Session Item

Online adaptive radiotherapy based on MR Linac (MRL) can potentially address inter and intra-fractional changes using images acquired daily during treatment. We report the intra- and inter-fractional movement for the first 25 fraction cervical cancer treatment delivered on the MRL in the UK. 

Following institutional approval, a woman with stage FIGO-IIB node-negative cervical cancer was consented to the MOMENTUM trial (NCT04075305) for 25 fractions of radiotherapy. Treatment was prescribed according to the EMBRACE II guidelines using adapt-to-shape (ATS) workflow. A comfortably full bladder was required for treatment.  

Three MRIs were acquired at each fraction: a setup image ( used for plan adaptation), a verification (verif) image (acquired immediately prior to ‘beam on’) and a post-treatment (post-treat) image (to assess intra-fractional changes). A single observer contoured the bladder and CTV on all available verif and post-treat MRIs. Intra-fraction changes in bladder and CTV were determined by measuring the volume differences between the setup and verif images and the volume differences between the verif and post-treat images. The inter-fraction changes in CTV and bladder volumes were determined by measuring the difference between the verif images from each fraction and the MRI used for initial treatment planning.  Intra-fractional changes were reported using Dice coefficients (DC) and inter-fraction changes reported using volume changes in cm³. Statistics have been presented as descriptive statistics and boxplots.

22 fractions and 67 MR images were available for review.  The mean treatment time was 68 min (range 51–82 min) from the time the patient entered the room until the patient left. The median time from the setup image to verif image was 43.5 min (range 31-72 min). The median time from verif image to post-treat image was 13 min (range 4-40 min). The median intra-fraction DC for the CTV was 0.62 (range 0.40-0.80) and 0.90 (range 0.57-0.95) for the setup to verif and verif to post-treat images respectively. The median intra-fractional DC in bladder volume from setup to verif images and from verif to post-treat images were 0.67 (range 0.44-0.83) and 0.90 (range 0.54-0.97) respectively. Figure 1 shows the intra-fraction variation from setup to verif and verif to post treat images for both bladder and CTV. CTV volume changes were greatest in the adaption time, and smaller during radiation delivery. For inter-fraction movement, median volume for CTV and bladder were 7.41cm³ (range -4.30-24.92) and -133cm³ (range-382.63-57), respectively. No consistent bladder or CTV volume patterns were observe.

Drinking protocols were not optimal to keep bladder volume constant. The low DC value justifies the need for an ATS workflow. Real-time adaptation based on MRL addressed inter and intra-fraction CTV and bladder changes by allowing an online correction that improves delineation and treats on the most updated plan.