ESTRO 2024 - Abstract Book

S3109

Physics - Autosegmentation

ESTRO 2024

Conclusion:

In this study, three AI-based autocontouring models were successfully trained for treatment planning of liver cancer patients on the 1.5 T MR-Linac system. Using MRI only, an AI-model was trained which yields high GTV and at the same time reasonable contouring accuracy. In a next step, the active integration of GTV-autocontouring within the clinical workflow should be initiated and a constant improvement of the model by active integration of unknown patient anatomies considered.

Keywords: Automatic annotations; Deep learning; MR-Linac

References:

[1] Nachbar M, lo Russo M, Gani C, et al. Automatic AI-based contouring of prostate MRI for online adaptive radiotherapy. Z Med Phys 2023; S0939-3889(23)00053-3.

[2] Jullian N, Benkhaled S, Paquier Z, et al. Evaluation of MIM ProtegeAI prostate: Time gain, accuracy and dosimetric consequences. Int J Rad Oncol Biol Phys 2022; 114: e98.

[3] Isensee F, Jaeger P, Kohl S, et al. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 2021; 18: 203–211.

2523

Digital Poster

Deep Learning Segmentation of Lymph Node Segmentation for Adaptive Radiotherapy using nnU-Net

Ricardo C Brioso 1 , Damiano Dei 2,3 , Nicola Lambri 2,3 , Daniele Loiacono 1 , Pietro Mancosu 3 , Marta Scorsetti 2,3