Purpose: The purpose of this study was to investigate the influence of lung tumor motion on the accuracy of dose delivery in both conventional and hypo fractionated regimens using VMAT delivery.
Methods: A custom tumor model 28 cc in volume with relative electron density of 0.97 was built and embedded in an anthromorphic lung phantom. The phantom was placed in a motion platform and 4D-CT was acquired to derive ITV for target motion ranging from 5mm to 20mm in 5mm increments. Three VMAT plans of incrementally increasing plan complexity were generated for both 2 Gy and 10 Gy prescription dose and plans were evaluated using RTOG-0813 criteria for low and high dose spillage. All plans were generated with the Monaco Treatment Planning System using Montecarlo dose calculation algorithm for a Elekta Versa Linac with Agility 160 leaf MLC. The actual dose delivered in a dynamic lung phantom was measured using EBT-3(2Gy) and EBT-XD (10Gy) films placed in the tumor fully encompassed in lung and compared against the dose received for each motion specific ITV and PTV calculated by TPS. Gamma analysis with a 3% dose difference and 2-or 3 mm distance to agreement was used to evaluate the accuracy of delivery.
Results: Gamma pass rates at 3%/3mm criteria were â‰¥ 95% for all plans up to 10mm target motion. The pass rates using 3% 2mm criteria were â‰¥ 95% only for simple modulation at 10mm tumor motion and this analysis criteria had greater sensitivity to plan modulation.
Conclusion: There is a threshold limit (10mm) for tumor motion beyond which dosimetric differences due to interplay effects become significant. The threshold limit is further dependent on the complexity of the treatment plan delivered. The accuracy of dose delivery was worse for conventionally fractionated dose delivery as the target motion increased beyond 10mm.