Radionuclide therapy extends the usefulness of radiation from localized disease of multifocal disease by combining radionuclides with disease-seeking drugs, such as antibodies or custom-designed synthetic agents. Like conventional radiotherapy, the effectiveness of targeted radionuclides is ultimately limited by the amount of undesired radiation given to a critical, dose-limiting normal tissue, most often the bone marrow. Because radionuclide therapy relies on biological delivery of radiation, its optimization and characterization are necessarily different than for conventional radiation therapy. However, the principals of radiobiology and of absorbed radiation dose remain important for predicting radiation effects. Fortunately, most radionuclides emit gamma rays that allow the measurement of isotope concentrations in both tumor and normal tissues in the body. By administering a small "test dose" of the intended therapeutic drug, the clinician can predict the radiation dose distribution in the patient. This can serve as a basis to predict therapy effectiveness, optimize drug selection, and select the appropriate drug dose, in order to provide the safest, most effective treatment for each patient. Although treatment planning for individual patients based upon tracer radiation dosimetry is an attractive concept and opportunity, practical considerations may dictate simpler solutions under some circumstances. There is agreement that radiation dosimetry (radiation absorbed dose distribution, cGy) should be utilized to establish the safety of a specific radionuclide drug during drug development, but it is less generally accepted that absorbed radiation dose should be used to determine the dose of radionuclide (radioactivity, GBq) to be administered to a specific patient (i.e., radiation dose-based therapy). However, radiation dosimetry can always be utilized as a tool for developing drugs, assessing clinical results, and establishing the safety of a specific radionuclide drug. Bone marrow dosimetry continues to be a "work in progress." Blood-derived and/or body-derived marrow dosimetry may be acceptable under specific conditions but clearly do not account for marrow and skeletal targeting of radionuclide. Marrow dosimetry can be expected to improve significantly but no method for marrow dosimetry seems likely to account for decreased bone marrow reserve.