ESTRO 2025 - Abstract Book

S3790

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

Conclusion: DNNs showed better results than conventional algorithms for IVIM modelling, with 4 b-values being sufficient for the analysis. The DNNs’ superiority for IVIM parameter quantification elicit an opportunity of using IVIM parameters in response-adaptive MRgRT.

Keywords: IVIM modelling, deep learning, head and neck

References: 1. Kaandorp MPT, Barbieri S, Klaassen R, et al. Improved unsupervised physics-informed deep learning for intravoxel incoherent motion modeling and evaluation in pancreatic cancer patients. Published online March 23, 2021. doi:10.48550/arXiv.2011.01689