Utilization of an MR-Linac adapted composite plan for dosimetric analysis in patients receiving MR-Linac guided SBRT

Purpose: To report our creation of an MR-Linac (MRL) adapted composite plan for patients treated with stereotactic body radiation therapy (SBRT) using online adaptation to shape (ATS) with live MRI-guidance during treatment on an Elekta Unity (Elekta AB, Stockholm, Sweden). Thus, we describe creation of our MRL-adapted composite with dosimetric analysis compared to the computed tomography (CT) reference plan.

Methodology: Patients were treated using a high-field 1.5T MRI. Patients underwent pre-treatment image acquisition consisting of CT simulation to create a CT reference plan and MRI simulation to create an MR Reference plan, by rigidly transferring contours from the planning CT. An IMRT treatment plan then was created. Prior to treatment, an online MRI image set was obtained for online adaptation to shape and optimization of planning, which was rigidly registered to the MR reference plan and plan was adapted daily. Target coverage and acceptable dose to organs at risk (OAR’s) were determined daily using the current MR image-set. Live MR-guidance was used the plan was adapted to shape (ATS) by a physicist, who optimized beam weight and aperture shape. Repeat adaptation to shape could be performed to account for inter-fraction changes in anatomy or target coverage. After completion of radiation therapy, an MRL-adapted composite plan was created through combination of daily treatment plans. Dosimetric analysis from the MRL-adapted composite plan was then compared to the CT reference plan to determine feasibility and accuracy.

Results: We identified 4 patients who received a total of 20 fractions SBRT. Average composite PTV volume was 45.1cc compared to 43.0cc on the reference plan. Ninety-five and ninety percent of the prescription dose covered respectively 90% and 95% of the treatment volume in our composite versus 96% and 98% in our reference plan. Mean maximum dose on the composite and reference plan was respectively 112.2% and 119.7% of prescription dose. Mean composite duodenal V35 was 0.69cc higher than the reference plan and V30 was 0.47cc higher than reference. Mean composite small bowel V35 was 0.83cc higher than reference and V30 was 2.51cc higher than reference. Composite kidney V12 was 1.8% lower than reference. No patients had a max spinal cord D1cc >20Gy on reference or composite.

Conclusions: Target volume coverage on our MRL-adapted composite plan was slightly lower, but similar to that of reference plan with decreased heterogeneity. Mean composite V35 and V30 of duodenum and small bowel were similar to that of the reference plan and all dose constraints to dynamic structures were met. Constraints for static structures were also met despite not optimizing the plan based on these structures. Creation of an MRL-adapted composite plan is a feasible and useful dosimetric application when delivering MRL-guided SBRT.

References: