ESTRO 2025 - Abstract Book

S3742

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2025

Keywords: RE, Delta radiomics, temporal sequencing

References: [1] Mireștean CC, Iancu RI, Iancu DPT. Image Guided Radiotherapy (IGRT) and Delta (Δ) Radiomics-An Urgent Alliance for the Front Line of the War against Head and Neck Cancers. Diagnostics (Basel). 2023;13(12):2045. [2] Tran WT, Suraweera H, Quiaoit K, et al. Quantitative ultrasound delta-radiomics during radiotherapy for monitoring treatment responses in head and neck malignancies. Future Sci OA. 2020;6(9):FSO624. doi: 10.2144/fsoa-2020-0073. [3] Fatima K, Dasgupta A, DiCenzo D, et al. Ultrasound delta-radiomics during radiotherapy to predict recurrence in patients with head and neck squamous cell carcinoma. Clin Transl Radiat Oncol. 2021;28:62-70.

1369

Proffered Paper Multi-centre external validation of NTCP models for predicting 6- and 12-month dysphagia in head and neck cancer patients Sarah W Stougaard 1,2 , Ruta Zukauskaite 3 , Richard Röttger 4 , Ebbe L Lorenzen 1,2 , Maximilian L Konrad 1,2 , Camilla P Nielsen 1,2 , Simon L Krogh 1 , Carsten Brink 1,2 , Jeanette FA Sommer 1 , Nis Sarup 1 , Jørgen Johansen 3,5 , Jesper G Eriksen 6,7 , Camilla K Lonkvist 8 , Jeppe Friborg 9 , Christian R Hansen 1,2,5 1 Laboratory of Radiation Physics, Odense University Hospital, Odense, Denmark. 2 Department of Clinical Research, University of Southern Denmark, Odense, Denmark. 3 Department of Oncology, Odense University Hospital, Odense, Denmark. 4 Department of Mathematics and Computer Science, University of Southern Denmark, Odense, Denmark. 5 Danish Centre for Particle Therapy, Aarhus University Hospital, Aarhus, Denmark. 6 Department of Oncology, Aarhus University Hospital, Aarhus, Denmark. 7 Department of Experimental Clinical Oncology, Aarhus University Hospital, Aarhus, Denmark. 8 Department of Oncology, Copenhagen University Hospital Herlev, Herlev, Denmark. 9 Department of Oncology, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark Purpose/Objective: Despite the increased growth in development of clinical prediction models, few models are implemented clinically, and even fewer are properly validated before implementation (1). An NTCP model aimed at predicting dysphagia at 6 months (2) is used to determine patient eligibility for proton/photon treatment randomisation in the DAHANCA35 trial (NCT04607694). This study externally validates this model using a real-world DAHANCA cohort. Material/Methods: This study included 1,948 pharyngeal and laryngeal squamous cell carcinoma patients treated consecutively between 2010-2015 at three DAHANCA centres. Data on dysphagia, survival, and patient-, tumour-, and RT characteristics were available. For patients with missing baseline dysphagia, a score of 0 was assumed. An AI segmentation model (3) was applied to ensure robust and consistent OAR segmentation. The NTCP models under validation were proposed by Van den Bosch et al. (2), which use logistic regression to predict dysphagia grade 2+ (ranging from only able to eat soft food to complete PEG tube dependency) at 6 and 12 months post-treatment. The models use the following predictors: mean doses for oral cavity and pharynx constrictor muscles upper, middle and lower, baseline dysphagia, and tumour site. Validation was performed using the closed testing procedure for model updating (4), sequentially testing three potential modifications: 1) intercept update, 2) recalibration, and 3) complete model revision. Original coefficients were adjusted to account for the current cohort's absence of oral cavity cancers. Model performance was assessed using AUC and calibration plots.

Results: The validation cohort included observations with complete information, resulting in 726 patients for the 6-month