The therapeutic use of radioisotopes in medicine as unsealed sources has a long history dating back to the 1930s. The established and continuing objectives are to provide radiation dose to the target tissue at the desired cytotoxic level while avoiding or minimizing toxic effects. Selected radionuclide therapy protocols including 32P for polycythemia vera, 131I for Graves' disease, and 131I for postsurgical ablation of thyroid remnants in the management of differentiated thyroid cancer are presented for historical review with the focus on protocols for administering the radiopharmaceuticals and the role played by dosimetry. The discussion also includes consideration of complications and the assessment of outcome for these diseases. The vista for radionuclide therapy today is reviewed along with the options for determining the administered activity. Patient specific dosimetry encompasses a number of levels ranging from basic measurement of relevant biokinetic parameters and use of standard models to calculate (and extrapolate) radiation dose to sophisticated three-dimensional techniques employing fusion of physiologic and high-resolution anatomic images coupled with advanced 3-D voxel patient representation and Monte Carlo techniques for use in radiation dose calculation. The role of patient specific dosimetry in clinical trials (Phase I, II, III trials) along with its utility in treatment planning, follow-up evaluation, and elucidation of dose-response relationships is discussed. The challenge ahead for those who advocate patient specific dosimetry is to assemble the outcome data and perform the analysis to support this contention.