ESTRO 2024 - Abstract Book
S4305
Physics - Intra-fraction motion management and real-time adaptive radiotherapy
ESTRO 2024
Purpose/Objective:
To evaluate accuracy of respiratory tracking in liver treatments with Cyberknife and to propose a strategy for adaptive margins based on the information of the first fraction of treatment
Material/Methods:
20 SBRT liver treatments (23 lesions) were evaluated in this work. For each patient 2 to 4 fiducials were placed in the surroundings of the tumor. The Cyberknife system performs respiratory tracking of the tumor by building a model of its position detected by its X-ray image system against three LEDs placed on the surface of the patient. This model is updated during treatment with displacements gathered by in-treatment follow up X-ray images. Thus, it is possible to calculate accuracy of tracking errors by comparing in-treatment displacements found by follow-up images against positions predicted by the tracking model. Cyberknife only performs a model for translations, only being able to correct for the average rotations found during the whole breathing cycle. Thus, as intratumor placement of fiducials is not recommended in liver lesions, rotations occurring during the breathing cycle may result in important displacements of the tumor position against the center of mass of the fiducials. Patient population statistics were analyzed for both fiducials and tumor positions (considering rotations). Mean displacements (µ), systematic errors (Σ) and inter- and intrafractional random errors (σinter, σintra) were obtained. Margin errors based on an 80% isodose level and beam penumbrae of σp=1.5mm were also calculated. A secondary goal is to analyze if margin adaptation is feasible based on data gathered during the first fraction. Thus, mean and standard deviation of residual errors for the first fraction and for the rest of the treatment were obtained. From these data, margins needed to account for these errors (mean displacement and margin reduction due to random errors) were calculated and compared.
Results:
Made with FlippingBook - Online Brochure Maker