Cone beam computed tomography guided online adaptive radiotherapy for cervical Cancer: Two-year single-Centre
Danique L J Barten1,2, Karin Goudschaal1, Karel A Hinnen1,2
1Amsterdam UMC, location University of Amsterdam, Department of Radiation Oncology, Meibergdreef 9, Amsterdam, the Netherlands.
Purpose:
Managing inter- and intrafraction motion remains a major challenge in cervical cancer radiotherapy. This study reports a two-year, single-centre experience with daily CBCT-guided cervical online adaptive radiotherapy (oART), evaluating the workflow feasibility of a radiation therapist (RTT)-led adaptive workflow, including treatment session duration and target coverage.
Materials And Methods:
Thirty patients with cervical cancer were treated with CBCT-guided oART between October 2023 and December 2025. After initial supervised sessions, treatment delivery transitioned to an RTT-led workflow supported by a traffic-light protocol with on-call multidisciplinary backup. Cervical oART was introduced in phases, progressing from single-isocentre cases to more complex treatments including two isocentres and para-aortic lymph nodes. Workflow evaluation was based on retrospective analysis of session logs, process timing and oART workflow-related interventions.
Results:
A total of 744 adaptive sessions were analysed. Median total session duration was 33.0 min (interquartile range 13.0). Imaging and beam delivery times were stable, whereas decision-intensive steps showed greater variability, requiring corrections to AI-generated contours in all patients. Median total session duration was comparable between the first five supervised sessions and the subsequent RTT-led sessions (32.5 vs. 32.0 min). Contour review showed the strongest correlation with total session duration (ρ = 0.73). Uterus displacement outside the PTV due to intrafraction motion was observed in approximately 11% of all post-treatment CBCT-verifications and was typically limited to small fundal regions.
Conclusion:
Daily CBCT-guided cervical oART can be efficiently implemented within an RTT-led workflow, demonstrating workflow feasibility while maintaining target coverage despite intra-fraction anatomical variability.
