Growth Modulation Index의 추정에 대한 연구: 잠재적 변수와 가속실패시간 모형의 적용
Estimation of Growth Modulation Index Using Latent Variable and Accelerated Failure Time Model
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Objectives: For cytostatic cancer trials, Growth Modulation Index (GMI) defined by an intrapatient progression-free survival (PFS) ratio, has been proposed to evaluate the efficacy of new target agent. The purpose of this study was to suggest new methods for the estimation of GMI with censored data in the first PFS (PFS1) interval, and subsequent second PFS (PFS2) interval. Methods: The proposed methods include latent variable approach based on Rank Preserving Structural Failure Time (RPSFT) model and Accelerated Failure Time (AFT) model. Simulations were conducted to compare the performance of proposed GMI estimates and estimates based on the Kaplan-Meier method in terms of bias and mean squared error (MSE) by varying dependency of two PFS and censoring rates. Results: Simulation results show that new GMI estimates using latent variable approach and AFT model exhibited smaller bias and MSE than the previous estimates based on the Kaplan-Meier survival function. As censoring rates increased in PFS1, bias and MSE increased in the previous GMI estimates. When the AFT model was applied in the case of high censoring rates, bias was relatively higher than those of latent variable approach. Conclusions: When using GMI as primary endpoint in cancer clinical trials, cautious statistical application and interpretation is needed, particularly for the presence of censored data in the first PFS interval.
Abstract
Objectives: For cytostatic cancer trials, Growth Modulation Index (GMI) defined by an intrapatient progression-free survival (PFS) ratio, has been proposed to evaluate the efficacy of new target agent. The purpose of this study was to suggest new methods for the estimation of GMI with censored data in the first PFS (PFS1) interval, and subsequent second PFS (PFS2) interval. Methods: The proposed methods include latent variable approach based on Rank Preserving Structural Failure Time (RPSFT) model and Accelerated Failure Time (AFT) model. Simulations were conducted to compare the performance of proposed GMI estimates and estimates based on the Kaplan-Meier method in terms of bias and mean squared error (MSE) by varying dependency of two PFS and censoring rates. Results: Simulation results show that new GMI estimates using latent variable approach and AFT model exhibited smaller bias and MSE than the previous estimates based on the Kaplan-Meier survival function. As censoring rates increased in PFS1, bias and MSE increased in the previous GMI estimates. When the AFT model was applied in the case of high censoring rates, bias was relatively higher than those of latent variable approach. Conclusions: When using GMI as primary endpoint in cancer clinical trials, cautious statistical application and interpretation is needed, particularly for the presence of censored data in the first PFS interval.
참고문헌 (15)
- Mick R, "Phase Ⅱ clinical trial design for noncytotoxic anticancer agents for which time to disease progression is the primary endpoint", Control Clin Trials, 21(4), pp.343~359 (2000)
- Kovalchik S, "Statistical methods for a phase Ⅱ oncology trial with a growth modulation index (GMI) endpoint", Contemp Clin Trials, 32(1), pp.99~107 (2011)
- Hoff V, "There are no bad anticancer agents, only bad clinical trial designs-twenty-first Richard and Hinda Rosenthal Foundation Award Lecture", Clin Cancer Res, 4(5), pp.1079~1086 (1998)
- Owen WJ, "A paired-data analysis for a lifetime distribution", Am Stat, 54(4), pp.252~256 (2000)
- Robins JM, "Correcting for non-compliance in randomized trials using rank preserving structural failure time models", Commun Statis Theory Meth, 23(8), pp.2609~2631 (1991)
- Jonsson L, "Analyzing overall survival in randomized controlled trials with crossover and implications for economic evaluation", Value Health, 17(6), pp.707~713 (2014)
- Wei LJ, "The Accelerated failure time model : a useful alternative to the cox regression model in survival analysis", Stat Med, 11(14-15), pp.1871~1879 (1992)
- Klein JP, "Survival analysis: Techniques for censored and truncated data" (2003), Springer-Verlag
- Cook RJ, "The statistical analysis of recurrent events" (2003), Springer-Verlag
- Owen WJ, "A power analysis of tests for paired lifetime data", Lifetime Data Anal, 11(2), pp.233~243 (2005)
- Penel N, "Growth modulation index as metric of clinical benefit assessment among advanced soft tissue sarcoma patients receiving trabectedin as a salvage therapy", Ann Oncol, 24(2), pp.537~542 (2013)
- Cousin S, "Correlation between overall survival and growth modulation index in pre-treated sarcoma patients: a study from the french sarcoma group", Ann Oncol, 24(10), pp.2681~2685 (2013)
- Cantor AB, "A test of the equality of survival distributions based on paired observations from conditionally independent exponential distributions", IEEE Transactions Reliability R, 34(4), pp.342~346 (1985)
- Kleinbaum DG, "Survival analysis: a self-learning text" (2005), Springer-Verlag
- Kim YJ, "Survival analysis" (2013), Free Academy