ESTRO 2024 - Abstract Book

S4054

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2024

1736

Poster Discussion

Online dose adaptation reduces irradiated volume without compromising target coverage

Erik van der Bijl 1 , Robert Jan Smeenk 1 , Uulke A. van der Heide 2 , Tomas M. Janssen 2

1 Radboud University Medical Center, Radiation Oncology, Nijmegen, Netherlands. 2 The Netherlands Cancer Institute, Radiation Oncology, Amsterdam, Netherlands

Purpose/Objective:

In conventional online adaptive radiotherapy (OART) each fraction is planned on the anatomy of the day as if the whole treatment is given according to this anatomy. The dose delivery of previous fractions is not taken into account although this would be possible. The objective of this work is to investigate the feasibility and possible gains of significantly reducing the PTV margins in OART by taking the already given accumulated dose into account. To this end we consider OART with MRI guided radiotherapy to the whole prostate (5 x 7 Gy) as an example. For simplicity in the dose accumulation, we assume the prostate to be a rigid volume. The considered dose-adaptive OART approach consists of the first four fractions with 0mm PTV margins followed by a new plan in the last fraction tailored to compensate for any accumulated underdosage to the target. The rationale of this approach is that by updating the plan to accommodate additional information about the observed motion and resulting delivered dose might substantially reduce the dose to surrounding organs at risk, without compromising target coverage due to further personalisation. The clinical delineations and all daily MRI images of ten prostate cancer patients treated in 5 fractions within the Hypo Flame 2.0 trial on a 1.5T MR-Linac at a single institution were used. Each fraction, four MRI scans were taken: before irradiation for planning and position verification and during and at the end of beam delivery for motion monitoring. Using local rigid registration on the prostate we extracted the per fraction residual setup error and per treatment intrafraction blur. For the dose accumulation the planned dose was blurred to simulate the actual given dose using a gaussian with its width given by the root mean square of the standard deviations of the intrafraction motion. These blurred dose distributions were subsequently accumulated whilst taking into account the residual rigid registration vectors. We created initial treatment plans on the pre-treatment MRI using 0mm CTV-PTV margins to a prescription dose of 5x7Gy homogenous to the whole prostate gland (PTV D98%≥35Gy). Dose was accumulated for the first 4 fractions. Treatment plans for the last fraction were optimized using a 3mm margin on both the CTV and compensation regions. These regions were automatically segmented from the accumulated dose to construct regions that required a higher dose, in 2Gy increments, for the last fraction. To measure the necessary inhomogeneity for compensation CTV+3mm(D2) were extracted. Material/Methods: