Abstract: Adaptive radiotherapy approaches have been adopted for select sites with the aim of achieving adequate target coverage and OAR sparing each day of treatment in the presence of anatomical variation. Considering the close proximity of highly mobile GI structures, abdominal disease sites are prime targets for online adaptation. The purpose-built Ethos adaptive radiotherapy platform has been commonly employed in the general pelvis, but motion-related artifacts in the abdomen are detrimental to the auto-contouring process on which the Ethos workflow relies. In this ongoing study, we use planning CT and daily CBCT data from a population of pancreatic cancer patients previously treated nonadaptive on Ethos to produce artifact-reduced synthetic CBCT images with which we simulate the adaptive workflow. Results from the simulated fractions indicate that adapted plans are able to accommodate anatomical changes observed in these patients a majority of the time, illustrating the potential promise of Ethos adaptive radiotherapy in this region. This promise is balanced by the requirement for artifact reduction in images and questions regarding auto-contouring performance in the abdomen.

About the speaker: Sven is currently a first-year physics resident in the Harvard Medical Physics Residency Program working with Dr. Jennifer Pursley during his research year. Prior to joining the residency program, Sven completed his PhD in the Department of Biomedical Engineering at Washington University in St. Louis, where his primary focus was deep learning-based synthetic CT reconstruction for MR-only radiotherapy.