In this paper, a new method of estimating tumorigenic potency is propo
sed that takes into account information on survival and, when availabl
e, the underlying cause of death. Specifically, Weibull distributions
are used to describe the time to tumor occurrence (X), the time to dea
th as a result of tumor occurrence (Y), and the time to death from cau
ses other than tumor occurrence (Z). The distributions of X and Y are
related via a lethality parameter rho lying between 0 (incidental tumo
rs) and 1 (rapidly fatal tumors). The index of tumorigenic potency is
defined as the dose inducing a specified excess tumor response rate (i
n the range of 25%-50%) following exposure for an extended period enco
mpassing most of the expected lifespan. Application of the proposed me
thod is illustrated using several data sets taken from the literature.