ESTRO meets Asia 2024 - Abstract Book

S402

RTT – Treatment planning, OAR and target definitions

ESTRO meets Asia 2024

plans were deemed to achieve similar level to or even better in plan quality than the original-accepted manual plans, as shown in Fig.1. The differences of DVH metrics between the automated plans and the manual plans were extracted for the objective comparison, and part of the results of the institution A and B are shown in Fig. 2.

Fig.1. Results of the subjective assessments of the ATP plans for common tumor sites from the institution A.

Fig.2. Plots of the differences of the DVH metrics (ATP plan minus manual plan) for rectal cases from the institution A and B (upper), and for lung cases from the institution A (bottom), respectively.

Conclusion:

The proposed method has brought the automation of treatment planning to a real-world practice. It is indicated initially that the characteristic planning styles in dose requirements among different institutions can be achieved by modifying the goal sheet within the same DL model. The clinically accessible and deliverable ATP solution will lead to higher efficiency and potentially more homogenous plans, and pave the way for facilitating online replanning in personalized adaptive radiotherapy.

Keywords: deliverable automated planning, deep learning

References: