Session Item

The PRISM-OUH protocol (Prostate Radiotherapy Integrated with Simultaneous MRI) has included patients with localized prostate cancer treated on the 1.5 T MR-linac (Unity, Elekta AB, Stockholm, Sweden). The protocol aimed to assess the feasibility, safety, and tolerability of the adaptive treatments. The primary endpoint was to evaluate whether on-line MR-guided radiotherapy (oMRgRT) could be delivered within a clinically feasible time-period (more than 90% of the fractions (F) delivered within 60minutes).

This study reports the treatment times from PRISM as well as time evolution of the manuel tasks during the study.

Localized prostate cancer patients were included from December 2018-to April 2020. Radiotherapy (RT) was planned to a total dose of 60 Gy in 20 F to the prostate and proximal 1cm of the seminal vesicles (SV), 48 Gy was given to an additional 1cm of the SV. The PTV margins were 5mm isotropic, except for 3mm posteriorly PTV 60 Gy. Organs at risk (OAR) were bladder, rectum, and bowel. Prostate, SV, and OAR were registered from the reference plan to the daily MR scan and manually adjusted if needed.

At each fraction, the total session time (in-room time) was registered together with time used for image fusion and conturing.

Two different types of workflow were used: Adapt-to-position (ATP) that used rigid contur propagation registration and translation of treatment fields, and adapt-to-shape (ATS) that uses deformable propagation of conturs, contur correction and a full treatment re-planning. 

The trial included 31 patients with a median age of 69 years (46-76), Gleason scores 6-7, and 78% had performance status 0. Of the planned 620 F, 26 F were delivered on standard linac (15: one patient had claustrophobia, 11: technical problems). The median in-room time and range were 25 (18–62)min for ATP, and 35 (21–69)min for ATS (fig 1a). The majority of the fractions (590/594) were delivered within 60 min. Treatment interruption with the need to have the patient off the couch occurred at 25 fractions (8: technical issue, 17: large filling of the rectum). In all cases, these treatments were delivered later during the same day. 

The median time for manual image registration and contouring was 2 min, range (1–11)min for ATP, and 9 min, range (1–35)min for ATS (fig 1b). There was no indication of decreased delineation time over the study period (fig 2).

From this study, we conclude that it is feasible to deliver daily oMRgRT for patients with prostate cancer since 99.3% of the fractions were delivered within 60min. The simpler ATP workflow is significantly faster to deliver than ATS that utilizes all the adaptive planning advantages.  Surprisingly, the time needed for the time-cosuming manual image registration and contouring has not declined over time; this might partly be due to the training of new staff during the study period. Improved deformable image registration algorithms are needed to reduce the treatment time significantly.