ESTRO 2024 - Abstract Book

S3018

Physics - Autosegmentation

ESTRO 2024

Conclusion:

This study highlights the significance of uncertainty estimation in strengthening the reliability of deep learning for clinical applications on auto-segmentation of HNC GTV-T and GTV-N. The results emphasize that while multiple methods can yield similar segmentation accuracy, their reliability, as measured by calibration errors, can vary markedly. Consequently, choosing an uncertainty estimation method like PhiSeg, which aligns confidence with its observed accuracy, is crucial for clinicians to use and trust.

Keywords: uncertianty, head and neck, GTV

References:

[1] V. Andrearczyk et al., “Overview of the HECKTOR Challenge at MICCAI 2022: Automatic Head and Neck Tumor Segmentation and Outcome Prediction in PET/CT,” Head Neck Tumor Segmentation Outcome Predict. Third Chall.