The goal of radiation therapy is to yield the greatest possible uncomplicated local and regional tumor control. An acceptable quality of life with few severe complications, anatomical defects or psychological disturbances should be an integral objective of excellent radiation therapy. Inasmuch as higher doses of irradiation and adequacy of treatment portals (volume irradiated) appear to correlate with greater probability of tumor control and major complications, there is a critical need to optimize treatment planning and quality assurance in radiation therapy. Along with this, techniques must be developed for reliable patient reposition and immobilization so that the optimized treatment plan can be translated into precise delivery of the irradiation. It is obvious that in clinical trials the basic parameters of therapy need to be optimally observed, in order to make comparison of experimental arms more reliable. It is easier to achieve uniformity of technical factors in a study at a single institution than in multi-institutional cooperative groups. Thus, stringent criteria must be incorporated into the protocol describing the requirements and techniques for planning and delivery of the radiation therapy. Furthermore, the dosimetry checks carried out by the Radiological Physics Center should be an integral part of this program. In some studies it is necessary to do an initial review of dosimetry factors and portal films. This alerts the radiation oncologist to variations from the protocol and, with prompt feedback and correction of the potential error, will ensure the complete evaluability of the case. Since an increasing number of reports point out that the doses of irradiation delivered and the volume treated may affect therapeutic results, a completed case review should always be carried out. Some cooperative group studies strongly suggest a correlation between the compliance with the protocol and the results of the trial. Costs of these programs in clinical trials should be justified since, for a small investment, the evaluability rate in a protocol can be increased by 10%. The increase of patient evaluability significantly decreases the cost per evaluable patient. Furthermore, the reliability of the data of the trial will be greatly enhanced by a strong quality assurance program.