Radiology artificial intelligence for prioritized imaging and diagnosis of lung cancer: qualitative interview analysis of stakeholder perspectives in Northern Ireland

Lung cancer is a leading cause of death internationally, with most cancers being diagnosed at an advanced stage. Initiatives, such as the Lung Cancer Policy Network promote best practice globally, such as development of screening programs, however, infrastructural issues such as staffing shortages may mean that changes to current practice may not be possible. Artificial intelligence (AI) has been proposed as a means to alleviate the pressure associated with additional imaging, triage and management. Principles of patient centered care should be adopted when considering any factor in healthcare. There is a dearth of literature on the patient and clinician perceptions of the impact of AI in the lung cancer pathway specifically, particularly in nations where this technology is being considered but not currently being utilized. This semi structured interview study recruited both patients and clinician volunteers who had responded to an initial survey on the same topic, resulting in seven members of the public and six clinicians. All participants reside in Northern Ireland, allowing for insight into a nation where AI had not yet been adopted in the lung cancer pathway. Interviews were coded and Braun and Clark’s recommendations for thematic analysis were followed, resulting in seven themes: 1. Person to person communication, 2. Use of AI in health – applications, 3. Validation, 4. Acceptability and variability of acceptance, 5. Education and training, 6. Patient consent AI in their care, 7. Workflow integration and infrastructural limitations. No themes were unique to either clinicians or public. Perception of the outlook for the future with AI in the lung cancer pathway was positive, and in many cased reported to be inevitable. Both clinicians and the members of the public highlighted the need for robust quality assurance to be in place. Opinions varied on the need for explicit patient consent to the use of AI in their pathway, with trust in the clinicians’ decision articulated by members of the public.