Surgeon Perception of Artificial Intelligence in Thoracic Surgery

J. Platz¹, D. Bryan², M. Ferguson², K. Naunheim^3
1Saint Louis University, Saint Louis, Missouri ²The University of Chicago, Chicago, Illinois ³Saint Louis University, St. Louis, Missouri

Purpose: The use of artificial intelligence (AI) in clinical medicine has increased dramatically.1 Its utilization will expand in cardiothoracic (CT) surgery in a manner that will likely alter current practice.2 We surveyed CT surgeons to gain insight into their beliefs, hopes, and fears regarding AI in cardiothoracic surgery.

Methods: Working with a dedicated survey lab, a 25-question instrument addressing AI use in CT surgery was created, focusing on the domains of role, value, effect, threat, and regulation of AI. In coordination with the Society of Thoracic Surgery (STS), the survey was electronically distributed and open during May and June, 2024 (1.5 months) to the society members and completed using the Qualtrics platform. De-identified responses were collected by the survey lab and analyzed by the lab and authors. Over 6500 invitations were sent, 297 (4.2%) non-duplicate responses were collected, and 91% of those (270) were complete or near complete. Likert scale answers were scored from 1 to 5 and data was collapsed into 3 categories for chi-squared analysis.

Results: Characteristics of the physicians who completed the survey are listed in the Table. There was a broad range of responses within domains and among individual statements that varied from very negative to very positive (Figure). There were few statistically significant differences among demographics and associated responses (20 among 90 response sets) and 12 (60%) of these occurred in two domains: the role of AI in the next 5 years and the value of AI in clinical care areas. Specifically, female surgeons envisioned a smaller role for AI in the future with less clinical value than male counterparts. This was echoed by young surgeons in comparison to trainees or senior surgeons. Cardiac surgeons saw greater value in AI utilization than general thoracic surgeons and anticipated a larger role for AI in patient evaluation and post-operative management. Generally, physicians felt that regulation of AI technology should come from professional societies and federal government, not state governments, hospitals, or practices. Surgeons supported a proactive approach to the integration of AI into cardiothoracic surgery. The majority endorsed cardiothoracic society engagement with industry (210; 76%), as well as curated literature (210; 76%) and database (184; 68%) sharing, enabling improved AI quality and reliability.

Conclusion: As artificial intelligence is further implemented in medicine, cardiothoracic surgery faces a changing landscape. Expectedly, feelings among physicians are mixed. There are, however, clear areas of agreement as to the benefit of AI integration and support for proactive societal approaches to integration.

Identify the source of the funding for this research project: No external funding. Internal departmental funding for survey lab involvement.