Interinstitutional variations in planning for stereotactic body radiation therapy for lung cancer

Purpose: The aim of this study was to assess interinstitutional variations in planning for stereotactic body radiation therapy (SBRT) for lung cancer before the start of the Japan Clinical Oncology Group (JCOG) 0403 trial. Methods and Materials: Eleven. institutions created virtual plans for four cases of solitary lung cancer. The created plans should satisfy the target definitions and the dose constraints for the JCOG 0403 protocol. Results: FOCUS/XiO (CMS) was used in six institutions, Eclipse (Varian) in 3, Cadplan (Varian) in one, and Pinnacle3 (Philips/ADAC) in one. Dose calculation algorithms of Clarkson with effective path length correction and superposition were used in FOCUS/XiO; pencil beam convolution with Batho power law correction was used in Eclipse and Cadplan; and collapsed cone convolution superposition was used in Pinnacle3. For the target volumes, the overall coefficient of variation was 16.6%, and the interinstitutional variations were not significant. For maximal dose, minimal dose, D95, and the homogeneity index of the planning target volume, the interinstitutional variations were significant. The dose calculation algorithm was a significant factor in these variations. No violation of the dose constraints for the protocol was observed. Conclusion: There can be notable interinstitutional variations in planning for SBRT, including both interobserver variations in the estimate of target volumes as well as dose calculation effects related to the use of different dose calculation algorithms. (C) 2007 Elsevier Inc.