Sargent, Daniel J.

A two‐stage model for evaluating both trial‐level and patient‐level surrogacy of correlated time‐to‐event endpoints has been introduced, using patient‐level data when multiple clinical trials are available. However, the associated maximum likelihood approach often suffers from numerical problems when different baseline hazards among trials and imperfect estimation of treatment effects are assumed. To address this issue, we propose performing the second‐stage, trial‐level evaluation of potential surrogates within a Bayesian framework, where we may naturally borrow information across trials while maintaining these realistic assumptions. Posterior distributions on surrogacy measures of interest may then be used to compare measures or make decisions regarding the candidacy of a specific endpoint. We perform a simulation study to investigate differences in estimation performance between traditional maximum likelihood and new Bayesian representations of common meta‐analytic surrogacy measures, while assessing sensitivity to data characteristics such as number of trials, trial size, and amount of censoring. Furthermore, we present both frequentist and Bayesian trial‐level surrogacy evaluations of time to recurrence for overall survival in two meta‐analyses of adjuvant therapy trials in colon cancer. With these results, we recommend Bayesian evaluation as an attractive and numerically stable alternative in the multitrial assessment of potential surrogate endpoints. Copyright © 2011 John Wiley & Sons, Ltd.

Abstract Purpose: Assessing impact of poor accrual on premature trial closure requires a relevant metric. We propose defining accrual sufficiency on apparent ability to address primary endpoints (PE) rather than attaining accrual targets. Experimental Design: All phase III trials open January 1, 1993, to December 31, 2002, by five U.S. oncology Clinical Trials Cooperative Groups (CTCG) were evaluated for accrual sufficiency and scientific results. Sufficient accrual included meeting accrual target, CTCGs documentation attesting adequate accrual, or conclusive results at interim analysis; insufficient accrual included poor accrual as cited closure reason or other reasons rendering a trial unable to address its primary endpoints. Closure rates based on our accrual sufficiency definition are compared with rates of meeting accrual targets and addressing the primary endpoints. A percentage of target accrual above which trials commonly answer the intended scientific question was identified to serve as an alternative to meeting full target accrual in designating accrual success. Results: Of 238 eligible trials, 158 (66%) closed with sufficient accrual. Among 80 trials with insufficient accrual, 70 (29%) closed specifically because of poor accrual. Inadequate accrual rates are overemphasized when defining accrual success solely by meeting accrual targets. Nearly 75% of trials conclusively addressed the primary endpoints with positive results in 39% of trials. Exceeding 80% of target accrual serves as a reliable proxy for answering the intended scientific question. Conclusions: Approximately one third of phase III trials closed with insufficient accrual to address the primary endpoints, primarily due to poor accrual. Defining accrual sufficiency broader than meeting accrual targets represents a fairer account of trial closures. Clin Cancer Res; 18(1); 256–62. ©2011 AACR.

PurposeProgression-free survival (PFS) is not an optimal end point to evaluate therapeutic strategies in advanced colorectal cancer (ACRC). Therefore, composite end points have been proposed to evaluate a chemotherapy strategy when sequential treatments are available: duration of disease control (DDC) and time to failure of strategy (TFS). The goal of this study was to evaluate these alternative end points and their potential surrogacy for overall survival (OS).MethodsWe pooled individual patient data from three randomized trials evaluating chemotherapy strategy, which accrued 1,042 patients with previously untreated ACRC. In these trials, first-line treatment was either oxaliplatin- or irinotecan-based chemotherapy. Compared with TFS, DDC included neither time interval between progression and next sequence of treatment nor time to progression if the best result of the next sequence of treatment was progression.ResultsThere was good correlation between DDC and OS (correlation of median: r, 0.62; correlation of hazard ratio [HR]: adjusted copula R2, 0.72) and between TFS and OS (correlation of median: r, 0.59; correlation of HR: adjusted copula R2, 0.67). There was no correlation between PFS and OS (correlation of median: r, 0.45; correlation of HR: adjusted copula R2, 0.47).ConclusionDDC and TFS roughly achieved the same results. Both are acceptable new end points to evaluate a therapeutic strategy in ACRC. Although TFS achieved a pragmatic evaluation of a multiline strategy, DDC captured the effect of a specific sequence in a therapeutic strategy.

Background Intermediate outcome variables can often be used as auxiliary variables for the true outcome of interest in randomized clinical trials. For many cancers, time to recurrence is an informative marker in predicting a patient's overall survival outcome and could provide auxiliary information for the analysis of survival times. Purpose To investigate whether models linking recurrence and death combined with a multiple imputation procedure for censored observations can result in efficiency gains in the estimation of treatment effects and be used to shorten trial lengths. Methods Recurrence and death times are modeled using data from 12 trials in colorectal cancer. Multiple imputation is used as a strategy for handling missing values arising from censoring. The imputation procedure uses a cure model for time to recurrence and a time-dependent Weibull proportional hazards model for time to death. Recurrence times are imputed, and then death times are imputed conditionally on recurrence times. To illustrate these methods, trials are artificially censored 2 years after the last accrual, the imputation procedure implemented, and a log-rank test and Cox model used to analyze and compare these new data with the original data. Results The results show modest, but consistent gains in efficiency in the analysis using the auxiliary information in recurrence times. Comparison of analyses show the treatment effect estimates and log-rank test results from the 2-year censored imputed data to be in between the estimates from the original data and the artificially censored data, indicating that the procedure was able to recover some of the lost information due to censoring. Limitations The models used are all fully parametric, requiring distributional assumptions of the data. Conclusions The proposed models may be useful in improving the efficiency in estimation of treatment effects in cancer trials and shortening trial length.

Abstract Purpose: The categorical definition of response assessed via the Response Evaluation Criteria in Solid Tumors has documented limitations. We sought to identify alternative metrics for tumor response that improve prediction of overall survival. Experimental Design: Individual patient data from three North Central Cancer Treatment Group trials (N0026, n = 117; N9741, n = 1,109; and N9841, n = 332) were used. Continuous metrics of tumor size based on longitudinal tumor measurements were considered in addition to a trichotomized response [TriTR: response (complete or partial) vs. stable disease vs. progression). Cox proportional hazards models, adjusted for treatment arm and baseline tumor burden, were used to assess the impact of the metrics on subsequent overall survival, using a landmark analysis approach at 12, 16, and 24 weeks postbaseline. Model discrimination was evaluated by the concordance (c) index. Results: The overall best response rates for the three trials were 26%, 45%, and 25%, respectively. Although nearly all metrics were statistically significantly associated with overall survival at the different landmark time points, the concordance indices (c-index) for the traditional response metrics ranged from 0.59 to 0.65; for the continuous metrics from 0.60 to 0.66; and for the TriTR metrics from 0.64 to 0.69. The c-indices for TriTR at 12 weeks were comparable with those at 16 and 24 weeks. Conclusions: Continuous tumor measurement–based metrics provided no predictive improvement over traditional response–based metrics or TriTR; TriTR had better predictive ability than best TriTR or confirmed response. If confirmed, TriTR represents a promising endpoint for future phase II trials. Clin Cancer Res; 17(20); 6592–9. ©2011 AACR.

Background Research on barriers to accrual has typically emphasized factors influencing participation after trial activation. Purpose We sought to identify factors influencing trial design and accrual predictions prior to trial activation associated with sufficient accrual. Methods A 30-question web-based survey was sent to the study chair and lead statistician for all 248 phase III trials open in 1993–2002 by five Clinical Trials Cooperative Groups. Questions addressed prior trial experience, trial design elements, accrual predictions, and perceived accrual influences. Accrual sufficiency categorization was derived from Clinical Trials Cooperative Group records: sufficient accrual included trials closed with complete accrual or at interim analysis, insufficient accrual included trials closed with inadequate accrual. Responses were analyzed by respondent role (study chair/lead statistician) and accrual sufficiency. Results Three hundred and nine eligible responses were included (response rate, 63%; lead statisticians, 81%; and study chairs, 45%), representing trials with sufficient (63%) and insufficient accruals (37%). Study chair seniority or lead statistician experience was not linked to accrual sufficiency. Literature review, study chair's personal experience, and expert opinion within Clinical Trials Cooperative Group most commonly influenced control arm selection. Clinical Trials Cooperative Group experience most influenced accrual predictions. These influences were not associated with accrual sufficiency. Among respondents citing accrual difficulties (41%), factors negatively influencing accrual were not consistently identified. Respondents credited three factors with positively influencing accrual: clinical relevance of study, lack of competing trials, and protocol paralleling normal practice. Limitations Perceptions of lead statisticians and study chairs may not accurately reflect accrual barriers encountered by participating physicians or patients. Survey responses may be subject to recall bias. Conclusion Consistent factors explaining poor accrual were not identified, suggesting reasons for poor accrual are not well understood and warrant further study. Alternate strategies for accrual prediction are needed since Clinical Trials Cooperative Group experience is linked to successful and unsuccessful accrual.

PurposeColorectal cancer predominantly occurs in the elderly, but approximately 5% of patients are 50 years old or younger. We sought to determine whether young age is prognostic, or whether it influences efficacy/toxicity of chemotherapy, in patients with advanced disease.MethodsWe analyzed individual data on 6,284 patients from nine phase III trials of advanced colorectal cancer (aCRC) that used fluorouracil-based single-agent and combination chemotherapy. End points included progression-free survival (PFS), overall survival (OS), response rate (RR), and grade 3 or worse adverse events. Stratified Cox and adjusted logistic-regression models were used to test for age effects and age-treatment interactions.ResultsA total of 793 patients (13%) were younger than 50 years old; 188 of these patients (3% of total patients) were younger than 40 years old. Grade 3 or worse nausea (10% v 7%; P = .01) was more common, and severe diarrhea (11% v 14%; P = .001) and neutropenia (23% v 26%; P < .001) were less common in young (younger than 50 years) than in older (older than 50 years) patients. Age was prognostic for PFS, with poorer outcomes occurring in those younger than 50 years (median, 6.0 v 7.5 months; hazard ratio, 1.10; P = .02), but it did not affect RR or OS. In the subset of monotherapy versus combination chemotherapy trials, the relative benefits of multiagent chemotherapy were similar for young and older patients. Results were comparable when utilizing an age cut point of 40 years.ConclusionYoung age is modestly associated with poorer PFS but not OS or RR in treated patients with aCRC, and young patients have more nausea but less diarrhea and neutropenia with chemotherapy in general. Young versus older patients derive the same benefits from combination chemotherapy. Absent results of a clinical trial, standard combination chemotherapy approaches are appropriate for young patients with aCRC.

Abstract Learning Objectives After completing this course, the reader will be able to: Explain the difference in survival prediction between response criteria (WHO) when used as a two-level variable (CR/PR vs. other) and as a three-level variable (CR/PR vs. SD vs. PD).Describe the limited benefit of using actual tumor measurements over traditional criteria (as a three-level variable) in predicting survival in colorectal cancer. This article is available for continuing medical education credit at CME.TheOncologist.com Purpose. We explored the prognostic value of actual tumor measurements (TM) versus World Health Organization (WHO) criteria as three-level (responder, stable, and progression) and two-level (responder and non-responder) variables at 12 and 24 weeks as predictors of survival in Intergroup Trial N9741, a phase III trial in metastatic colorectal cancer (CRC). Methods. All patients with measurable disease (N = 1,188) were included. The percentage changes in TM from baseline to 12 and 24 weeks were calculated. The prognostic values of TM versus WHO criteria (as three- and two-level variables) at 12 and 24 weeks were compared, using Cox models for overall survival (OS) in a landmark analysis, adjusting for baseline tumor size, performance status, and treatment arm. Results. Tumor status at 12 weeks by WHO criteria (three or two levels) or actual TM were all strongly associated with OS. Actual TM provided no meaningful additional benefit compared with the three-level WHO criteria. Tumor status at 24 weeks was also strongly associated with survival, but added no additional prognostic value compared with the 12-week assessment. At 12 weeks, actual TM improved prognostic characterization of patients with WHO status of response, but provided no additional value in patients with stable disease or progression. Conclusions. In N9741, the use of actual TM, or following tumor status beyond 12 weeks, did not improve survival prediction compared with a single three-level response assessment at 12 weeks, suggesting that 12-week tumor status could be an appropriate phase II trial endpoint in metastatic CRC.

The majority of patients with gastrointestinal cancers are over the age of 65. This age group comprises the minority of the patients enrolled in clinical trials, and it is unknown whether older patients achieve similar results as younger patients in terms of survival benefit and tolerability. In addition, there are few studies specifically designed for patients over 65 years. Subset analyses of individual trials and studies using pooled patient data from multiple trials provide some understanding on outcomes in older patients with gastrointestinal cancers. This article reviews the evidence on chemotherapeutic regimens in the elderly with colorectal, pancreatic, and gastroesophageal cancers, and discusses a practical approach to provide the best outcomes for older patients.