Will AI widen the gap? Exploring lung cancer patient perceptions of AI.

Abstract

400 Background: Artificial intelligence (AI) is being integrated throughout the cancer care continuum, from administrative functions to diagnostic and monitoring processes, yet substantial concerns regarding implementation and quality remain. One important concern is that it will exacerbate inequities already present in healthcare. This study aims to explore lung cancer patients’ perceptions of AI implementation and utilization in their cancer care and treatment decision-making across different age groups (65 and younger vs. 66 and older) and geographic settings (large city/suburbs vs. small town/rural). Methods: We administered a cross-sectional anonymous pen and paper survey to attendees of a patient and caregiver meeting hosted by a lung cancer patient organization in 2026 (response rate 80%). This study focused on questions about how informed patients were about AI, a questionnaire on the role of AI in their cancer care using a Likert-type agreement scale, and an open-ended question about AI’s role in decision making with doctors. Patient responses (n=88) were summarized using descriptive statistics. Results: Patients were primarily living with metastatic cancer (57%), 71% were <65, 74% living in a large city or suburb while 26% lived in small towns or rural areas, and 76% had a college degree. More than half of the patients (56%) were somewhat to very informed on the uses and applications of AI. Over 40% of patients agreed that AI would improve the accuracy of their treatment and help their oncologist make better decisions. However, regardless of age or geography most patients disagreed with the following statements “I have access to the necessary technology to use AI diagnostic tools” (48%), “my oncologist provides support for understanding AI” (58%) and “the cancer clinic has resources to implement AI” (49%). Notably, for patients 65+, 71% disagreed with “my oncologist provides support for understanding and using AI” compared to 61% of patients ≤65. However, no patients from small towns/rural areas agreed with the statement that their oncologist provides support for understanding AI, compared with 17% in cities and suburbs. Conclusions: Patients in this study were well informed about the concrete use and application of AI technology and believed AI could meaningfully improve their cancer care and clinical decision-making. However, they indicated a lack of clinical infrastructure to support the use of AI. As a patient in a smalltown rural areas noted, “My cancer center doesn’t use AI. We can’t even send messages through the portal.” This perceived lack of readiness was consistent across demographic groups and was greatest among older adults and those in rural areas. The successful implementation of AI-assisted cancer care will require targeted investments in clinics serving diverse patients and those in non-city/suburban areas, equitable access will need to be deliberately engineered into all applications of AI integration.