ESTRO 2024 - Abstract Book

S3162

Physics - Autosegmentation

ESTRO 2024

Conclusion:

This study demonstrates that both systems, despite their minor differences, can be employed in clinical practice, significantly reducing contouring time compared to manual methods. As seen in other studies, these algorithms exhibit limitations in contouring tubular structures like the rectum [ 6 ]. Furthermore, no differences were observed in DVH parameters with clinical implications.

For future studies, expanding the number of observers and employing more interesting qualitative metrics such as the Dice coefficient (DICE) or overlap index, commonly used in this kind of studies, is recommended

Keywords: Deep Learning, Autocontouring, Clinical usability

References:

[1] BAROUDI, Hana, et al. Automated Contouring and Planning in Radiation Therapy: What Is ‘Clinically Acceptable’?. Diagnostics, 2023, vol. 13, no 4, p. 667.