Table 5.

Human dose at various specified tumor incidences estimated by linear extrapolation and by the probabilistic approach based on treating tumors as deterministic quantal versus stochastic quantal effects for the example tumor dataset.

Population tumor incidence (tumor dataset)	Linear extrapolation from allometrically scaled BMDLa (mg/kg/day)	Human dose assuming deterministic quantal effect [5th and 95th percentiles (mg/kg/day)]	Human dose assuming stochastic quantal effect [5th and 95th percentiles) (mg/kg/day)]
5%	0.11	(0.029, 0.67)	(0.020, 0.37)
1%	0.022	(0.011, 0.47)	(0.0062, 0.17)
0.1%	0.0022	(0.0034, 0.33)	(0.0012, 0.078)
0.01%	0.00022	(0.0013, 0.25)	(0.00018, 0.040)
aBased on U.S. EPA (2005) default approach, where the point of departure is the lower (one-sided) 95% confidence limit on the benchmark dose at a 10% extra risk, scaled to a human equivalent by multiplying by (BWanimal/BWhuman)0.25.