Simultaneous analysis of different tumor types in a long-term carcinogenicity study with scheduled sacrifices.

A statistical analysis of a long-term carcinogenicity study with noninbred CF-1 mice exposed to DDT for different durations of time is presented. The experiment included a high proportion of animals with scheduled sacrifices and thus permitted the use of statistical models for tumor prevalence and lethality. Three types of tumors, lymphomas and lung and liver tumors, were considered. The analysis attempted to model the joint behavior of these tumors in relation to time and treatment group. An apparently negative association in the occurrence of lymphomas and liver tumors was confirmed through separate analyses by accounting effectively for intercurrent mortality.