Session Item

The Varian ETHOS system enables online adaptive radiotherapy (oART) based on CBCT with daily AI-assisted adaption of the treatment plan to the anatomy of the day. Most clinical treatments focused on the pelvic region in entities like prostate, bladder and anal cancer (Calmels et al. 2022). Cervical cancer is another challenging target for oART due to significant interfractional changes in the position of the uterus and tumor shrinkage and secondly due to changes of rectum and bladder filling. Here, we report the first institutional experiences with cervical cancer oART on the ETHOS system.

Five patients with cervical cancer have been treated with primary chemo-radiotherapy. In total 124 oART fractions between March 2022 and October 2022 have been applied. Treatment of the cervical cancer and the pelvic lymph drainage was prescribed with 25 to 28 fractions, with two patients receiving a homogeneous PTV dose of 1.8 Gy, while three patients received a simultaneous integrated boost to pelvic lymph node metastasis with a second dose level of 2.2 Gy. Additionally, patients received a brachytherapy boost to the primary tumor. Comparison of originally scheduled and online-adapted plans in terms of DVH metrics was performed after manual import of each fraction into Eclipse treatment planning system (version 15.6). Subsequent analysis was done in Origin (v. 9.6).

The overall duration of the oART workflow from the start of CBCT acquisition to the end of treatment was 32.9 min (SD = 6.3 min). Figure 1 shows a case for the differences in dose distribution between the scheduled (Fig. 1a) and corresponding adapted plan (Fig. 1b). A clear benefit in target coverage was achieved by adaptation to the actual uterus position. Figure 1c shows the DVHs of the most relevant structures, which were created by averaging all scheduled (dashed lines) and the corresponding adapted plans (solid lines). Improvements in target coverage differed between patients, as seen in Fig. 2a. Applying the scheduled instead of the adapted plan would result in greatly reduced DVH metrics for PTV D95% and D98%. Except for one patient dose to the rectum was reduced by oART (Fig. 2b). By applying the adapted plan the average reduction of mean rectum dose was 5.8%. An overall dose reduction to the bladder was achieved for three patients (Fig. 2b), with a negligible change in mean bladder dose of <0.1%. This conflicting picture for bladder sparing may be a consequence of prioritizing target volume coverage over OAR sparing and secondly of differences in bladder filling.

Initial clinical evidence shows that CBCT-based oART can provide significant dosimetric benefits in cervical cancer. This may translate in meaningful improvement in local control and OAR sparing.