S1468
Interdisciplinary - Other
ESTRO 2026
cancer diagnoses but remain less likely to receive guideline-concordant treatments(1). This could relate to the higher incidence of significant comorbidities, and frailty which can predict for worse cancer treatment outcomes in this group(2). However, this is also in part due to the underrepresentation of older patients in trials leading to a lack in detailed information regarding the safety and effectiveness of treatments(3). However, chronological age alone is not a sufficient indicator to omit treatment(4). This systematic review aims to identify and evaluate interventions developed to improve access to appropriate treatment (i.e surgery, radiotherapy (RT) or systemic anti-cancer treatment (SACT)) in older
quasi-experimental designs and poor implementation reporting. Embedding quality improvement frameworks and expanding multicentre evaluations are critical to scale equitable, evidence-based cancer care for older populations. References: 1. Swaminathan D, Swaminathan V. Geriatric oncology: problems with under-treatment within this population. Cancer Biol Med. 2015;12(4):275-83.2. Chen RC, Royce TJ, Extermann M, Reeve BB. Impact of age and comorbidity on treatment and outcomes in elderly cancer patients. Seminars in radiation oncology. 2012;22(4):265-71.3. Mishkin GE, Denicoff AM, Best AF, Little RF. Update on Enrollment of Older Adults Onto National Cancer Institute National Clinical Trials Network Trials. J Natl Cancer Inst Monogr. 2022;2022(60):111-6.4. Soto-Perez- de-Celis E, Li D, Yuan Y, Lau YM, Hurria A. Functional versus chronological age: geriatric assessments to guide decision making in older patients with cancer. The Lancet Oncology. 2018;19(6):e305-e16. Keywords: geriatric oncology, multidisciplinary care, equity Evaluating machine learning fairness across publicly available radiotherapy datasets Shreya Chappidi 1,2 , Andra V Krauze 1 1 Radiation Oncology Branch, National Cancer Institute, Bethesda, USA. 2 Department of Computer Science and Technology, University of Cambridge, Cambridge, United Kingdom Purpose/Objective: Machine learning (ML) algorithms are increasingly being applied to radiotherapy (RT) data, including for autosegmentation and contouring, treatment planning, and even clinical outcome prediction. At the same time, clinical ML algorithms have been contested over issues relating to poor cross-cohort Digital Poster Highlight 5178 generalizability, low adoption by clinicians, and concerns over fairness and disparate performance across patient demographics such as race, gender, and socioeconomic status1. To support calls for improved ML fairness and generalizability, we evaluate publicly available RT datasets under the lens of ML fairness. Material/Methods: We perform a systematic review investigating publicly available datasets across cancer sites for protected attribute availability and potential issues inhibiting ML fairness evaluations. The PubMed research database was queried for research articles mentioning (radiotherapy data AND clinical AND publicly available) in any field. Exclusion criteria included papers not
patients with cancer. Material/Methods:
A systematic search of MEDLINE and EMBASE was conducted for studies published from January 2000 to June 2025, in accordance with PRISMA guidelines and a PROSPERO protocol. Eligible studies evaluated interventions aiming to improve access to appropriate treatment for older adults (≥60 years) with cancer, defined as any surgical or oncological therapy. Data was extracted on study design, tumour type, intervention characteristics (using the TIDieR framework), outcomes, and funding. Risk of bias was assessed. Results: From 16,377 records screened, 26 studies were included. Most studies were conducted in Europe (n=17 (65.4%)). They focused on three major cancers and the following modalities (all treatment modalities including RT (42.3%), surgery (30.8%), SACT (23.7%), surgery and SACT (3.8%)). Most were single-centre (n=21) and only three randomised or cluster- randomised trials were identified. Six main intervention types were identified: (1) geriatric assessment (GA)-focused (n=19), (2) prehabilitation with or without GA, (3) inclusion of a geriatrician in the multidisciplinary team (MDT), (4) development of shared decision-making pathways e.g. joint interdisciplinary clinics, (5) decision support tools, and (6) electronic “nudge” interventions. GA-based interventions led to treatment modifications in 6–75% of patients, most often towards less intensive but individualised care. These interventions involved delivery of GA in varying formats from a self- administered GA to involvement of a geriatrician or geriatric oncologist, or a multi-disciplinary onco- geriatric team. None explicitly used quality improvement methodologies in intervention design or evaluation. Conclusion: Interventions incorporating geriatric assessment, MDT integration, or structured decision support can improve treatment access for older adults with cancer. However, evidence is constrained by single-centre,
Made with FlippingBook - Share PDF online