ESTRO 2024 - Abstract Book

S4372

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

ESTRO 2024

2691

Digital Poster

Use of the EPID to actively monitor single fraction stereotactic treatments

Paul J Doolan 1 , Afxentia Karousiou 2 , Natassa Pittas 2 , Eleni Hadjitoouli 3 , Constantinos Zamboglou 4,5,6 , Konstantinos Ferentinos 4,6 1 German Oncology Center, Medical Physics, Limassol, Cyprus. 2 University of Cyprus, Physics, Nicosia, Cyprus. 3 University of Surrey, Physics, London, United Kingdom. 4 German Oncology Center, Radiation Oncology, Limassol, Cyprus. 5 University of Freiberg, Radiation Oncology, Freiberg, Germany. 6 European University Cyprus, Medicine, Nicosia, Cyprus

Purpose/Objective:

Stereotactic treatments involve delivering a larger dose per fraction over a reduced number of fractions, making correct patient positioning, patient anatomy consistent with planning CT and machine performance, critical. Typically, this is ensured with additional pre-treatment machine quality assurance, more restrictive patient immobilisation devices and tighter CBCT tolerances. Using the electronic portal imaging device (EPID) to capture the exit dose allows for treatment monitoring, by comparison with a predicted dose. However, a failing EPID would not be rectifiable for single fraction treatments. We propose a method to use the EPID proactively, by splitting the original plan into two separate plans. Provided the agreement between predicted and measured EPID doses are within specified limits for the first part of the plan, the treatment could proceed with confidence and the second part (remainder) of the plan would be delivered.

Material/Methods:

The Python package Pydicom was used to create a routine that can take as its input a DICOM RT plan with a treatment beam, and generate as its output two DICOM RT plans, each with a split component of the original treatment beam. The original plans were created in RayStation TPS v12.0, using VMAT beams that were composed of many control points. Upon splitting into 10%-90% proportions, the first plan would contain a beam with the control points corresponding to the first 10% of dose, while the second would have the control points corresponding to the remaining 90% of dose. There are many DICOM tags that must be accounted for in order for each of these new plans to be deliverable (e.g. Final Cumulative Meterset Weight (DICOM tag 300A,0134) of each Control Point Item, or renumbering the Control Point Index (DICOM tag 300A,0112) among others). Ideally, the lowest dose should be delivered to the patient before confirmation, so lower dose proportions between 2% and 30% were tested (and corresponding high dose proportions of 98% to 70%), on a variety of phantoms with different target shapes and sizes. In testing different dose proportions, a relationship was established (quadratic regression) that allowed the user to predict the passing rate based on the dose proportion. If, when using this relation, the measured passing rate is less than predicted by the relationship, it is suggestive that there is an issue with the treatment and the operator should not proceed.