Population and patient-specific target margins for 4D adaptive radiotherapy to account for intra- and inter-fraction variation in lung tumour position

In this work, five 4D image-guidance strategies (two population, an offline adaptive and two online strategies) were evaluated that compensated for both inter- and intra-fraction variability such as changes to the baseline tumour position and respiratory pattern. None of the strategies required active motion compensation such as gating or tracking; all strategies simulated a free-breathing-based treatment technique. Online kilovoltage fluoroscopy was acquired for eight patients with lung tumours, and used to construct inter- and intra-fraction tumour position variability models. Planning was performed on a mid-ventilation image acquired from a respiration-correlated CT scan. The blurring effect of tumour position variability was included in the dose calculation by convolution. CTV to PTV margins were calculated for variability in the cranio-caudal direction. A population margin of 9.0 +/- 0.7 mm was required to account for setup error and respiration in the study population without the use of image-guidance. The greatest mean margin reduction was introduced by the offline adaptive strategy. A daily online correction strategy produced a small reduction (1.6 mm) in the mean margin from the offline strategy. Adaptively correcting for an inter-fraction change in the respiratory pattern had little effect on margin size due to most patients having only small daily changes in the respiratory pattern. A daily online correction strategy would be useful for patients who exhibit large variations in the daily mean tumour position, while an offline adaptive strategy is more applicable to patients with less variation.