ESTRO 2025 - Abstract Book

S3409

Physics - Machine learning models and clinical applications

ESTRO 2025

2832

Proffered Paper Engaging patients’ experience in a cumulative toxicity score as endpoint for clinical trials: an example from cervical cancer Marta Pelizzola 1 , Kari Tanderup 1,2 , Kathrin Kirchheiner 3 , Supriya Chopra 4 , Remi Nout 5 , Sofia Spampinato 5 1 Danish Centre for Particle Therapy, Aarhus University Hospital, Aarhus, Denmark. 2 Department of Clinical Medicine, Aarhus University, Aarhus, Denmark. 3 Department of Radiation Oncology, Comprehensive Cancer Center, Medical University of Vienna, Vienna, Austria. 4 Department of Radiation Oncology, Advanced Centre for Treatment Research and Education in Cancer, Tata Memorial Centre, Homi Bhabha National Institute, Navi Mumbai, India. 5 Department of Radiotherapy, Erasmus MC Cancer Institute, Rotterdam, Netherlands Purpose/Objective: To develop a cumulative toxicity score (CTS) integrating patient-reported outcome and physician-assessed toxicity for applications in clinical trials. Material/Methods: Physician-assessed toxicities, patient-reported symptoms and quality of life (QOL) data recorded in a large prospective study (EMBRACE-I) (1,2,3) on locally advanced cervical cancer were used. The CTS was constructed as a weighted sum including toxicities/symptoms related to treatment. Toxicities/symptoms to be included in the CTS and weights were selected based on their association with both QOL and radiation treatment intensity (RTI). For this purpose, 20 toxicities (CTCAEv3.0) and 23 patient-reported symptoms (EORTC-C30/CX24) recorded in EMBRACE-I were selected to build two CTS based on physician-assessed toxicity (CTS CTCAE ) and patient-reported symptoms (CTS EORTC ). Follow-ups every 3 months during the first year after treatment and every 6 months thereafter until 36 months were considered. For each CTS, two random forest (RF) models ranked toxicities/symptoms based on their association with: 1) Global-Health/QOL (EORTC-C30), and 2) intensity of EBRT (nodal boosting, para-aortic irradiation, body V43Gy) and brachytherapy (bladder/rectum D 2cm3 , isodose-surface-volume 60Gy EQD2 ). CTS CTCAE at follow-up t was calculated as:

th CTCAE toxicity; and Grade

CTCAE-i severity for the i th

with N CTCAE number of toxicities; w CTCAE-I weight assigned to the i

toxicity at t. CTS t

th symptom at follow-up t:

EORTC was calculated using EORTC weights (w EORTC-i ) and scores for the i

CTS CTCAE and CTS EORTC were tested within the EMBRACE-I cohort evaluating differences between patient subgroups based on: nodal boosting (yes/no), para- aortic irradiation (yes/no) and EBRT prescribed dose (>46Gy/≤46Gy) at each follow-up using t-test to identify differences (p-value<0.05). Results: 1040 patients were included in the analysis. According to RF rankings, 8 CTCAE toxicities and 15 EORTC symptoms were included in CTS CTCAE and CTS EORTC , respectively (Table1). Combined weights taking into account both QOL and RTI used in the calculation are reported in Table1.