Separating the dosimetric consequences of changing tumor anatomy from positional uncertainty for conventionally fractionated lung cancer patients

Purpose: To separate the dosimetric consequences of changing tumor volume from positional uncertainty for patients undergoing conventionally fractionated lung radiation therapy (RT) and to quantify which factor has a larger impact on dose to target volumes and organs at risk (OAR). Methods and materials: Clinical treatment plans from 20 patients who had received conventionally fractionated RT were retrospectively altered by replacing tumor and atelectasis with lung equivalent tissue in the treatment planning system calculations. To simulate positional uncertainty, the isocenter was shifted in both the altered and original plans by 2 and 5 mm in 6 directions. Rotational uncertainty was introduced by rotating each computed tomographic image set by ±. 3 degrees about a superior-inferior axis extending through patient center. Additionally, after rotation the isocenter was translated back to its original point within the patient to evaluate whether purely translational corrections could minimize dosimetric consequences due to rotations. Results: Dosimetric statistics for each altered plan were compared with the original. Average changes in the planning target volume (PTV) receiving 95% of prescription dose (PTV V95%) resulting from changing tumor anatomy alone were approximately 0.1%. Average changes in PTV V95% resulting from positional uncertainty were greater (0.2%-4.2%) but were largely independent of whether or not the original tumor volume was present. For 3 patients, increases in volumes receiving 110% of the prescription dose were seen but were largely limited to within the PTV. Translational corrections for patient rotations were effective in minimizing differences in target coverage but had less effect on reducing the maximum spinal cord dose. Conclusions: Anatomic changes alone, such as reductions in tumor volume and atelectasis, had minimal effect on the overall dose distribution. Greater dosimetric consequences were seen with positional uncertainty. With accurate patient localization, replanning during the course of treatment for conventionally fractionated lung cancer patients may not be necessary.