ESTRO 2024 - Abstract Book

S36

Invited Speaker

ESTRO 2024

This session will focus on:

•

Diverse Research Activities Involving RTTs

•

The Need for an RTT-Research Network

•

Discussing the gaps and challenges for RTTs in research

•

How to tailor the network to RTTs' needs

•

How we could utilise the network to fill in the gaps

3334

Fairness towards ALL users

Stine S Korreman 1 , Mathis E Rasmussen 2 , Jesper G Eriksen 3

1 Aarhus University, Department of Clinical Medicine, Aarhus, Denmark. 2 Aarhus University Hospital, Department of Oncology, Aarhus, Denmark. 3 Aarhus University Hospital, Department of Experimental Clinical Oncology, Aarhus, Denmark

Abstract:

Artificial intelligence (AI) has the potential to facilitate automation of time-consuming and complex tasks for increased utilization of radiotherapy on a global scale. In that sense, AI can be a tool in favor of fairness for patients by providing increased access to radiotherapy for all, including in low- and middle-income countries (LMICs). Seen from the point of view of medical professionals in low-resource settings where there is a lack of staff and where time-consuming tasks are a bottleneck, automation by AI can give an increased opportunity for providing standard-of-care treatment thus balancing the grounds for all users. However, the access to AI based tools requires acquisition of relevant software and hence comes with a price tag, which may be prohibitive to actually benefitting from AI. If use of high level AI tools for complex tasks is available only to those able to afford a high price tag, there is a risk of 'automating inequality'. Most of the research and development of AI based tools is conducted in high-resource clinics. However, in LMICs, local factors such as patient abundance, clinical and financial resources, technical expertise, and mind-sets of clinicians may influence how such tools are implemented and used. In a recent study, conducted in collaboration with the International Atomic Energy Agency, we investigated use of AI-assisted contouring, with 90+ radiation oncologists from 23 clinics in LMICs across five continents. We demonstrated that use of AI-assistance in contouring of organs at risk in head and neck cancer gave less inter-observer variation, fewer delineation outliers and closer correspondence with expert delineations, than in manual delineation (abstract by Mathis Rasmussen et al). Also, time spent on delineation was shorter when AI assistance was used compared to manual delineation. This all points to the potential benefits of providing access to AI tools on a global scale at reasonable or even no cost. With open-source access to AI-based tools, provided by clinics and research groups with resources for research, development and deployment, it may be possible to mitigate risk of automating inequality and move towards fairness for all patients and all users.