Many radiopharmaceuticals are excreted from the body through the gastrointestinal (GI) tract. The doses to the walls of the organs involved often are very significant. As significant fractions of the administered activity pass through them, these organs may receive the highest doses in the body for many radiopharmaceuticals. The absorbed dose to these walled organs, from activity in their contents, is typically calculated as 50% of the average absorbed dose to the contents, for nonpenetrating emissions. The internal surface of the GI tract, and to a certain extent the urinary bladder, is lined with a variable thickness of mucus. In addition, the radiosensitive cell populations (crypt or stem cells) are located at some depth into the mucosa. These two factors suggest that the surface dose, often used to characterize the clinically relevant absorbed doses for walled organs, may represent an overestimate in some cases. METHODS In this study, the radiation transport code MCNP was used to simulate the deposition of energy from nonpenetrating emissions of several radionuclides of interest: 90Y, 99mTc,123I and 131I. Absorbed doses as a function of distance from the wall-contents interface were calculated for three geometric shapes representing different organs along the routes of excretion. RESULTS The absorbed dose from nonpenetrating emissions to the sensitive cell populations was consistently lower than estimated by the standard model assumption. The simulated absorbed doses to radiosensitive cells in the GI tract for 99mTc and 123I are tenfold lower; those for 131I are fivefold lower and those for 90Y are 20% lower. CONCLUSIONS This study demonstrates that the normally reported dose to the walls of hollow organs probably should be modified to account for the attenuation of these nonpenetrating emissions in the linings of the walls. This study also demonstrates that Monte Carlo codes continue to be useful in the evaluation of the dose to sensitive cells in walled organs.