ESTRO 2024 - Abstract Book

S3009

Physics - Autosegmentation

ESTRO 2024

Conclusion:

We produced an algorithm tailored for automated thoracic OAR delineation from national consensus definitions. The resulting dose distributions to OARs were in excellent agreement with those generated using STAPLE ground truth delineations. This paves the way for development of national standards for and implementation of consistent and fast automated OAR delineation in clinical practice.

Keywords: Lung cancer, deep learning, organs at risk