Session Item

The  1.5T MR-linac (MRL) enables on-treatment imaging, providing superior soft tissue contrast than of CT, without ionising radiation. Daily MR imaging facilitates daily plan adaptation, which can be achieved using the adapt-to-position (ATP) or adapt-to-shape (ATS) workflow. ATP uses rigid registration to adjust the beam weight and shape, matching the position of the target whereas ATS performs daily replanning, using deformable registration or manual contouring. When compared to daily IGRT on a C-arm linac, adaptation  on the MRL provides the potential for improvement in dose coverage for patients with OARs which closely border target volumes (Dunlop et al, 2020). However, this can necessitate longer session times, limiting patient throughput. Investigations into the session time for these workflows found a median of 26 minutes for ATP and 42 minutes for ATS over multiple treatment sites (Bertelsen et al, 2019 ). In this work, we investigate the effect daily replanning has on dose accumulation for prostate cancer patients in the target volume, rectum and bladder.

The daily MRI and dose for 7 prostate patients treated on the MRL using the ATS workflow  were collected from the multi-centre MOMENTUM dataset. To compare ATS versus ATP workflows the following comparison was made: 1) for ATS, dose from each fraction was non-rigidly registered, using Niftyreg b-spline deformable registration algorithm, to fraction one by registering the daily images and applying the deformable vector field (DVF) to the associated dose. 2) to emulate an ATP workflow the planned dose was copied across all treatment images , taking a rigid match into account, assuming dose invariance. Dose was than transformed back to fraction one using the same DVFs to evaluate the cumulative dose on fraction one. The difference  in the mean  dose in the PTV, rectum, bladder and bowel  for the ATS and emulated ATP workflow was evaluated, highlighting the difference caused by the choice of workflow.

The cumulative difference between the mean dose of the ATS workflow and the emulated ATP workflow is shown in Table 1. The mean (SD) difference across all patients is 0.42 (0.53) Gy for the PTV, -0.22 (0.81) Gy for the bladder and -0.77 (2.81) Gy for the rectum, i.e. ATS has the largest influence on the rectum, with two patients showing more than 2Gy difference.

The differences in dose delivered to the PTV and two OARs between the ATS and emulated ATP workflow are small in most cases, showing limited benefit from the increased time of the fully adaptive workflow in prostate cancer. However, in some instances, the difference exceeds 2Gy, suggesting that the ATS workflow benefits some patients. Further work to develop decision making tools to identify patients that benefit from ATS may allow for more widespread use, by alleviating concerns about that extra time spent.