Biomedical subjects

Sin-Ho Jung

Publications and source records attributed to Sin-Ho Jung.

14 recordsLinked to original sources

PubMed2004

On the estimation of the binomial probability in multistage clinical trials.

Due to the optional sampling effect in a sequential design, the maximum likelihood estimator (MLE) following sequential tests is generally biased. In a typical two-stage design employed in a phase II clinical trial in cancer drug screening, a fixed number of patients are enrolled initially. The trial may be terminated for lack of clinical efficacy of treatment if the observed number of treatment responses after the first stage is too small. Otherwise, an additional fixed number of patients are enrolled to accumulate additional information on efficacy as well as on safety. There have been numerous suggestions for design of such two-stage studies. Here we establish that under the two-stage design the sufficient statistic, i.e. stopping stage and the number of treatment responses, for the parameter of the binomial distribution is also complete. Then, based on the Rao-Blackwell theorem, we derive the uniformly minimum variance unbiased estimator (UMVUE) as the conditional expectation of an unbiased estimator, which in this case is simply the maximum likelihood estimator based only on the first stage data, given the complete sufficient statistic. Our results generalize to a multistage design. We will illustrate features of the UMVUE based on two-stage phase II clinical trial design examples and present results of numerical studies on the properties of the UMVUE in comparison to the usual MLE.

Clinical Trials, Phase II as Topic↗

PubMed2004

Admissible two-stage designs for phase II cancer clinical trials.

In a typical two-stage design for a phase II cancer clinical trial for efficacy screening of cytotoxic agents, a fixed number of patients are initially enrolled and treated. The trial may be terminated for lack of efficacy if the observed number of tumour responses after the first stage is too small, thus avoiding treatment of patient with inefficacious regimen. Otherwise, an additional fixed number of patients are enrolled and treated to accumulate additional information on efficacy as well as safety. The minimax and the so-called 'optimal' designs by Simon have been widely used, and other designs have largely been ignored in the past for such two-stage cancer clinical trials. Recently Jung et al. proposed a graphical method to search for compromise designs with features more favourable than either the minimax or the optimal design. In this paper, we develop a family of two-stage designs that are admissible according to a Bayesian decision-theoretic criterion based on an ethically justifiable loss function. We show that the admissible designs include as special cases the Simon's minimax and the optimal designs as well as the compromise designs introduced by Jung et al. We also present a Java program to search for admissible designs that are compromises between the minimax and the optimal designs.

Bayes Theorem↗

PubMed2004

Characterization of gains, losses, and regional amplification in testicular germ cell tumor cell lines by comparative genomic hybridization.

We have performed comparative genomic hybridization on 12 testicular germ cell tumor (TGCT) cell lines and one paraffin-embedded surgical specimen to identify and characterize genome-wide gains and losses of chromosomes in these specimens. All specimens demonstrated overrepresentation of 12p. Other significant chromosomal gains, apart from 12p, included the X chromosome and chromosome arms 1q and 20q. Chromosomal losses were observed for chromosomes 4 and 18 and chromosome arms 2q, 9q, and 13q. Genomic differences were observed between an embryonal carcinoma component of a mixed tumor, 833K, and its cisplastin-resistant derivative line, 64CP, including losses of 6q23 approximately qter and 9p22 approximately q21. Five lines also demonstrated gain of 12p and additional 12p12 approximately p13 material. Similarly, two lines demonstrated gain of 12p and additional 12p11.2 approximately p12 material. The data supports the consistent gain of 12p in adult TGCT cell lines and additional regional amplification of 12p in some lines. This regional amplification has been observed in both primary tumor specimens and TGCT cell lines and may support a hypothesis that at least two different regions of 12p, one proximal and one distal, harbor genes important for the pathogenesis of testicular germ cell neoplasia.

Cell Line, Tumor↗

PubMed2003

Assessment of quality of life in outpatients with advanced cancer: the accuracy of clinician estimations and the relevance of spiritual well-being--a Hoosier Oncology Group Study.

PURPOSE: To evaluate the association between quality-of-life (QOL) impairment as reported by patients and QOL impairment as judged by nurses or physicians, with and without consideration of spiritual well-being (SWB). PATIENTS AND METHODS: A total of 163 patients with advanced cancer were enrolled onto a therapeutic trial, and cross-sectional data were derived from clinical and demographic questionnaires obtained at baseline, including assessment of patient QOL and SWB. Clinicians rated the QOL impairment of their patients as mild, moderate, or severe. Clinician-estimated QOL impairment and patient-derived QOL categories were compared. Correlation coefficients were estimated to associate QOL scores using different instruments. The analysis of variance method was used to compare Functional Assessment of Cancer Therapy-General scores on categorical variables. RESULTS: There was no significant association between self-assessment scores and marital status, education level, performance status, or predicted life expectancy. However, a strong relationship between SWB and QOL was noted (P <.0001). Clinician-estimated QOL impairment matched the level of patient-derived QOL correctly in approximately 60% of cases, with only slight variation depending on the method of categorizing patient-derived QOL scores. The accuracy of clinician estimates was not associated with the level of SWB. Interestingly, a subset analysis of the inaccurate estimates revealed an association between lower SWB and clinician underestimation of QOL impairment (P =.0025). CONCLUSION: Clinician estimates of QOL impairment were accurate in more than 60% of patients. SWB is strongly associated with QOL, but it is not associated with the overall accuracy of clinicians' judgments about QOL impairment.

Activities of Daily Living↗

PubMed2003

Fluoxetine versus placebo in advanced cancer outpatients: a double-blinded trial of the Hoosier Oncology Group.

PURPOSE: To determine whether fluoxetine improves overall quality of life (QOL) in advanced cancer patients with symptoms of depression revealed by a simple survey. PATIENTS AND METHODS: One hundred sixty-three patients with an advanced solid tumor and expected survival between 3 and 24 months were randomly assigned in a double-blinded fashion to receive either fluoxetine (20 mg daily) or placebo for 12 weeks. Patients were screened for at least minimal depressive symptoms and assessed every 3 to 6 weeks for QOL and depression. Patients with recent exposure to antidepressants were excluded. RESULTS: The groups were comparable at baseline in terms of age, sex, disease distribution, performance status, and level of depressive symptoms. One hundred twenty-nine patients (79%) completed at least one follow-up assessment. Analysis using generalized estimating equation modeling revealed that patients treated with fluoxetine exhibited a significant improvement in QOL as shown by the Functional Assessment of Cancer Therapy-General, compared with patients given placebo (P =.01). Specifically, the level of depressive symptoms expressed was lower in patients treated with fluoxetine (P =.0005), and the subgroup of patients showing higher levels of depressive symptoms on the two-question screening survey were the most likely to benefit from treatment. CONCLUSION: In this mix of patients with advanced cancer who had symptoms of depression as determined by a two-question bedside survey, use of fluoxetine was well tolerated, overall QOL was improved, and depressive symptoms were reduced.

Ambulatory Care↗

PubMed2003

Sample size estimation for GEE method for comparing slopes in repeated measurements data.

Sample size calculation is an important component at the design stage of clinical trials. Controlled clinical trials often use a repeated measurement design in which individuals are randomly assigned to treatment groups and followed-up for measurements at intervals across a treatment period of fixed duration. In studies with repeated measurements, one of the popular primary interests is the comparison of the rates of change in a response variable between groups. Statistical models for calculating sample sizes for repeated measurement designs often fail to take into account the impact of missing data correctly. In this paper we propose to use the generalized estimating equation (GEE) method in comparing the rates of change in repeated measurements and introduce closed form formulae for sample size and power that can be calculated using a scientific calculator. Since the sample size formula is based on asymptotic theory, we investigate the performance of the estimated sample size in practical settings through simulations.

Data Interpretation, Statistical↗

PubMed2003

A parametric model for long-term follow-up data from phase III breast cancer clinical trials.

We propose a parametric version of a univariate gamma frailty model. The proposed model is shown to be flexible enough to model long-term follow-up survival data from breast cancer clinical trials when the treatment effect diminishes as time progresses, a case for which neither the proportional hazards nor proportional odds assumptions are satisfied. The observed information matrix is computed to evaluate the variances of parameter estimates. A simple parametric test statistic to test proportional odds assumption is also constructed. The model is applied to a data set from a phase III clinical trial on breast cancer.

Antineoplastic Agents, Hormonal↗

PubMed2003

Update on late relapse of germ cell tumor: a clinical and molecular analysis.

PURPOSE: Analysis of patients with late relapse (LR) of germ cell tumor (GCT) with reports on clinical characteristics, outcomes, and molecular and cytogenetic features. PATIENTS AND METHODS: Eighty-three patients evaluated at Indiana University from 1993 through 2000 for relapse of GCT more than 2 years from initial therapy were reviewed. Available specimens were investigated for expression of the transcription regulator FoxD3 and apurinic/apyrimidinic endonuclease and the presence of chromosome 12 abnormalities. RESULTS: Median interval from initial presentation to LR was 85 months. Forty-three of 49 LR patients who underwent surgery were rendered disease free (NED), and 20 (46.5%) remain continuously NED. Thirty-two patients received chemotherapy, but only six (18.8%) obtained a complete remission. Five of these patients remain continuously NED after chemotherapy alone, including three who were chemotherapy naïve. Eighteen of these 32 patients were successfully rendered NED by postchemotherapy surgery, and 12 remain continuously NED. Two patients continue on observation with no treatment for their LR. Overall, 69 of the 81 treated patients (85.2%) ultimately achieved an NED state, and 38 (46.9%) remain continuously NED with median follow-up from LR therapy of 24.5 months (range, 1 to 83 months), whereas nine other patients are currently NED after therapy for subsequent relapses. Because of the small numbers of specimens tested, we were unable to draw any definitive conclusions from the molecular and cytogenetic analyses. CONCLUSION: GCT patients require lifetime follow-up. At the time of LR, surgical resection alone remains our preferred therapy.

Adult↗

PubMed2003

Results of the American College of Surgeons Oncology Group Z0050 trial: the utility of positron emission tomography in staging potentially operable non-small cell lung cancer.

OBJECTIVES: The American College of Surgeons Oncology Group undertook a trial to ascertain whether positron emission tomography with 18F-fluorodeoxyglucose could detect lesions that would preclude pulmonary resection in a group of patients with documented or suspected non-small cell lung cancer found to be surgical candidates by routine staging procedures. METHODS: A total of 303 eligible patients registered from 22 institutions underwent positron emission tomography after routine staging (computed tomography of chest and upper abdomen, bone scintigraphy, and brain imaging) had deemed their tumors resectable. Positive findings required confirmatory procedures. RESULTS: Positron emission tomography was significantly better than computed tomography for the detection of N1 and N2/N3 disease (42% vs 13%, P =.0177, and 58% vs 32%, P =.0041, respectively). The negative predictive value of positron emission tomography for mediastinal node disease was 87%. Unsuspected metastatic disease or second primary malignancy was identified in 18 of 287 patients (6.3%). Distant metastatic disease indicated in 19 of 287 patients (6.6%) was subsequently shown to be benign. By correctly identifying advanced disease (stages IIIA, IIIB, and IV) or benign lesions, positron emission tomography potentially avoided unnecessary thoracotomy in 1 of 5 patients. CONCLUSIONS: In patients with suspected or proven non-small cell lung cancer considered resectable by standard staging procedures, positron emission tomography can prevent nontherapeutic thoracotomy in a significant number of cases. Use of positron emission tomography for mediastinal staging should not be relied on as a sole staging modality, and positive findings should be confirmed by mediastinoscopy. Metastatic disease, especially a single site, identified by positron emission tomography requires further confirmatory evaluation.

Adult↗

PubMed2003

Rank tests for clustered survival data.

In a clinical trial, we may randomize subjects (called clusters) to different treatments (called groups), and make observations from multiple sites (called units) of each subject. In this case, the observations within each subject could be dependent, whereas those from different subjects are independent. If the outcome of interest is the time to an event, we may use the standard rank tests proposed for independent survival data, such as the logrank and Wilcoxon tests, to test the equality of marginal survival distributions, but their standard error should be modified to accommodate the possible intracluster correlation. In this paper we propose a method of calculating the standard error of the rank tests for two-sample clustered survival data. The method is naturally extended to that for K-sample tests under dependence.

Cluster Analysis↗

PubMed2003

Chi-square test for R x C contingency tables with clustered data.

In ophthalmologic or dental studies, observations are frequently taken from multiple sites (called units), such as eyes or teeth, of each subject. In this case, observations within each subject (called clusters) may be dependent, although those from different subjects are independent. When a categorical observation is made from each site, application of the usual Pearson chi-square tests is invalid since sites within the same subject tend to be dependent. We propose a modified chi2 statistic for testing no treatment effect in these cases. The proposed methods do not require correct specification of the dependence structure within cluster. Simulation studies are conducted to show the finite-sample performance of the new methods. The proposed methods are applied to real-life data.

Animals↗

PubMed2003

Patterns of initial recurrence and prognosis after sentinel lymph node biopsy and selective lymphadenectomy for melanoma.

The histologic status of the sentinel lymph node is a highly significant prognostic factor for patients with clinically localized cutaneous melanoma. The patterns of initial treatment failure of patients with positive sentinel lymph node biopsy versus those with negative results have not been well described. The purpose of this study was to determine the relative prognostic importance of sentinel lymph node status and to compare patterns of initial treatment failure and prognosis of node-positive versus node-negative cutaneous melanoma patients staged by sentinel lymph node biopsy and selective lymphadenectomy. The authors reviewed the pertinent demographic and surgical data in a consecutive series of patients with cutaneous melanoma who underwent sentinel lymph node staging of nonpalpable regional nodes. Sentinel lymph node biopsy was performed using a combination of blue dye and radiolocalization. Patients with positive biopsy results underwent selective lymphadenectomy, whereas those with negative results were observed. Site(s) and date(s) of initial recurrence and death were determined, and disease-free and overall survival probabilities were compared between positive and negative groups using the log-rank test and multivariable Cox regression analysis. Between February of 1994 and August of 2000, 408 patients with melanoma underwent sentinel lymph node biopsy to stage 518 regional lymph node basins. Mean Breslow tumor thickness was 2.27 mm (range, 0.2 to 14.0 mm). Eighty-five patients (20.8 percent) had at least one histologically positive sentinel lymph node, and selective lymphadenectomy yielded additional positive lymph nodes in 18 of 84 patients (21.4 percent). Recurrences were noted in 70 patients (17 percent) at a median follow-up period of 31.4 months. Recurrences were more frequent in patients with positive biopsy results (36.5 percent) than in those with negative results (12.1 percent, p < 0.0001). Distant sites of initial recurrence were more likely in the positive group than in the negative group (71 percent versus 49 percent of recurrences, respectively; p = 0.06). The false-negative rate for sentinel lymph node staging was 4.5 percent and overall accuracy was 99 percent compared with clinical follow-up. Disease-free and overall survival correlated significantly with tumor thickness, ulceration, sentinel lymph node status, and the number of tumor-positive lymph nodes (two-sided p < 0.0001 for all comparisons). Multivariable analysis revealed that sentinel lymph node status (p = 0.003), tumor thickness (p = 0.016), ulceration (p = 0.006), and age (p = 0.003) were significant independent predictors of survival for the entire group. Tumor thickness and ulceration were significant predictors of recurrence and survival in sentinel node-negative patients but not in sentinel node-positive patients. Sentinel lymph node histology is possibly the most important negative predictor of early recurrence and survival in patients with American Joint Committee on Cancer stage I and II melanoma. The number of positive lymph nodes provides additional prognostic information. Although sentinel node-negative patients are a prognostically favorable group, various combinations of local and regional recurrences comprise the most common pattern of initial relapse after a negative sentinel lymph node biopsy result.

Adolescent↗

PubMed2002

Sample size calculation for rank tests comparing K survival distributions.

Rank tests, such as logrank or Wilcoxon rank sum tests, have been popularly used to compare survival distributions of two or more groups in the presence of right censoring. However, there has been little research on sample size calculation methods for rank tests to compare more than two groups. An existing method is based on a crude approximation, which tends to underestimate sample size, i.e., the calculated sample size has lower power than projected. In this paper we propose an asymptotically correct method and an approximate method for sample size calculation. The proposed methods are compared to other methods through simulation studies.

Algorithms↗

PubMed2002

Application of an adjusted chi2 statistic to site-specific data in observational dental studies.

BACKGROUND: When a binary response is observed on teeth from each subject belonging to 2 or more exposure groups, application of the usual Pearson chi2 tests is invalid, since such responses within the same subject are not independent. Consequently, special statistical methods are needed to control for the correlation among teeth (sites) within the same subject. A simple adjustment to the Pearson chi2 statistic has been proposed for comparing proportions in site-specific data. However, the required assumptions for this statistic have not yet been thoroughly addressed. These assumptions are guaranteed to hold in experimental comparisons, but may be violated in some observational studies. METHOD: We investigate the conditions under which the adjusted chi2 statistic is valid and examine the performance of the adjusted chi2 statistic when these conditions are violated. RESULTS: Our simulation study shows that the adjusted chi2 statistic generally produces good empirical type I errors under the assumption of a common intracluster correlation coefficient. Even if the intracluster correlations are different, the adjusted statistic performs well when the groups have equal numbers of clusters (subjects). CONCLUSION: The discussion is illustrated using an observational study of caries on the roots of teeth.

Analysis of Variance↗