临床前啮齿类动物致癌性实验常用统计方法简介

药物致癌性实验的目的是通过考察药物在动物体内的潜在致癌作用，从而评价和预测其可能对人体造成的危害，是药物临床前安全性评价的重要组成部分。致癌性实验肿瘤发生率及死亡率数据的统计分析是致癌实验结果解释的关键部分，目前用于评估致癌实验数据的统计方法很多，但还没有哪一种统计方法得到该领域科研人员的公认。鉴于上述情况，本文介绍了目前国际常用致癌数据的统计存在的问题，美国毒性病理学会及统计学家建议的致癌数据统计方法，例如Peto检验、Poly-3检验及Poly-k检验，以及美国毒性病理学会用于Peto检验的肿瘤准确性、一致性分类建议等内容，以期为我国临床前药物致癌性实验的数据统计提供参考。

The objective of drug carcinogenicity studies is to identify the potential tumorigenic effects of drugs in animals and to assess and predict the potential hazard to human body.Carcinogenicity studies are important part of the preclinical safety evaluation of drugs.The statistical analyses of the neoplasm incidence and motality rate are key parts for interpretation of the results of carcinogenicity studies.So far,many statistical methods have been developed and available for analysis of the preclinical rodent carcinogenicity studies;however,there has not been consensus on what constitutes the ‘optimal’ approach to statistical analysis of tumor data from carcinogenesis bioassays,and no single approach has received broad support from the scientific community.In this paper,we made a brief introduction to the issues in the application of statistical methods in carcinogenicity studies and the statistical methods for analysis and interpretation of carcinogenicity data recommended by Society of Toxicologic Pathology (STP) of USA and statisticians,such as Peto test,Poly-3 test and Poly-k test,as well as recommendations for accurately and consistently classifying neoplasms,so as to provide reference for the data statistics of preclinical carcinogenicity studies.

[1] Bickis M,Krewski D.Statistical issues in the analysis of the long-term carcinogenicity bioassay in small rodents:an empirical evaluation of statistical decision rules[J].Fund Appl Toxicol, 1989,12:202
[2] Ten Berge WF.Kaplan-Meier tumor probability as a starting point for dose-response modeling provides accurate lifetime risk estimates from rodent carcinogenicity studies[J].Ann N Y Acad Sci, 1999,895:112
[3] Morton D.The society of toxicological pathology's recommendations on statistical analysis of rodent carcinogenicity studies[J].Toxicol Pathol,2002,30(3):415
[4] United States Food and Drug Administration.Draft guidance for industry:Statistical aspects of the design,analysis,and interpretation of chronic rodent carcinogenicity studies of pharmaceuticals[EB/OL].[2001-4-23].http://www.fda.gov/cder/guidance/index.htm
[5] Morton D,Lee PN,Fry JS,et al.Statistical methods for carcinogenicity studies[J].Toxicol Pathol,2002,30(3):403
[6] Peto R,Pike MC,Day NE,et al.Guidelines for simple sensitive significance tests for carcinogenic effects in long-term animal experiments//IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans,Suppl.2.Long-term and Short-term Screening Assays for Carcinogens:A Critical Appraisal[S].Lyon:International Agency for Research on Cancer,1980:311
[7] Lagakos SW,Ryan LM.On the representativeness assumption in prevalence tests of carcinogenicity[J].App Stat,1985,34:54
[8] Peto Analysis Working Group of STP.Draft recommendations on classification of rodent neoplasms for Peto analysis[J].Toxicol Pathol,2001,29(2):265
[9] Cochran W.Some methods for strengthening the common x2 tests[J].Biometrics,1954,10:417
[10] Armitage P.Tests for linear trends in proportions and frequencies[J].Biometrics,1955,11:375
[11] Bailer A,and C Portier.Effects of treatment-induced mortality on tests for carcinogenicity in small samples[J].Biometrics, 1988,44:417
[12] Dinse GE.A comparison of tumor incidence analyses applicable in single-sacrifice animal experiments[J]. Stat Med,1994,13:689
[13] Bieler GS,Williams RL.Ratio estimates,the Delta method,and quantal response tests for increased carcinogenicity[J].Biometrics,1993,49:793
[14] Elmore SA,Peddada SD.Points to consider on the statistical analysis of rodent cancer bioassay data when incorporating historical control data[J].Toxicol Pathol,2009,37(5):672
[15] Stammberger I,Essermeant L.Insulin glargine:a reevaluation of rodent carcinogenicity findings[J].Int J Toxicol,2012,31(2):137