Biostatistical methodology in carcinogenicity studies

This paper addresses the design, conduct, and statistical analysis of carcinogenicity studies, especially in the context of drug products for human use. It contains suggestions concerning the choice of dose levels, number of animals, methods of slide reading, and the ensuing statistical analysis, focusing on the significance testing approach. The purpose of this document is to describe the current thinking of statisticians and others who work in the area of carcinogenicity studies. The authors represent experience gained in the pharmaceutical industry, regulatory agencies, and academia. © 1998, Drug Information Association. All rights reserved.