Work

A Survey on Perspectives Toward Artificial Intelligence Among Italian Interventional Cardiologists

Year: 2026

Type: article

Abstract: Background: Artificial intelligence (AI) is increasingly being integrated into cardiovascular medicine, with potential applications across image analysis, procedural planning, risk stratification, decision support, and workflow optimization. However, its adoption in interventional cardiology remains heterogeneous and may be influenced by several factors. We aimed to conduct a nationwide survey to assess attitudes towards AI among Italian interventional cardiologists. Methods: We conducted a nationwide, cross-sectional, web-based survey of Italian interventional cardiologists. A structured questionnaire collected information on professional characteristics, familiarity with and current use of AI, perceived clinical applications, expected benefits, trust, implementation barriers, and training needs. Conditional branching was used to obtain additional details from respondents who reported current use of AI-based tools, while all responses were collected voluntarily and analyzed in anonymized, aggregate form. Categorical variables and Likert-scale responses were summarized using descriptive statistics, with exploratory comparisons performed across prespecified professional and institutional subgroups. Results: Among 129 respondents, 70.5% reported at least moderate familiarity with AI and 77.5% reported some current use, although only 60.5% reported regular or occasional professional use, and applications were concentrated mainly in research, education, and information synthesis rather than direct procedural support. Nearly half (48.1%) expected AI to become standard in many procedures within 5 years, while 69.0% anticipated either routine use or particular value in complex cases. Attitudes were broadly favorable, with 85.3% agreeing that AI could improve diagnostic and procedural precision, 86.8% expressing strong interest in future use, and 76.7% stating that AI should support rather than replace physician judgment. The leading barriers were medico-legal uncertainty (45.0%), poor integration with existing clinical systems (34.9%), and cultural resistance or operator distrust (29.5%), whereas preservation of physician control was the most frequently cited requirement for adoption (58.9%). Greater AI familiarity was independently associated with current AI use (p < 0.001) and good or high trust (p < 0.001). Compared with no prior training, one and multiple AI training experiences were independently associated with good or high familiarity (both p < 0.05). Conclusions: Italian interventional cardiologists showed substantial exposure to AI, strong interest in future adoption, and generally favorable expectations regarding its contribution to diagnostic precision, workflow, and procedural support. Acceptance remained conditional on physician oversight, stronger clinical validation, reliable interoperability, and clear medico-legal governance, and previous AI-focused education appeared independently associated with greater familiarity.

Source: Journal of Clinical Medicine

DOI: 10.3390/jcm15187050