ESTRO 2024 - Abstract Book

S4498

Physics - Machine learning models and clinical applications

ESTRO 2024

However, more research is needed to prevent overfitting. Furthermore, looking into recontouring instead of contour propagation is needed to improve the ground-truth classification. Enhancing the multi-label classification or using transfer learning could potentially also better distinguish the errors.

Keywords: in-vivo EPID-dosimetry, 2D-gamma, DVH-metrics

References:

[1] Rozendaal RA, Mijnheer BJ, Van Herk M, Mans A. In vivo portal dosimetry for head-and-neck VMAT and lung IMRT: linking γ-analysis with differences in dose-volume histograms of the PTV. Radiother Oncol 2014;112:396–401. doi:10.1016/J.RADONC.2014.03.021. [2] Olaciregui-Ruiz I, Beddar S, Greer P, Jornet N, McCurdy B, Paiva-Fonseca G, et al. In vivo dosimetry in external beam photon radiotherapy: Requirements and future directions for research, development, and clinical practice. Phys Imaging Radiat Oncol 2020;15:108–16. doi:10.1016/J.PHRO.2020.08.003.

[3] Mans A, Wendling M, McDermott LN, Sonke JJ, Tielenburg R, Vijlbrief R, et al. Catching errors with in vivo EPID dosimetry. Med Phys 2010;37:2638–44. doi:10.1118/1.3397807.

[4] Nijsten SMJJG, Mijnheer BJ, Dekker ALAJ, Lambin P, Minken AWH. Routine individualised patient dosimetry using electronic portal imaging devices. Radiother Oncol 2007;83:65–75. doi:10.1016/J.RADONC.2007.03.003.

[5] Nijsten SMJJG, Van Elmpt WJC, Jacobs M, Mijnheer BJ, Dekker ALAJ, Lambin P, et al. A global calibration model for a-Si EPIDs used for transit dosimetry. Med Phys 2007;34:3872–84. doi:10.1118/1.2776244.

[6] Lin TY, Goyal P, Girshick R, He K, Dollar P. Focal Loss for Dense Object Detection. IEEE Trans Pattern Anal Mach Intell 2020;42:318–27. doi:10.1109/TPAMI.2018.2858826.

1807

Digital Poster

Deep learning-based dose prediction for breast and regional lymph node radiotherapy

Akseli Leino 1 , Janne Heikkilä 1 , Tuomas Virén 1 , Juuso TJ Honkanen 1 , Jan Seppälä 1 , Henri Korkalainen 1,2

1 Kuopio University Hospital, Center of Oncology, Kuopio, Finland. 2 University of Eastern Finland, Department of Technical Physics, Kuopio, Finland

Purpose/Objective:

Treatment planning is a time-consuming and complex process requiring an iterative optimization process based on trial and error to obtain a feasible plan. Moreover, the outcome may vary in consistency and quality due to human