Session Item

Abdominal compression greatly mitigates the intrafraction motion of abdominal tumors and their surroundings during MR-guided SBRT (MRgSBRT) delivery. Nevertheless, residual motion can be of concern. The purpose of this study was to quantitatively assess the dosimetric impact of this residual motion for both the target and abutting organs at risk as part of our QA protocol.

Twenty patients were included that underwent online adaptive MRgSBRT for upper abdominal malignancies in five fractions of 8 Gy on a 1.5 T MR-linac. Patients were wearing a custom fitted polyurethane corset during treatment delivery for abdominal compression. During the entire beam-on time, interleaved coronal and sagittal cine MRIs were acquired at 2.8 Hz, from which the GTV motion was extracted with deformable image registration. This motion timeline was synchronized to the linac log files to translate the individual beams at each point in time during treatment delivery, which were summed to obtain the delivered dose (see fig. 1). Key DVH parameters of the GTV and neighboring OARs (duodenum, bowel structures and stomach) were compared between the planned dose and the delivered dose.

The key DVH parameters from the delivered dose relative to the planned dose are shown in fig. 2. Each point in the scatterplot represents a fraction, characterized by a cranio-caudal (CC) peak-to-peak respiratory amplitude and the maximum CC drift as measured during treatment.

In more than 80% of the fractions, the measured D_99% of the GTV coverages was more than 95% of the planned D_99%. In only two fractions the actual D_99% of the GTV was lower than 90% of the planned D_99% due to a lateral 2.5 mm drift GTV drift during dose delivery towards areas with sharp dose gradients. The dose hotspots in the OARs remained below 110% of the planned D_0.5cc, as they were mostly resolved in the delivered dose maps due to the blurring effect of motion.

We have successfully implemented a workflow that retrospectively calculates the delivered dose for abdominal MRgSBRT treatments based on the intrafraction motion. Analyses revealed that for the vast majority of fractions/patients the intrafraction motion observed during treatment delivery only modestly impacted the dose to the target and organs at risk. This increased our confidence that MRgSBRT can be safely executed for patients with abdominal tumors, potentially allowing dose escalation strategies. Furthermore, this workflow helps us to identify patients that, based on the intrafraction motion, would benefit from a re-assessment of the dose planning objectives/constraints for the remaining fractions if tolerances tend to be become exceeded.