Session Item

The aim of this work is to perform comparative treatment planning to investigate the potential dosimetric benefits of daily adaptive treatments, utilizing step-and-shoot MR-Linac (MRL) plans, over current standard non-adaptive treatment plans, conventional VMAT photons and scanned proton beams, for rectal cancer patients receiving a short-course of radiotherapy (5Gy x 5 fractions).

Anonymised data from twelve female rectal cancer patients treated at Uppsala University Hospital (UUH) with conventional VMAT photon plans were utilised. One reference plan per patient was created for proton and VMAT plans, while three reference plans were created for MRL each with a different PTV margin. The PTV A was a 7 mm isotropic margin coinciding with the current clinical practice for both MRL and conventional Linac treatments. The PTV B was 6 mm cranio-caudal, 6 mm anterior, 4 mm posterior, 4 mm left-right and the PTV C was 3 mm cranio-caudal, 3 mm anterior, 2 mm posterior, 2 mm left-right. The proton plans were created according to the PRORECT clinical study and the VMAT plans to the clinical practice at UUH. Moreover, for all proton and MRL plans the electron density of rectal gas was overridden to the value of the surrounding tissue.

To study the effect on the delivered dose by the daily anatomical changes, synthetic CTs were created from the daily CBCTs through image registration and post processing techniques. The proton and VMAT reference plans were then recalculated, whereas the MRL plans were adapted on the daily synthetic CTs. We evaluated how many fractions failed to fulfil the target and OARs constraints as well as the potential dosimetric benefits of the daily adaption. Dose was calculated as dose to water since it affected the evaluation of the dose to the sacrum bone. A qualitative evaluation of the daily anatomy variation based on the synthetic CTs was performed using the Dice similarity coefficient.

The analysis of the fraction plans showed that every MRL plans fulfilled the target and OAR constraints in contrast to the proton and VMAT plans that failed to fulfil the target constraints for some of the fractions. The VMAT fraction plans also failed to fulfil the constraints in the OAR in 60% of the cases. The dosimetric comparison of the three MRL plans to the VMAT plans showed that although there is a decrease in dose to the bowel bag for MRL PTV A, the doses to both the bladder and the sacrum were increased. A decrease in dose to the OARs is present in the reduced PTV margin plans, although the OARs in the proton plans received the lowest dose compared to the other two modalities. Future work will show if the results are sensitive to the synthetic CT quality and if so, how much.

This study shows that daily adaptive radiotherapy using the MRL delivers full target coverage and that OAR doses may be reduced by using reduced PTV margins.