ESTRO 2024 - Abstract Book

S4488

Physics - Machine learning models and clinical applications

ESTRO 2024

Conclusion:

Using a co-creation strategy, we were able to rapidly prototype a novel human-interpretable system to detect treatment errors and aid the decision-making process. The simplicity of the system interface allows users to rapidly and efficiently evaluate the complex IVD data, resulting in a workable clinical system. The prototype sparked enthusiastic responses from end-users. Part of the project is to further develop a verification framework for the AI reported error and automated AI trend analysis.

Keywords: AI-based error detection, Design

References:

[1] I. Olaciregui-Ruiz et al., “In vivo dosimetry in external beam photon radiotherapy: Requirements and future directions for research, development, and clinical practice,” Phys. Imaging Radiat. Oncol., vol. 15, pp. 108–116, Jul. 2020, doi: 10.1016/j.phro.2020.08.003. [2] C. J. A. Wolfs, R. A. M. Canters, and F. Verhaegen, “Identification of treatment error types for lung cancer patients using convolutional neural networks and EPID dosimetry,” Radiother. Oncol., vol. 153, pp. 243–249, Dec. 2020, doi: 10.1016/j.radonc.2020.09.048.

[3] J. P. Roberts, T. R. Fisher, M. J. Trowbridge, and C. Bent, “A design thinking framework for healthcare management and innovation,” Healthcare, vol. 4, no. 1, pp. 11–14, Mar. 2016, doi: 10.1016/j.hjdsi.2015.12.002.

1596

Mini-Oral

The impact of dosimetric features on interpretable overall survival prediction in NSCLC radiotherapy

Joshua R Astley, James Reilly, Stephen Robinson, Jim M Wild, Matthew Q Hatton, Bilal A Tahir

The University of Sheffield, Division of Clinical Medicine, Sheffield, United Kingdom

Purpose/Objective:

The Radiation Therapy Oncology Group 0617 revealed that a high-dose regimen of 74 Gray over 37 fractions (74Gy/37fr) unexpectedly reduced overall survival (OS) compared to the standard 60Gy/30fr regimen in non-small cell lung cancer (NSCLC) patients [1]. This diminished OS has been attributed to increased radiation exposure to the heart. Additionally, inadvertent radiation to other organs-at-risk (OARs), i.e. the lungs and oesophagus, can give rise to radiation-induced conditions like pneumonitis and oesophagitis that correlate with reduced OS in NSCLC patients receiving radiotherapy [1-3].