ESTRO 2024 - Abstract Book

S4347

Physics - Intra-fraction motion management and real-time adaptive radiotherapy

ESTRO 2024

the dose from the dose cube at its respective position. Hence, the point clouds represent the delivered dose for each structure whilst accounting for its motion.

The IBSWO algorithm takes the delivered dose point clouds and calculates the segment weights for the next beam, such that the difference between the originally planned dose at the end of the next beam and the sum of the previously delivered dose plus the re-weighted dose of the next beam is minimised. To calculate the next beam dose, the IBSWO accounts for target and OAR motion by updating the patient geometry and associated electron density map and by shifting the MLC shapes according to the last reported target position. Segment doses are calculated, and weights are assigned by solving the minimization problem described above using an analytical least-squares formulation. While it is possible to use point-cloud doses from all structures to perform the minimization, in this study we only used the PTV point-cloud.

Results:

Figure 1 shows the DVHs for the GTV, the PTV, and the Duodenum for the first patient using MLC tracking and MLC tracking in combination with our IBSWO algorithm. The online dose reconstruction of the simulated MLC-tracked treatment (blue) results in a higher dose to the target structures. This is partially due to making MLC trackable, where closed leaf-pairs are moved into the centre of the high-dose region. Using inter-beam segment weight optimization (orange) results in a DVH curve fitting the originally planned dose (black) to the target more closely. At the same time IBSWO substantially lowers the delivered dose to the duodenum in this case.

Table 1. shows selected dose-volume parameters for the same structures. In most cases the dose to the target shows lower differences to the planned dose after using IBSWO, when compared to not using the optimization.