Radiation oncology patients' perceptions of artificial intelligence and machine learning in cancer care: A multi-centre cross-sectional study.

AIM: The use of artificial intelligence (AI) and machine learning (ML) is increasingly widespread in radiation oncology. However, patient engagement to date has been poor. Respect for persons in the healthcare setting and the principle of informed consent requires recognition of patient perspectives. The aim of this study was to provide a baseline understanding of patient views about the use of AI/ML in the specific context of radiotherapy to contribute towards future governance of the technology.

METHODS

We developed a new questionnaire regarding AI/ML use in radiotherapy. Radiation oncology patients were surveyed from June to October 2024 at two public hospitals in Australia. Questions were on a five-point Likert scale and grouped into six topics. A free text item allowed participants to comment further.

RESULTS

We analysed 474 completed questionnaires (474/811, 58 % completion rate). Most participants supported using AI/ML to help physicians with radiation oncology specific tasks (Median Score (MS) 4.3) and held positive views on the general benefits of AI/ML (MS 4.0). Patients also strongly expressed a preference to be aware and informed (MS 2.2). Significant uncertainty remained about whether AI/ML use would enable retention of the human touch and equity in care (MS 3.1).

CONCLUSION

This is the largest questionnaire study to date of radiation oncology patients' perceptions of AI/ML, establishing a clear baseline. These results can inform future governance around AI/ML in radiotherapy. Actionable steps include informing patients of AI/ML use in their care and engaging physicians during development and regulation of the technology.