Gamma scintillation cameras for diagnostic imaging have been in use for many years, and the criteria for selecting a radionuclide are well documented. Similar criteria also pertain to the early use of a handheld probe for external scanning of deep lesions. For intraoperative detection of cancer, however, different considerations apply. The probe is not needed for visible or palpable tumors. The objective is to locate hidden tumors. We have therefore conducted studies both clinically and in the laboratory to determine the properties of a radiolabel that would optimize the Radioimmunoguided Surgery(TM) system (RIGS(R) ). The laboratory investigations included phantom studies in an environment simulating that present during surgery. Of the nuclides studied, I-125, I-131, In-111, and Tc-99m were emphasized since these are commonly available in hospitals. Phantom data were obtained for point sources, spheres and cylinders. Each of these sources was tested in air and tissue (water) to separate the effects of tissue attenuation and scattering from inverse square law effects. The major factors one must consider when choosing a radionuclide are the half-life and the photon energy levels. The energy has a direct effect upon the detector efficiency, the magnitude of Compton scattering, and the attenuation of emitted photons by tissue. Detection of the smallest possible lesions requires minimizing tissue background as well as providing a statistically acceptable tumor to background ratio. A useful way to reduce background count rates is to use a low energy nuclide to increase tissue attenuation. The physical half-life of the nuclide must be several times longer than the biological clearance half-life of the carrier substance. These studies and extensive clinical practice indicate that I-125 is the most effective of the commonly available radionuclides in this application.