy of the approximation in moderate sample sizes are reported.