ESTRO 2023 - Abstract Book

S408

Sunday 14 May 2023

ESTRO 2023

Materials and Methods Three methods were used for calculation of total OAR dose: A) EBRT (assumed OAR dose = target prescription) + original BT B) EBRT (estimate of delivered OAR dose from daily CBCT) + original BT C) EBRT (estimate of delivered OAR dose from daily CBCT) + optimised BT

The method was used for one patient as a proof of concept. The contours were propagated onto the CBCT data set with a rigid and deformable registration, manually edited and checked by qualified staff. The planned EBRT treatment was recalculated in Pinnacle via a validated CBCT to ED curve [7]. The D2cc was reported for each OAR (excluding high dose regions near involved nodes not relevant to BT). This was repeated on all CBCTs with D2cc doses summed giving a maximum estimate of the accumulated EBRT OAR dose. This was converted to EQD2 and used to replan the original BT plan in BrachyVision. Results The process was used successfully to give an estimate of the maximum OAR EBRT delivered dose. The results in Table 1 show a difference in the combined EBRT+BT dose calculated by method A and B. When the EBRT dose was changed from the original assumption to the best-estimate of delivered dose, the rectum dose exceeded the mandatory constraint. The rectum dose was brought back into tolerance by replanning the BT with the estimate of delivered EBRT OAR dose (method C, Table 1) at the expense of reducing the HRCTV dose by 0.6 Gy (EQD2). A clinician was consulted and favoured the rectum sparing plan. In all cases the bladder dose exceeded mandatory tolerance as per a clinical decision to accept higher dose due to bladder involvement.

Conclusion We proposed a novel method using an estimate of delivered EBRT OAR doses when planning BT. In the case presented, a differently optimised BT plan would have been used clinically by the CCO. Further work will determine the magnitude of changes in OAR doses that will lead to clinically meaningful differences in the optimisation of BT planning and determine the related uncertainty and decision criteria.

PD-0500 Dose prediction in HDR brachytherapy for cervical cancer using 3D transformer-based deep learning W. Jian 1 , L. Zhu 1 , Y. Zhang 2 , B. Zhang 1 , X. Wang 1 1 The Second Affiliated Hospital, Guangzhou University of Chinese Medicine, Department of Radiation Therapy, Guangzhou, China; 2 Guangzhou University of Chinese Medicine, School of Medical Information Engineering, Guangzhou, China Purpose or Objective High-dose-rate (HDR) brachytherapy plays a critical role in radiotherapy of patient with locally advanced cervical cancer. However, the process of HDR brachytherapy treatment planning is time-consuming and user-dependent. A clinical QA tool is necessary to propose for detecting the suboptimal needle placement in HDR interstitial brachytherapy. This study aimed to introduce a 3D transformer-based deep learning method for dose prediction in HDR brachytherapy for cervical cancer, with the goal of improving the quality of treatment plans and facilitating the workflow of HDR brachytherapy.