ESTRO 2023 - Abstract Book

S407

Sunday 14 May 2023

ESTRO 2023

Conclusion Deep-learning-based automatic segmentation can produce high-quality segmentations. Our method, which produces multiple segmentation variants instead of just one, was evaluated to give the best results, ranking better even than the reference segmentations. PD-0499 Optimising cervix brachytherapy planning using OAR dose estimates from external beam RT daily CBCTs E. Tassano-Smith 1 , A.L. Palmer 1 , W. Polak 1 1 Portsmouth Hospitals University Hospitals NHS Trust, Medical Physics, Portsmouth, United Kingdom Purpose or Objective Cervical cancer is prescribed EBRT and brachytherapy (BT) to achieve a D90% for the HRCTV above 85 Gy [1]. Current BT practice assumes OARs (bladder, rectum, sigmoid, bowel) receive a dose equal to the EBRT target prescription dose. This assumption lacks supporting data [1-6]. The actual delivered OAR dose may differ from the assumed dose and lead to sub-optimal BT plan optimisation for OAR sparing and HRCTV dose escalation. No previous studies have been found investigating this issue. We propose a methodology for optimising BT planning by considering an estimate of the actual EBRT OAR dose. This is derived by recalculating the treatment plan on daily CBCT images, with deformed OAR contours. The dose to the highest irradiated volume (D2cc) on each daily CBCT is accumulated to determine a maximum estimate of delivered EBRT OAR doses.