Exploring Radiologists’ Perceptions and Experiences Regarding Integration of Artificial Intelligence in Diagnostic Imaging Practices

Abstract

Background: Artificial intelligence has rapidly entered diagnostic radiology, offering opportunities to enhance workflow efficiency, support interpretation, and strengthen clinical decision-making. Despite these advancements, radiologists’ real-world experiences and perceptions remain central to understanding how effectively AI systems integrate into daily practice. Exploring these perspectives is essential for developing safe, practical, and clinician-aligned implementation strategies. Objective: To qualitatively explore radiologists’ perspectives, perceived benefits, and challenges regarding the integration of AI technologies in diagnostic radiology. Methods: A qualitative study was conducted over four months in South Punjab, involving twelve practicing radiologists selected through purposive sampling. Semi-structured interviews were used to collect data, focusing on experiences with AI-enabled imaging tools. Transcribed interviews were analyzed using thematic analysis supported by a structured coding framework. Descriptive statistics were applied to summarize demographic variables, and normality was confirmed through the Shapiro–Wilk test. Results: Participants reported noticeable improvements in workflow efficiency and reporting timeliness, with high mean scores in both areas. Diagnostic support was viewed positively, whereas error reduction received moderate ratings. Challenges centered around trust, inconsistent system performance, and integration issues, which were reflected in higher challenge scores. Adoption likelihood varied among participants, with five radiologists demonstrating high confidence, four moderate confidence, and three low readiness for long-term AI adoption. Experiences indicated that familiarity with AI tools strongly influenced acceptance, while technical concerns and medicolegal uncertainties contributed to caution. Conclusion: The study found that radiologists acknowledged AI as a supportive tool that improves workflow and enhances diagnostic processes, yet meaningful concerns about reliability, trust, and integration persisted. These findings emphasize that successful AI adoption requires balanced implementation, structured training, and ongoing evaluation to maintain confidence and ensure safe integration into radiological practice