ESTRO 2024 - Abstract Book

S5108

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2024

Keywords: Diffusion-weighted MRI, Imaging biomarker, LAPC

References:

[1] de Mol van Otterloo SR, Christodouleas JP, Blezer ELA, et al. The MOMENTUM Study: An International Registry for the Evidence-Based Introduction of MR-Guided Adaptive Therapy. Front Oncol. 2020;10.

[2] Rahbek S, Madsen KH, Lundell H, et al. Data-driven separation of MRI signal components for tissue characterization. J Magn Reson. 2021;333:107103.

[3] Rahbek S, Mahmood F, Tomaszewski M, et al. Decomposition-based framework for tumor classification and prediction of treatment response from longitudinal MRI. Phys Med Biol. 2023;68.

2237

Poster Discussion

Imputation method comparison on longitudinal patient reported toxicity scores

Luuk van der Hoek 1 , Suzanne P.M. de Vette 1 , Hans Paul van der Laan 1 , Arjan van der Schaaf 1 , Lisa G.L.J. van den Bosch 1 , Hendrike Neh 1 , Nanna M. Sijtsema 1 , Johannes A. Langendijk 1 , Peter M.A. van Ooijen 1,2 , Lisanne V. van Dijk 1,3 1 University Medical Centre Groningen, Radiation Oncology, Groningen, Netherlands. 2 Rijksuniversiteit Groningen, Data Science Centre in Health, Groningen, Netherlands. 3 University of Texas MD Anderson Cancer Centre, Radiation Oncology, Houston, USA

Purpose/Objective:

Toxicity evaluation before, during and after head and neck radiotherapy reflects the treatment-induced burden to patients and is essential for the development of Normal Tissue Complication (NTCP) models. NTCP-models improve our understanding of the relation between radiation dose and toxicity, and may be helpful to guide treatment planning optimization strategies. However, missing values in toxicity scoring datasets challenge their use for advanced data-driven modelling (e.g. machine/deep learning). Imputation methods can systematically generate data for missing values aiming to mitigate this issue. The toxicity used for this study was patient-rated xerostomia (dry mouth), which is a common toxicity for head and neck cancer (HNC) patients following radiotherapy. The aim of this study was to compare the performance of multiple imputation methods for longitudinal toxicity data for 16 time points during and after treatment, with xerostomia as a use-case.

Material/Methods:

Xerostomia scores were available for 1442 HNC patients treated with primary radiotherapy at a Dutch medical centre. Xerostomia scores were rated using an ordinal 4-point Likert scale (not at all, a little, quite a bit, very much; EORTC