Evaluating ChatGPT's Role in Interventional Cardiology: Assessing Responses Related to CAD and STEMI in PAD Patients

Vivie Tran, Hoang Ho, Steven Daley, Simon Williams, MD, Aliakbar Arvandi, MD, Mohammad M. Ansari, MD

PAD Center of Excellence, Texas Tech University Health Sciences Center, Lubbock, TX, USA

Background

With ChatGPT's recent rise, many have sought to explore the relevance of this technology in medicine. We seek to investigate the potential applicability of ChatGPT in interventional cardiology by assessing the appropriateness of its responses to fundamental questions pertaining to CAD and STEMI in PAD patients.

Methods

For the purpose of this study, 20 questions centered around CAD and STEMI were asked to ChatGPT. Inquires covered areas such as risk factor counseling, prevention, test result analysis, and treatment, including medications and PCI. All responses underwent thorough evaluation by a panel of 3 experienced interventional cardiologists that also perform vascular procedures Experts examined the responses based on clinical knowledge, classifying them as appropriate, inappropriate, or unreliable. Two teams were created and results were combined and analyzed for correctness.

Results

ChatGPT delivered appropriate responses for 13/15 questions (86.7%). No responses from the AI software were classified as inappropriate. In 2/15 questions (13.3%), responses were considered unreliable due to a lack of comprehensive detail surrounding the topic asked. 

Conclusion

Our study demonstrated ChatGPT’s capacity to offer mostly suitable answers to inquiries related to CAD and STEMI. Given the continual progress in AI, the utilization of ChatGPT may benefit patient education and experience over complex pathologies in cardiology. Additional investigations are essential to gain a comprehensive understanding of ChatGPT's role. Our next step is to assess the knowledge and incidence of acute limb ischemia in CAD patients.