PURPOSE: 3-D conformal radiation therapy (3DCRT) holds promise in allowing safe escalation of radiation dose to increase the local control of prostate cancer. Prospective evaluation of this new modality requires strict quality assurance (QA). We report the results of QA review on patients receiving 3DCRT for prostate cancer on a cooperative group trial.MATERIALS [amp ] METHODS: In 1993 the NCI awarded the ACR/RTOG and nine institutions an RFA grant to study the use of 3DCRT in the treatment of prostate cancer. A phase I/II trial was developed to: a) test the feasibility of conducting 3DCRT radiation dose escalation in a cooperative group setting; b) establish the maximum tolerated radiation dose that can be delivered to the prostate; and c) quantify the normal tissue toxicity rate when using 3DCRT. In order to assure protocol compliance each participating institution was required to implement data exchange capabilities with the RTOG 3D QA center. The QA center reviews at a minimum the first five case from each participating center and spot checks subsequent submissions. For each case review the following parameters are evaluated: 1) target volume delineation, 2) normal structure delineation, 3) CT data quality, 4) field placement, 5) field shaping, and 6) dose distribution.RESULTS: Since the first patient was registered on August 23, 1994, an additional 170 patients have been accrued. Each of the nine original approved institutions has participated and three other centers have recently passed quality assurance bench marks for study participation.Eighty patients have been treated at the first dose level (68.4 Gy minimum PTV dose) and accrual is currently ongoing at the second dose level (73.8 Gy minimum PTV dose). Of the 124 cases that have undergone complete or partial QA review, 30 cases (24%) have had some problems with data exchange. Five of 67 CT scans were not acquired by protocol standards. Target volume delineation required the submitting institution's correction and resubmission in 7 of 67 (10.4%) reviewed cases. Normal tissues required correction in 6 of 67 (8.9%) of cases. Initial field shaping differed from the submitted treatment plan by more than 5 mm in significant regions of the field in only 2% of the cases. Isocenter shifts of more than 5 mm on at least one of the treated fields was identified in 7% of initial port films examined. Dosimetry review has demonstrated that 14 of 86 cases (16.3%) had minor variations in target volume coverage ([lt ]100% of the target volume coverage by the prescription isodose) and 3.4% had major variation in dose coverage ([lt ]95% coverage of target volume by prescription isodose). Nineteen of 93 cases (20%) had more than 7% heterogeneity of dose within the planning target volume.CONCLUSION: 3DCRT can be studied and implemented in a cooperative group setting. Although data exchange problems in this study have been frequent, most of these problems occurred early in the trial and have been resolved in most circumstances. A significant amount of variation has been identified in the definition of target volumes and organs at risk. Similarly, field shaping and port film evaluation showed occasional errors. It is our impression that quality assurance is a critical component of 3DCRT in the cooperative group setting. As experience in the planning of patients with 3DCRT increases, it is expected that the frequency of planning variations will diminish.