S1469
Interdisciplinary - Other
ESTRO 2026
reporting clinical patient data, not involving radiotherapy data, reviews, and where data were inaccessible/insufficiently documented. We analyzed included publicly available datasets across cancer sites, patient cohort characteristics (size, years treated, institutions), reported data modalities, clinical data availability, and treatment characteristics, as these factors can impact data quality and related ML evaluation performance. Results: Our systematic review included 141 papers (2006- 2025) with 9 final publicly available clinical RT sets selected for further analysis, including RADCURE2, NHS RTDS3, and LUND-Probe4. Datasets included head and neck (n=2), prostate (n=2), central nervous system (n=2), lung(n=1), and mixed site cancers (n=2), Median patient cohort size was 317 (range 45-22679) with most repositories sharing imaging data (n=8) in either MRI or planning/simulation CT, alongside RTSTRUCT data (n=7). Clinical data was rarely available, with 2 datasets reporting outcomes (e.g., survival) and another reporting adverse events. Race/ethnicity were not reported in any dataset, while sex/gender were not reported for any head and neck datasets and patient age was missing from approximately half of datasets. Conclusion: Our systematic review of publicly available datasets reveal challenges in effectively employing and evaluating RT-driven algorithms for ML fairness. We identified no publicly available RT datasets that document race/ethnicity, limiting application of ML fairness methods and potentially enabling "fairness through unawareness." Many datasets lacked patient ages and treatment years, making it difficult to contextualize treatment plan choices given evolving guidelines. Lastly, differences in reported imaging modalities, sequences, and scan quality can reduce algorithmic generalizability across cohorts. We advocate for standardized clinical data reporting requirements that protect patient privacy whilst enabling fairness evaluations across RT algorithms. References: 1. Chen, R. J., et al (2023). Algorithmic fairness in artificial intelligence for medicine and healthcare. Nature biomedical engineering, 7(6), 719–742. https://doi.org/10.1038/s41551-023-01056-82. Welch, M. L., et al. (2024). RADCURE: An open-source head and neck cancer CT dataset for clinical radiation therapy insights. Medical physics, 51(4), 3101–3109. https://doi.org/10.1002/mp.16972Sandhu, S., et al. (2023). Cohort profile: radiotherapy dataset (RTDS) in England. BMJ open, 13(6), e070699. https://doi.org/10.1136/bmjopen-2022-0706993. Rogowski, V., et al. (2025). LUND-PROBE – LUND Prostate Radiotherapy Open Benchmarking and Evaluation dataset. Sci Data12, 611.
https://doi.org/10.1038/s41597-025-04954-5 Keywords: AI/ML, fairness, public radiotherapy data
Digital Poster 5189 Well-Being and Burnout among Women Working in Radiotherapy: The Tunisian Experience Farah Nadia Liouane, Rim Abidi, Khedija Ben Zid, Hajer Zelaiti, Nesrine Sallemi, Najla Attia, Alia Mousli, Asma Ghorbel, Hadhemi Ayadi, Semia Zaraa, Chiraz Nasr radiotherapy oncology, institute of Salah Azaiez, tunis, Tunisia Purpose/Objective: Working in a radiotherapy department exposes healthcare professionals to high levels of technical, emotional, and organizational stress. Women, who represent a large proportion of the radiotherapy workforce, may face specific challenges in balancing professional duties with personal and family responsibilities. This study aimed to assess the psychological and social impact of working in radiotherapy on Tunisian female healthcare professionals. Material/Methods: A cross-sectional observational study was conducted among women working in radiotherapy departments across Tunisia.A structured questionnaire was used to collect sociodemographic and occupational data and to assess stress, anxiety, fatigue, work–life balance, and coping strategies. Results: The study included 30 Tunisian female healthcare professionals, predominantly from the public sector (93%), with a median age of 28 years [25–53]. Twenty- seven had less than five years of professional experience. Among the 16 married participants, 11 had children and relied partly on external support for family duties. Despite this, 20 participants reported difficulties maintaining work–life balance, and 22 often thought about work outside regular hours.Psychologically , 20 participants reported work- related stress with consequences extending to their private lives. Anxiety symptoms were reported by 20 participants, while 24 described emotional exhaustion and loss of motivation. Chronic physical and emotional fatigue affecting life outside work was noted by 20 participants. Manifestations of this fatigue included end-of-day exhaustion, sleep disturbances in 19 participants, and insufficient recovery between working days. Overall, 25 participants reported a feeling of burnout.Only 8 participants had consulted a mental health professional or used anxiolytics. Meanwhile, 13 practiced relaxation techniques and 8 engaged in physical activity as coping strategies. Conclusion:
Made with FlippingBook - Share PDF online