The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

Analysis of Tumorigenicity Data with Informative Censoring

종속적인 중도절단을 가진 동물종양 자료의 분석을 위한 모형

  • Kim, Jin-Heum (Department of Applied Statistics, University of Suwon) ;
  • Kim, Youn-Nam (Graduate School of Public Health, Yonsei University)
  • Received : 20100600
  • Accepted : 20100800
  • Published : 2010.10.31
Download PDF
⟨ Previous Next ⟩

Abstract

In animal tumorigenicity data, the occurrence time of tumor is not observed because the existence of a tumor is examined only at either time of natural death or time of sacrifice for the animal. A three-state model (Health-Tumor onset-Death) is widely used to model the incomplete data. In this paper, we employed a frailty effect into the three-state model to incorporate the dependency of death on tumor occurrence when the time of natural death works as an informative censoring against the tumor onset time. For the inference of parameters, then the EM algorithm is considered in order to deal with missing quantities of tumor onset time and random frailty. The proposed method is applied to the bladder tumor data taken from Lindsey and Ryan (1993, 1994) and a simulation study is performed to show the behavior of the proposed estimators.

동물종양 실험에서는 종양발생 시간이 직접 관찰되지 않고 단지 자연사로 인한 관찰 시점이나 강제적으로 희생시킨 시점 이전에 종양이 발생했는지 유무만을 알 수 있다. 이와 같은 형태의 결측을 가진 자료를 분석하기 위해 3단계(건강$\rightarrow$종양발생$\rightarrow$사망) 모형이 널리 사용되고 있다. 본 논문에서는 자연사로 인한 사망 시간이 종속적인 중도절단으로 작용하여 사망 시간과 종양발생 시간이 종속될 때, 이를 모형에 반영하기 위해 감마 프레일티 효과를 도입하였다. 모수 추정은 종양발생 시간과 프레일티 효과의 결측을 다루기 위해 EM 알고리즘 방법을 사용하였다. 제안한 추정량의 소표본 성질을 살펴보기 위해 제안한 방법을 Lindsey와 Ryan (1993, 1994)의 방광암 자료에 적용하여 모수를 추정하였으며, 그 추정값을 바탕으로 모의실험을 수행하였다.

Keywords

Acknowledgement

Supported by : 한국연구재단

References

  1. Dempster, A. P., Laird, N. M. and Rubin, D. B. (1977). Maximum likelihood for incomplete data via the EM algorithms (with discussion), Journal of the Royal Statistical Society, Series B, 39, 1-38.
  2. French, J. L. and Ibrahim, J. G. (2002). Bayesian methods for a three-state model for rodent carcinogenicity studies, Biometrics, 58, 906-916. https://doi.org/10.1111/j.0006-341X.2002.00906.x
  3. Golub, G. H. and Welsch, J. H. (1969). Calculation of Gauss quadrature rules, Mathematics of Computation, 23, 221-230. https://doi.org/10.2307/2004418
  4. Hastings, W. K. (1970). Monte Carlo sampling methods using Markov chains and their applications, Biometrika, 57, 97-109. https://doi.org/10.1093/biomet/57.1.97
  5. Huang, X. and Wolfe, R. A. (2002). A frailty model for informative censoring, Biometrics, 58, 510-520. https://doi.org/10.1111/j.0006-341X.2002.00510.x
  6. Kim, J., Kim, Y., Nam, C., Choi, E. and Kim, Y. J. (2010). A analysis of tumorigenicity data using a normal frailty effect, Proceedings for the Spring Conference, 2010, The Korean Statistical Society, 13.
  7. Lagakos, S. W. and Louis, T. A. (1988). Use of tumor lethality to interpret tumorigenicity experiments lacking cause-of-death data, Applied Statistics, 37, 169-179. https://doi.org/10.2307/2347336
  8. Lindsey, J. C. and Ryan, L. M. (1993). A three-state multiplicative model for rodent tumorigenicity experiments, Applied Statistics, 42, 283-300. https://doi.org/10.2307/2986233
  9. Lindsey, J. C. and Ryan, L. M. (1994). A comparison of continuous - and discrete time three state models for rodent tumorigenicity experiments, Environmental Health Perspective Supplements, 102, 9-17.
  10. Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H. and Teller, E. (1953). Equations of state calculations by fast computing machines, Journal of Chemical Physics, 21, 1087-1091. https://doi.org/10.1063/1.1699114
  11. Sun, J. (2006). The Statistical Analysis of Interval-Censored Failure Time Data, Springer, New York.