As exemplified in Figure 7, the linearized multistage models upper-bound q₁* potency is very nonresponsive to the information on the shape of the dose-response relationship in observed data.  The q₁* value is primarily determined by the highest dose level in the data and is not primarily determined by the observed dose-response pattern in the data.  However, the highest dose in experimental data is usually the maximum tolerated dose which is determined by some simple gross measure of toxicity and not cancer or the specified adverse health effect of regulatory concern.  Thus, the nonresponsiveness of the linearized multistage model to the observed dose-response data on cancer means that the q₁* value used to regulate cancer is not determined by the corresponding cancer dose-response data but rather is primarily determined by the more general measure of toxicity used to determine the maximum tolerated dose.  In Figure 7, five hypothetical experimental outcomes have very different dose-response relationships, yet the largest of the five corresponding linearized multistage model potency measures differs from the smallest potency measure by less than one order of magnitude (10 fold).

>> Figure 7