Session Item

MR-Linac (MRL) hybrid devices are a new development in radiation oncology. The unique workflow and specific time management represent a challenge for all involved professions (RTTs, physicists and physicians). Herein we present the data of our prospective workflow analysis for the first 80 patients treated at a high-field MRL.

Duration of all the daily working steps at the MRL was prospectively collected since the beginning of treatments in September 2018 for all patients by the responsible RTT. Indications for treatment at the MRL were: seven lymph node metastasis, 17 liver metastasis, seven head and neck cancers, eleven rectal cancers, three bladder cancers, 14 prostate cancers, three pancreatic cancers, eleven partial breast irradiation and four patients with various indications. Two adaptation methods are available at the MRL: the adapt-to-position (ATP) workflow consists of an isocenter shift depending on the image fusion offset for each segment followed by a segment weight optimization. The adapt-to-shape (ATS) workflow includes after image registration a contour propagation step in the workflow with the possibility to re-contour relevant structures on the daily MR followed by a constraint based re-optimization starting from the fluence map. Treatment times needed for each working step (cf. table 1) were analyzed independently for ATS- and ATP-based RT adaptation workflows. Differences between ATP and ATS were assessed with T-Test in Matlab ver. 2019a.

ATS was mainly used for prostate and bladder cancer and for RT of lymph node metastasis. Mean in-room time (range) was 32.1 min (22.8-44.8) and 45.0 min (34.7-62.6) for ATP and ATS (p<0.001), respectively. Mean beam on time was 5.3 min (2.3-15) and 4.9 minutes (3.0-9.8) for ATP and ATS, respectively. Beam on time was depending on the absolute dose delivered, with a maximum of 15 min for a single dose of 15 Gy and overall comparable to standard linacs. Mean time for image fusion, contouring and plan adaptation was significantly longer for ATS with 5.0 min (2.8–11.5) and 19.7 min (11.7–35.2) for ATP and ATS (p<0.001), respectively. Also plan assessment takes longer for ATS (p=0.03), whereas increased times for post-RT imaging is due to entity specific imaging protocols. An overview of all gathered mean times for the different working steps of the two adaptation methods is shown in table 1 and figure 1.

A large intra- and interpatient variability was seen in the analysis. As expected the difference between ATP and ATS arose mainly from image fusion, contouring and plan adaptation taking about 15 min longer in ATS, due to the complexity of the procedure. The new processes and close inter-professional cooperation at the MRL were quickly implemented for daily clinical routine.