摘要
据WHO官方网站公布的信息,恶性肿瘤目前在世界范围内是导致死亡的最主要的原因之一。2008年约有760万人死于恶性肿瘤,约占总死亡人数的13%;预计到2030年,世界范围内死于恶性肿瘤的人数将超过1100万人。调查显示,我国恶性肿瘤发病数正以年均3%~5%的速度递增。在城市,肺癌、胃癌和乳腺癌的发病率最高;在农村,胃癌、食管癌和肝癌是前三位的高发癌症。因此,针对抗肿瘤治疗方法的研究越来越引起了社会各界的广泛重视。
寻找抗肿瘤治疗效果证据的最安全最有效的途径是临床试验。临床试验的主要目的是研究治疗的有效性与安全性。肿瘤临床试验作为临床试验的一个特殊分支,具有自己的特点:首先,结局指标往往是某结局事件发生的时间;其次,随访时间通常较长,病人在随访中更容易发生删失,因此最常用的分析方法是生存分析。
由于肿瘤临床试验的特殊性,在肿瘤临床试验的设计和数据分析中,可能遇到各种各样的问题。到目前为止,仍然存在一些问题尚未解决。本论文的目的就是探索与研究在肿瘤临床试验中遇到的三个问题的解决方法。
第一个问题,在一类探索新治疗有效性的研究中,研究者为在常规治疗下发生了进展(一般定义为肿瘤增大25%)的带瘤病人尝试新的治疗,为了在有限的成本下能够对新治疗的效果进行评价,研究者怎样利用这些病人在之前接受的常规治疗下的肿瘤进展时间或其相关函数作为对照,与新治疗下的肿瘤进展时间或其相关函数进行比较,来对新治疗的效果进行评价。
在比例风险假定成立时,我们给出了相关生存数据在非劣性假设下,用非参数的加权秩检验进行分析的方法,对新治疗的效果进行评价。并且通过随机模拟的方法对检验的效果进行了评估,发现效果很好。该方法不仅限于以上情形,更可以应用于任何配对或相关生存数据的非劣性假设检验中。
另外,我们在无删失和存在删失的条件下,分别给出了非劣性假设检验相关生存数据所需样本量的计算公式,并在计算出的样本量下通过随机模拟并计算实际的统计效能(power)。从结果中不难看出计算出的样本量完全能够满足power的要求。因此,我们给出的样本量公式可以为临床试验的设计提供参考。我们利用一个乳腺癌临床试验的例子说明了检验的方法及样本量计算公式的可行性。
第二个问题,在临床试验得到的配对或相关生存数据中,如果不满足比例风险假定,我们对随访过程中任意时间点上的风险比可能更感兴趣。针对在比例风险假定不满足时,如何计算任意时间点上的风险比的问题,我们给出了不需要比例风险模型假定的非参数的核方法来估计任意时间点上的风险比的及其置信区间。目前为止,这种核方法仅被用于当两个治疗组的生存时间独立的情形之下。
我们将非参数核方法用于估计边际风险率,并将近年来提出的MOVER方法用在构造两个相关的生存数据在随访过程中任意时间点的边际风险率之比的置信区间中。在构造单个边际风险率的置信区间时,我们分别采用了"delta"法与"transform"法。通过随机模拟,我们可以看出当样本量很大并且时间点不是很靠近边界时,两种方法的效果都很好。我们也将这两种方法运用于实际的临床试验数据中,用以说明其可行性。
第三个问题,在平行对照的临床试验中,将病人随机分为两组,新治疗组与对照治疗组,在中期分析中,发现所试验的治疗比对照治疗好,使得原来设计的试验不得不终止,并允许对照治疗的病人选择开始接受新治疗。此时,由于随访时间较短,得到的仅是治疗的短期疗效指标的结果,为了观察新疗法的长期效果,因此继续对病人进行观察。我们希望能得到如果没有发生治疗转换,一直继续随访,最终如何评价两种治疗效果的差异。
我们比较了利用反事实风险率构造的SCC法、利用逆概率删失加权的IPCW方法和简单ITT法、在ITT集上调整了与治疗转换和结局都有关的协变量的COX回归相比的效果。通过随机模拟,我们发现SCC法与IPCW法都能校正治疗转换效应,并且与ITT和COX相比更加能够反映两种治疗的真实差异。于是,虽然SFDA要求使用ITT法进行分析,但在遇到病人发生治疗转换时,SCC与IPCW法可依作为ITT法的补充。我们也将SCC法与IPCW法用于解决加拿大国家癌症研究所的临床试验组(NCIC CTG)负责的一个随机双盲安慰剂对照的肿瘤临床试验中遇到的问题。
According to the World Health Organization's official website, cancer is a leading cause of death worldwide. In 2008, cancer accounted for 7.6 million deaths, which is around 13% of all deaths. Deaths from cancer worldwide are projected to continue to rise to over 11 million in 2030. Recently, surveys revealed that the number of newly developed cancer patients is increasing by 3% to 5% per year in China. In urban dis-tricts, the top three cancers with the highest incidence rate are lung cancer, stomach cancer and breast cancer, whereas in the rural areas, stomach cancer, esophageal can-cer and liver cancer are the three with the highest incidence rate.
The safest and most effective way to find the evidence of anti-cancer treatment is through clinical trials. The main purpose of clinical trials is to evaluate the efficacy and safety of anti-cancer treatment. Cancer clinical trial as a special branch of clinical trials has its own features:first, the follow-up time is usually relatively long; second, the most common endpoints are time-to-event; in addition, the major analyzing me-thod of cancer clinical trials is survival analysis.
Because of the features above, during the design and the data analysis of cancer clinical trials, one might come across all sorts of problems. Until now, there still are some of these problems being unsolved. The main purpose of this thesis is to study on three problems we met in cancer clinical trials.
First, in exploratory research, to make a preliminary evaluation of the efficacy of the new treatment, researchers will try the new treatment on those cancer patients with cancer progression (usually defined as 25% increase of the tumor size) under the standard treatment. To control the research cost, how to evaluate the efficacy of the new treatment by using the performance of treatment before the new treatment as a control to be compared with the performance of the new treatment.
When proportional hazard assumption holds, we derived a class of non-parametric weighted rank tests for paired survival times to evaluate the efficacy of the new treatment. Monte-Carlo simulations were performed to evaluate the per-formance of the proposed tests and the results were satisfactory. The proposed me- thods are not only limited to the scenario above, they could be used to verify any non-inferiority hypothesis for paired or correlated survival data.
In addition, we also derived the sample size formula under no-censoring or cen-soring situation for these tests. Through simulations, the statistical powers were gen-erated under the calculated sample size. The results showed the sample size could gain enough statistical power. So, the sample size formula could be used as a refer-ence to design clinical trials. As an example, a phaseⅡclinical trial of metastatic breast cancer was taken to demonstrate the proposed tests and to illustrate the feasibil-ity of the sample size formula.
Second, if the proportional hazard assumption does not be satisfied by the paired or correlated survival data from the clinical trials, the inferences on hazard ratio at each specific time point of follow-up may be of more interest. How to detect the dif-ference between the two treatment groups? We used the recently proposed non-parametric kernel method, which doesn't need the proportional harzard assump-tion, to estimate the time-dependent hazard ratio and its confidence interval. Until now, this kernel method had only been used in the situation when the survival times of two treatment groups are independent.
We used the non-parametric kernel method to estimate each marginal hazard rate and the variance estimates recovery (MOVER) method to construct the confidence intervals of time-dependent hazard ratios based on confidence limits of each hazard rate. Two methods were proposed when constructing confidence limits for each mar-ginal hazard rate:"delta" method and "transform" method. Simulation results showed that both methods can perform well, especially when sample size is large and the time points are not too close to the boundary. The proposed methods were also applied to datasets from real clinical trials to show their feasibility.
Third, in parallel controlled clinical trials, we randomly assign patients into two groups, the new treatment group and the control group. After the interim analysis, if a short-term endpoint of the new treatment, such as disease free survival, is significant-ly better than the controlled treatment, the trial has to be stopped, and those patients in the control arm are offered a chance to switch to the new treatment. The conclusion we drawn after interim analysis is only the short-term effect of the treatment. To ob-serve the long-term effect of the treatment, we continue to follow up those patients. Our concern is if those patients who crossed over to the new treatment group who did not switch and were continued to be followed up, how to detect the real difference between the two treatment groups.
We compared the SCC method which based on counterfactual hazard rates and the IPCW method which uses the inverse probability of censoring rate to weigh the patients to simple ITT method and the COX regression adjusting the covariate both related to switch probability and outcome based on ITT set. Through simulations, we found the SCC and the IPCW method can both adjust the treatment crossover effect and outperform the ITT and the COX method in revealing the real difference between the two treatments. So, although the State Food and Drug Administration recommends the ITT method in the clinical trial analysis, we still could use the results from the SCC and the IPCW method as supplemental analysis when there is treatment crossov-er. Real data from a randomized double-blind placebo-controlled clinical trial from the National Cancer Institute of Canada Clinical Trial Group is used to support the results.
引文
[1]Mick R, Crowley JJ, Carroll RJ. Phase II clinical trial design for noncytotoxic an-ticancer agents for which time to disease progression is the primary endpoint[J]. Controlled Clinical Trials,2000,21:343-359.
[2]Von Hoff DD. There are no bad anticancer agents, only bad clinical trial designs: twenty-first Richard and Hinda Rosenthal Foundation Award Lecture[J]. Clinical Cancer Research,1998,4:1079-1086.
[3]Tu D. Under-Smoothed Kernel Confidence Intervals for the Hazard Ratio Based on Censored Data[J]. Biometrical Journal,2007,3:474-483.
[4]Zou GY, Donner A. Construction of confidence limits about effect measures:A general approach[J]. Statistics in Medicine,2008,27:1693-1702.
[5]Ramasundarahettige CF, Donner A, Zou GY. Confidence interval construction for a difference between two dependent intraclass correlation coefficients [J]. Statistics in Medicine,2009,28:1041-1053.
[6]Tang ML, Ling MH, Lin L, et al. Confidence intervals for a difference between proportions based on paired data[J]. Statistics in Medicine,2010,29:86-96.
[7]Morden JP, Lambert PC, Latimer N, et al. Assessing methods for dealing with treatment switching in randomized controlled trials[R]. Technical Reports 2010/BS01, Department of Health Sciences, University of Leicenster.
[8]Branson M, Whitehead J:Estimating a treatment effect in survival studies in which patients switch treatment[J]. Statistics in Medicine,2002,21:2449-2463.
[9]Robins JM, Tsiatis AA. Correcting for non-compliance in randomized trials using rank preserving structural failure time models. [J] Communications in Statistics-Theory and Methods,1991,20(8):2609-2631.
[10]Shao J, Chang M, Chow SC. Statistical inference for cancer trials with treatment switching[J]. Statistics in Medicine,2005,24:1783-1790.
[11]Robins JM, Finkelstein DM. Correcting for noncompliance and dependent cen-soring in an AIDS clinical trial with inverse probability of censoring weighted (IPCW) log-rank tests[J]. Biometrics,2000,56(3):779-788.
[12]Colleoni M, Giobbie-Hurder A, Regan MM et al. Analyses adjusting for selective crossover show improved overall survival with adjuvant Letrozole compared with Tamoxifen in the BIG 1-98 study [J]. Published online before print February 14, 2011,doi:10.1200/JCO.2010.31.6455.
[13]Yoshida M, Matsuyama Y, Ohashi Y et al. Estimation of treatment effect adjust-ing for dependent censoring using the IPCW method:an application to a large primary prevention study for coronary events (MEGA study)[J]. Clinical Trials, 2007,4:318-328.
[14]Folkman J. Fighting cancer by attacking its blood supply. Scientific American Magazine,1996,275:150-154.
[15]Kovalchik S, Mietlowski W. Statistical methods for a phase II oncology trial with a growth modulation index (GMI) endpoint[J]. Contemporary Clinical Trials, 2011,32(1):99-107.
[16]Bonetti A, Zaninelli M, Leone R, et al. Use of the ratio of time to progression following first-and second-line therapy to document the activity of the combina-tion of oxaliplatin with 5-uorouracil in the treatment of colorectal carcinoma[J]. Annal of Oncology,2001,12:187-191.
[17]Bachet J, Mitry E, Lievre A, et al. Second- and third-line chemotherapy in pa-tients with metastatic pancreatic adenocarcinoma:feasibility and potential benefits in a retrospective series of 117 patients[J]. Gastroenterologie Clinique Biologique, 2009,33:1036-1044.
[18]Comella P, Casaretti R, Crucitta E, et al. Oxaliplatin plus raltitrexed and leucovo-rin-modulated 5-fluorouracil i.v. bolus:a salvage regimen for colorectal cancer patients[J]. British Journal of Cancer,2002,86:1871-1875.
[19]Gebbia V, Maiello E, Giuliani F, et al. Second-line chemotherapy in advanced pancreatic carcinoma:a multicenter survey of the gruppo oncologico italia meri-dionale on the activity and safety of the folfox4 regimen in clinical practice[J]. Annal of Oncology,2007,18(Supplement 6):vi124-vi127.
[20]Demetri G, Chawla S, von Mehren M, et al. Efficacy and safety of trabectedin in patients with advanced or metastatic liposarcoma or leiomyosarcoma after failure of prior anthracyclines and ifosfamide:results of a randomized phase Ⅱ study of two different schedules[J]. Journal of Clinical Oncology,2009,27:4188-4196.
[21]Scott WF. The estimation of the hazard ratio in clinical trials and in meta-analysis [J]. The Mathematical Scientist,2000,25:54-58.
[22]Wei LJ, Lin DY, Weissfeld L. Regression analysis of multivariate incomplete failure time data by modeling marginal distributions[J]. Journal of the American Statistical Association,1989,84:1065-1073.
[23]Lee EW, Wei LJ, Amato DA. COX-type regression analysis for large numbers of small groups of correlated failure time observations[A]. In:Klein J.P., Goel P.K. (eds) Survival Analysis:State of the Art. Dordrecht:Kluwer Academic Publishers, 1992:237-247.
[24]Lee EW, Wei LJ, Ying Z. Linear regression analysis for highly stratified failure time data[J]. Journal American Statistial Association,1993,88:557-565.
[25]Liang KY, Self SG, Chang YC. Modelling marginal hazards in multivariate fail-ure time data[J]. Journal Royal Statistical Society,1993,55:441-453.
[26]Cai J, Prentice RL. Estimating equations for hazard ratio parameters based on correlated failure time data[J]. Biometrika,1995,82:151-164.
[27]Oakes D. Bivariate survival models induced by frailties[J]. Journal of the Ameri-can Statistical Association,1989,84(406):487-493.
[28]Hougaard P. Frailty models for survival data. Lifetime Data Analysis,1995, 1:255-273
[29]Shih JH, Louis TA. Assessing gamma frailty models for clustered failure time data[J]. Lifetime Data Analysis,1995,1:205-220.
[30]Xie X. Bivariate frailty model for the analysis of multivariate survival time[J]. Lifetime Data Analysis,1996,2:277-289.
[31]Parner E. Asymptotic Theory for the Correlated Gamma-Frailty Model[J]. The Annals of Statistics,1998,26:183-214.
[32]Lin JS, Wei LJ. Linear regression analysis for multivariate failure time observa-tions[J]. Journal of American Statistical Association,1992,87:1091-1097.
[33]Cai T, Cheng SC, Wei LJ. Semiparametric mixed-effects models for clustered failure time data[J]. Journal of the American Statistical Association,2002, 97:514-522.
[34]Zeng D, Lin, DY, Lin X. Semiparametric transformation models with random effects for clustered failure time data[J]. Statistica Sinica 2008; 18:355-377.
[35]Woolson RF, Lachenbruch PA. Rank tests for censored matched pairs[J]. Biome-trika,1980,67:597-606.
[36]Wei LJ. A generalized Gehan and Gilbert test for paired observations that are subject to arbitrary right censorship [J]. Journal of the American Statistical Associ-ation,1980,75:634-637.
[37]O'Brien PC, Fleming TR. A Paired Prentice-Wilcoxon Test for Censored Paired Data[J]. Biometrics,1987,43:169-180.
[38]Jung SH. Rank tests for matched survival data[J]. Lifetime Data Analysis,1999, 5:67-69.
[39]Dallas MJ, Rao PV. Testing equality of survival functions based on both paired and unpaired censored data[J]. Biometrics,2000,56:154-159.
[40]Jones MP, Yoo B. Linear signed-rank tests for paired survival data subject to a common censoring time[J]. Lifetime Data Analysis,2005,11:351-365.
[41]Oakes D, Feng C. Combining stratified and unstratified log-rank tests in paired survival data[J]. Statistics in Medicine,2010,29:1735-1745.
[42]Jung SH, Kang SJ, McCall LM, Blumenstein, B. Sample size computation for two-sample noninferiority log-rank test[J]. Journal of Biopharmaceutical Statistics, 2005,15:969-979.
[43]Jung SH. Sample size calculation for the weighted rank statistics with paired sur-vival data[J]. Statistics in Medicine,2008,27:3350-3365.
[44]Fleming TR, Harrington DP. Counting Processes and Survival Analysis[M]. New York:Wiley,1991.
[45]Gumbel EJ. Bivariate exponential distributions[J]. Journal of the American Sta-tistical Association,1960,55:698-707.
[46]Lee L. Multivariate distributions having Weibull properties[J]. Journal of Multi-variate Analysis,1979,9:267-277.
[47]Khoo K, Brandes L, Reyno L et al. Phase Ⅱ trial of N,N-diethyl-2-[4-(phenylmethyl)phenoxy]ethanamine.HCl and doxorubicin che-motherapy in metastatic breast cancer:A National Cancer Institute of Canada clinical trials group study [J]. Journal of Clinical Oncology,1999,17:3431-3437.
[48]Reyno L, Seymour L, Tu D, et al. Phase Ⅲ study of N,N-Diethyl-2-[4-(Phenylmethyl) Phenoxy] Ethanamine (BMS-217380-01) com-bined with doxorubicin versus doxorubicin alone in metastatic/recurrent breast cancer:National Cancer Institute of Canada Clinical Trials Group Study MA.19[J]. Journal of Clinical Oncology,2004,22:269-276.
[49]Batchelor JR, Hackett M. HL-A matching in treatment of burned patients with skin allografts[J]. The Lancet,1970,296(7673):581-583.
[50]Huster WJ, Brookmeyer R, Self SG. Modelling Paired Survival Data with Co-variates. Biometrics,1989,45:145-156.
[51]Rothmann M, Li N, Chen G, et al. Design and analysis of non-inferiority mortal-ity trials in oncology [J]. Statistics in Medicine,2003,22:239-264
[52]Freitag G. Methods for assessing noninferiority with censored data[J]. Biometri- cal Journal,2005,47:88-98.
[53]Freitag G, Lange S and Munk A. Nonparametric assessment of non-inferiority with Censored data[J]. Statistics in Medicine,2006,25,1201-1217.
[54]Stablein DM, Carter WH, Novak JW. Analysis of survival data with nonpropor-tional hazard functions[J]. Controlled Clinical Trials,1981,2:149-159.
[55]Abrahamowicz M, MacKenzie T, Esdaile JM. Time-dependent hazard ratio: Modeling and hypothesis testing with application in lupus nephritis [J]. Journal of the American Statistical Association,1996,91:1432-1439.
[56]Martinussen T, Scheike TH, Skovegaard IM, et al. Efficient Estimation of Fixed and Time-varying Covariate Effects in Multiplicative Intensity Models[J]. Scan-dinavian Journal of Statistics,2002,29:57-74.
[57]Jung SH, Su JQ. Non-parametric estimation for the difference of ratio of median failure times for paired observations [J]. Statistics in medicine,1995,14:275-281.
[58]Tanner MA, Wong WH. The estimation of the hazard function from randomly censored data by the kernel method[J]. The Annals of Statistics,1983, 11(3):989-993.
[59]Yandell BS. Nonparametric inference for rates with censored survival data[J]. The Annals of Statistics,1983,11(4):1119-1135.
[60]Cheng M, Hall P, Tu D. Confidence bands for hazard rate under random censor-ship[J]. Biometrika,2006,93:357-366.
[61]Yan J. Enjoy the Joy of Copulas:With a Package copula[J]. Journal of Statistical Software,2007,21:1-21.
[62]Nelsen RB. An introduction to copulas[M]. New York:Springer,2006:158.
[63]Singer J. Letter to the editor:Construction of confidence limits about effect measures:A general approach[J]. Statistics in Medicine,2010,29:1757-1759.
[64]Liang KY, Self SG, Chang YC. Modeling marginal hazards in multivariate fail-ure time data[J]. Journal Royal Statistical Society,1993,55:441-453.
[65]Cheng MY, Qiu P, Tan X, Tu D. Confidence intervals for the first crossing point of two hazard functions[J]. Lifetime Data Analysis,2009,15:441-454.
[66]Fleischer F, Gaschler-Markefski B, Bluhmki E. A statistical model for the de-pendence between progression-free survival and overall survival [J]. Statistics in Medicine,2009,28:2669-2686.
[67]Zou GY, Donner A. Construction of confidence limits about effect measures:A general approach-reply [J]. Statistics in Medicine,2010,29:1760.
[68]Gilbert PB, Wei LJ, Kosorok MR and Clemens JD. Simultaneous inferences on the contrast of two hazard functions with censored observations[J]. Biomet-rics,2002,58:773-780.
[69]Hall P and Owen AB. Empirical likelihood confidence bands in density estima-tion[J]. Journal of Computational and Graphical Statistics,1993,2:273-289.
[70]Bickel P. and Rosenblatt M. On some global estimate of the deviations of density estimation[J]. Annals of Statistics,1993,1:1071-1095.
[71]Cao R, Janssen P and Veraverbeke N. Relative hazard rate estimation for right censored and left trucated data[J]. Test,2005,14:257-280.
[72]Lewis R, Bagnall AM, Forbes C, et al. A rapid and systematic review of the clin-ical effectiveness and costeffectiveness of trastuzumab for breast cancer[R]. http://www.nice.org.uk/nicemedia/live/11445/32317/32317.pdf,2001.
[73]Abrams K, Palmer S, Wailoo A. Bevacizumab, sorafenib, sunitinib and temsiro-limus for renal cell carcinoma[R]. Decision Support Unit Report, http://www.nice.org.uk/nicemedia/pdf/RenalCellCarcinomaExtraWorkPreparedBy DSU.pdf,2008.
[74]Law MG, Kaldor JM. Survival analyses of randomized clinical trials adjusted for patients who switch treatments[J]. Statistics in Medicine,1996,15:2069-2076.
[75]White IR. Letters to the editor:Survival analyses of randomized clinical trials adjusted for patients who switch treatments [J]. Statistics in Medicine,1997, 16:2619-2625.
[76]Loeys T, Goetghebeur E. A causal proportional hazards estimator for the effect of treatment actually received in a randomized trial with all-or-nothing compliance [J]. Biometrics,2003,59:100-105.
[77]Robins JM. Information recovery and bias adjustment in proportional hazards regression analysis of randomized trials using surrogate markers. In:American Statistical Association. Proceedings of the Biopharmaceutical Section[M], Alex-andria, Virginia:American Statistical Association,1993,24-33.
[78]Karnon J, Kaura S. Updated survival-based analysis using inverse probability of censoring weighted analysis (IPCW) to estimate the cost-effectiveness of letrozole and anastrozole versus tamoxifen as adjuvant therapy in postmenopausal women with early breast cancer[J]. Journal of Clinical Oncology,2010,28(Supplement 15):656.
[79]Robins JM. Causal inference from complex longitudinal data. In:Berkane M. Latent Variable Modeling and Applications to Causality[M]. New York:Springer-Verlag,1997:69-117.
[80]Robins JM, Hernan MA, Brumback B. Marginal Structural Models and Causal Inference in Epidemiology[J]. Epidemiology.2000, 11(5):550-560.
[81]D'Agostino RB, Lee ML, Belanger AJ, Cupples LA, Anderson K, Kannel WB. Relation of pooled logistic regression to time dependent COX regression analysis: the Framingham Heart Study [J]. Statistics in Medicine.1990,9(12):1501-1515.
[82]Collett D. Modelling survival data in medical research[M]. London:Chapman and Hall,2003:200-206.
[83]Goss PE, Ingle JN, Martino S, et al. A randomized trial of letrozole in postmeno-pausal women after five years of Tamoxifen therapy for early-stage breast can-cer[J]. New England Journal of Medicine,2003,349:1793-1802.
[84]Goss PE, Ingle JN, Martino S, et al. Randomized trials of Letrozole following Tamoxifen as extended adjuvant therapy in receptor positive breast cancer:up-dated findings from NCIC CTG MA.17[J]. Journal of National Cancer Institute, 2005,97:1262-1271.
[85]Ingle JN, Tu D, Pater JL, et al. Intent-to-treat analysis of the placebo-controlled trial of letrozole for extended adjuvant therapy in early breast cancer:NCIC CTG MA.17[J]. Annals of Oncology,2008,19:877-882.
[86]Goss PE, Ingle JN and Pater JL. Late extended adjuvant treatment with letrozole improves outcome in women with early-stage breast cancer completing 5 Years of tamoxifen[J]. Journal of Clinical Oncology,2008,26:1948-1955.
[87]Korn EL and Freidlin B. Causal inference for definitive clinical end points in a randomized clinical trial with intervening nonrandomized treatments [J]. Journal of Clinical Oncology,2010,28:3808-3815.
[1]Mick R, Crowley JJ, Carroll RJ. Phase Ⅱ clinical trial design for noncytotoxic an-ticancer agents for which time to disease progression is the primary endpoint[J]. Controlled Clinical Trials,2000,21:343-359.
[2]Von Hoff DD. There are no bad anticancer agents, only bad clinical trial designs: twenty-first Richard and Hinda Rosenthal Foundation Award Lecture [J]. Clinical Cancer Research,1998,4:1079-1086.
[3]Kovalchik S, Mietlowski W. Statistical methods for a phase Ⅱ oncology trial with a growth modulation index (GMI) endpoint[J]. Contemporary Clinical Trials, 2011,32(1):99-107.
[4]Bonetti A, Zaninelli M, Leone R, et al. Use of the ratio of time to progression fol-lowing first-and second-line therapy to document the activity of the combination of oxaliplatin with 5-uorouracil in the treatment of colorectal carcinoma[J]. Annal of Oncology,2001,12:187-191.
[5]Bachet J, Mitry E, Lievre A, et al. Second-and third-line chemotherapy in patients with metastatic pancreatic adenocarcinoma:feasibility and potential benefits in a retrospective series of 117 patients[J]. Gastroenterologie Clinique Biologique, 2009,33:1036-1044.
[6]Comella P, Casaretti R, Crucitta E, et al. Oxaliplatin plus raltitrexed and leucovo-rin-modulated 5-fluorouracil i.v. bolus:a salvage regimen for colorectal cancer patients[J]. British Journal of Cancer,2002,86:1871-1875.
[7]Gebbia V, Maiello E, Giuliani F, et al. Second-line chemotherapy in advanced pancreatic carcinoma:a multicenter survey of the gruppo oncologico italia meri-dionale on the activity and safety of the folfox4 regimen in clinical practice[J]. Annal of Oncology,2007,18(Supplement 6):vi124-vi127.
[8]Demetri G, Chawla S, von Mehren M, et al. Efficacy and safety of trabectedin in patients with advanced or metastatic liposarcoma or leiomyosarcoma after failure of prior anthracyclines and ifosfamide:results of a randomized phase Ⅱ study of two different schedules[J]. Journal of Clinical Oncology,2009,27:4188-4196.
[9]Scott WF. The estimation of the hazard ratio in clinical trials and in meta-analysis [J]. The Mathematical Scientist,2000,25:54-58.
[10]Wei LJ, Lin DY, Weissfeld L. Regression analysis of multivariate incomplete failure time data by modeling marginal distributions[J]. Journal of the American Statistical Association,1989,84:1065-1073.
[11]Lee EW, Wei LJ, Amato DA. COX-type regression analysis for large numbers of small groups of correlated failure time observations[A]. In:Klein J.P., Goel P.K. (eds) Survival Analysis:State of the Art. Dordrecht:Kluwer Academic Publishers, 1992:237-247.
[12]Lee EW, Wei LJ, Ying Z. Linear regression analysis for highly stratified failure time data[J]. Journal American Statistial Association,1993,88:557-565.
[13]Liang KY, Self SG, Chang YC. Modelling marginal hazards in multivariate fail-ure time data[J]. Journal Royal Statistical Society,1993,55:441-453.
[14]Cai J, Prentice RL. Estimating equations for hazard ratio parameters based on correlated failure time data[J]. Biometrika,1995,82:151-164.
[15]Oakes D. Bivariate survival models induced by frailties[J]. Journal of the Ameri-can Statistical Association,1989,84(406):487-493.
[16]Hougaard P. Frailty models for survival data. Lifetime Data Analysis,1995, 1:255-273
[17]Shih JH, Louis TA. Assessing gamma frailty models for clustered failure time data[J]. Lifetime Data Analysis,1995,1:205-220.
[18]Xie X. Bivariate frailty model for the analysis of multivariate survival time[J]. Lifetime Data Analysis,1996,2:277-289.
[19]Parner E. Asymptotic Theory for the Correlated Gamma-Frailty Model [J]. The Annals of Statistics,1998,26:183-214.
[20]Lin JS, Wei LJ. Linear regression analysis for multivariate failure time observa-tions[J]. Journal of American Statistical Association,1992,87:1091-1097.
[21]Cai T, Cheng SC, Wei LJ. Semiparametric mixed-effects models for clustered failure time data[J]. Journal of the American Statistical Association,2002, 97:514-522.
[22]Zeng D, Lin, DY, Lin X. Semiparametric transformation models with random effects for clustered failure time data[J]. Statistica Sinica 2008; 18:355-377.
[23]Woolson RF, Lachenbruch PA. Rank tests for censored matched pairs[J]. Biome-trika,1980,67:597-606.
[24]Wei LJ. A generalized Gehan and Gilbert test for paired observations that are subject to arbitrary right censorship[J]. Journal of the American Statistical Associ-ation,1980,75:634-637.
[25]O'Brien PC, Fleming TR. A Paired Prentice-Wilcoxon Test for Censored Paired Data[J]. Biometrics,1987,43:169-180.
[26]Jung SH. Rank tests for matched survival data[J]. Lifetime Data Analysis,1999, 5:67-69.
[27]Dallas MJ, Rao PV. Testing equality of survival functions based on both paired and unpaired censored data[J]. Biometrics,2000,56:154-159.
[28]Jones MP, Yoo B. Linear signed-rank tests for paired survival data subject to a common censoring time[J]. Lifetime Data Analysis,2005,11:351-365.
[29]Oakes D, Feng C. Combining stratified and unstratified log-rank tests in paired survival data[J]. Statistics in Medicine,2010,29:1735-1745.
[30]Jung SH, Kang SJ, McCall LM, Blumenstein, B. Sample size computation for two-sample noninferiority log-rank test[J]. Journal of Biopharmaceutical Statistics, 2005,15:969-979.
[31]Jung SH. Sample size calculation for the weighted rank statistics with paired sur-vival data[J]. Statistics in Medicine,2008,27:3350-3365.
[32]Batchelor JR, Hackett M. HL-A matching in treatment of burned patients with skin allografts[J]. The Lancet,1970,296(7673):581-583.
[33]Huster WJ, Brookmeyer R, Self SG. Modelling Paired Survival Data with Co-variates. Biometrics,1989,45:145-156.
[34]Jung SH, Su JQ. Non-parametric estimation for the difference of ratio of median failure times for paired observations[J]. Statistics in medicine,1995,14:275-281.
[35]Stablein DM, Carter WH, Novak JW. Analysis of survival data with nonpropor-tional hazard functions[J]. Controlled Clinical Trials,1981,2:149-159.
[36]Abrahamowicz M, MacKenzie T, Esdaile JM. Time-dependent hazard ratio: Modeling and hypothesis testing with application in lupus nephritis[J]. Journal of the American Statistical Association,1996,91:1432-1439.
[37]Martinussen T, Scheike TH, Skovegaard IM, et al. Efficient Estimation of Fixed and Time-varying Covariate Effects in Multiplicative Intensity Models[J]. Scan-dinavian Journal of Statistics,2002,29:57-74.
[38]Tu D. Under-Smoothed Kernel Confidence Intervals for the Hazard Ratio Based on Censored Data[J]. Biometrical Journal,2007,3:474-483.
[39]Zou GY, Donner A. Construction of confidence limits about effect measures:A general approach[J]. Statistics in Medicine,2008,27:1693-1702.
[40]Ramasundarahettige CF, Donner A, Zou GY. Confidence interval construction for a difference between two dependent intraclass correlation coefficients[J]. Statistics in Medicine,2009,28:1041-1053.
[41]Tang ML, Ling MH, Lin L, et al. Confidence intervals for a difference between proportions based on paired data[J]. Statistics in Medicine,2010,29:86-96.
[42]Morden JP, Lambert PC, Latimer N, et al. Assessing methods for dealing with treatment switching in randomized controlled trials[R]. Technical Reports 2010/BS01, Department of Health Sciences, University of Leicenster.
[43]Law MG, Kaldor JM. Survival analyses of randomized clinical trials adjusted for patients who switch treatments[J]. Statistics in Medicine,1996,15:2069-2076.
[44]White IR. Letters to the editor:Survival analyses of randomized clinical trials adjusted for patients who switch treatments [J]. Statistics in Medicine,1997, 16:2619-2625.
[45]Loeys T, Goetghebeur E. A causal proportional hazards estimator for the effect of treatment actually received in a randomized trial with all-or-nothing compliance [J]. Biometrics,2003,59:100-105.
[46]Robins JM, Tsiatis AA. Correcting for non-compliance in randomized trials using rank preserving structural failure time models. [J] Communications in Statistics-Theory and Methods,1991,20(8):2609-2631.
[47]Branson M, Whitehead J:Estimating a treatment effect in survival studies in which patients switch treatment[J]. Statistics in Medicine,2002,21:2449-2463.
[48]Shao J, Chang M, Chow SC. Statistical inference for cancer trials with treatment switching[J]. Statistics in Medicine,2005,24:1783-1790.
[49]Robins JM. Information recovery and bias adjustment in proportional hazards regression analysis of randomized trials using surrogate markers. In:American Statistical Association. Proceedings of the Biopharmaceutical Section[M], Alex-andria, Virginia:American Statistical Association,1993,24-33.
[50]Colleoni M, Giobbie-Hurder A, Regan MM et al. Analyses adjusting for selective crossover show improved overall survival with adjuvant Letrozole compared with Tamoxifen in the BIG 1-98 study[J]. Published online before print February 14, 2011,doi:10.1200/JCO.2010.31.6455.
[51]Karnon J, Kaura S. Updated survival-based analysis using inverse probability of censoring weighted analysis (IPCW) to estimate the cost-effectiveness of letrozole and anastrozole versus tamoxifen as adjuvant therapy in postmenopausal women with early breast cancer[J]. Journal of Clinical Oncology,2010,28(Supplement 15):656.
[52]Yoshida M, Matsuyama Y, Ohashi Y et al. Estimation of treatment effect adjust-ing for dependent censoring using the IPCW method:an application to a large primary prevention study for coronary events (MEGA study)[J]. Clinical Trials, 2007,4:318-328.