ESTRO 2024 - Abstract Book

S3693

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2024

2656

Digital Poster

Automatic treatment planning of cervical cancer using RayStation treatment planning system

Mathilde Hirsum Lystad 1 , Trine Martens 2 , Taran Paulsen Hellebust 1 , Kjersti Bruheim 2

1 Oslo University Hospital, Department of Medical Physics, Oslo, Norway. 2 Oslo University Hospital, Department of Oncology, Oslo, Norway

Purpose/Objective:

Patients with locally advanced cervical cancer (LACC) are treated with concurrent chemoradiotherapy followed by an image-guided brachytherapy (BT) boost to the cervix and residual tumor. In definitive radiotherapy, the tumor related target volume includes the primary cervical tumor, the uterus, parametria and the upper vagina. The elective nodal volume includes pelvic- and, if indicated, para-aortic or inguinal lymph nodes (LNs). According to international guidelines, a dose of 45-46 Gy using intensity-modulated radiotherapy/volumetric modulated arc therapy is recommended (1). Additionally, pathological LNs receive a boost of 55 or 57.5 Gy depending on their proximity to the BT target volume, using the coverage probability principle. Several organs at risk (OARs) are located near the target volume and the dose to these should be as low as possible to minimize late toxicity. Manual treatment planning entails an iterative process where treatment objectives are defined and refined within the treatment planning system (TPS). While treatment planners are highly skilled, individual preferences and experiences may lead to variations in plan quality and consistency. Moreover, due to its complexity, manual treatment planning for LACC is time-consuming. Automatic treatment planning is an ongoing area of research, and includes dose planner mimicking as one potential solution to automate the treatment planning process. Automated treatment planning for LACC could potentially lead to plan quality improvement, less inter-planner variation and a substantial time gain in the clinic. This project aims to develop and implement a Python script for automatic treatment planning of LACC using RayStation TPS. In June 2022, an in-house-developed dose planner mimicking Python script was implemented at St.Olavs Hospital in Trondheim, Norway, for LACC using RayStation TPS (2). This script has been modified to fit the radiotherapy equipment at Oslo University Hospital (OUS). The script mimics the procedure of a manual planning process. In the first phase, four optimizations are run (40+40+100+40 iterations) with the aim of minimizing the dose to OARs. After each optimization, the objectives are adapted according to specified criteria depending on the OAR. Similarly, a second phase is run, also with four optimizations (50 iterations each) with the aim of optimizing the dose to the target volumes. After fine-tuning the script to fulfil the planning aims for LACC according to the Embrace II protocol (3), the script was run on seven cases without pathological LNs and subsequently on another seven cases with pathological LNs (1-7 LNs). For each case, the auto-generated (auto) plan was compared with the clinically delivered (clinical) plan, using dose-volume histogram (DVH) parameters (D98%, D95%, D50%, D1%, V40Gy, V30Gy, average dose). Material/Methods:

Results: