Performance of a clinical chatbot in the interpretation of coronary angiography: a comparative analysis with interventional cardiologists
DOI:
https://doi.org/10.59681/2175-4411.v18.2026.1577Keywords:
Cardiology, Cardiovascular Health, Biomedical TechnologyAbstract
Objective: To investigate the accuracy and predictive capacity of an AI chatbot in formulating clinical management based on coronary angiography results, comparing its recommendations to those of interventional cardiologists. Method: A comparative, mixed-methods study was conducted in the hemodynamics sector in the state of Paraíba, Brazil. Fifteen interventional cardiologists were selected to evaluate three clinical cases constructed from coronary angiography (CAG) examinations of patients with ACS. Results: Agreement was predominantly observed in questions Q1 (92%) and Q3 (77%), indicating good acceptance of the procedures in more linear scenarios aligned with protocols. In Q2, 54% of participants disagreed with the procedure, highlighting a greater gap between the AI and the consensus of experts in more complex revascularization decisions. Conclusion: The chatbot showed overall satisfactory performance in selected situations, and can act as complementary support for the interventional cardiologist and as a training resource in extension activities.
Downloads
References
Brasil. Ministério da Saúde. Cerca de 400 mil pessoas morreram em 2022 no Brasil por problemas cardiovasculares. Brasília, DF: Ministério da Saúde; 2023. Disponível em: https://bvsms.saude.gov.br/cerca-de-400-mil-pessoas-morreram-em-2022-no-brasil-prproblemas-cardiovasculares/. Acesso em: 25 abr. 2025.
Ferreira AP, Silva ER, Pereira EC de O. Fatores sociodemográficos e comportamentais associados ao risco cardiovascular em adultos: estudo de base populacional em Campinas, São Paulo, Brasil. Rev Bras Epidemiol. 2021;24:e210008. DOI: 10.1590/1980-549720210013.supl.2. Acesso em: 17 jun. 2025.
World Health Organization. Cardiovascular diseases (CVDs) [Internet]. Geneva: World Health Organization; 2025. Disponível em:https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds). Acesso em: 17 jun. 2025.
Agência Brasil. Doenças cardiovasculares matam 400 mil brasileiros por ano. Brasília (DF): Agência Brasil; 29 set. 2024. Disponível em: https://agenciabrasil.ebc.com.br/saude/noticia/2024-09/doencas-cardiovasculares-matam400-mil-brasileiros-por-ano. Acesso em: 25 abr. 2025.
Cerci, R. J. et al. O Impacto da COVID-19 no Diagnóstico de Doenças Cardíacas na América Latina Uma Subanálise do INCAPS COVID. Arquivos Brasileiros de Cardiologia, [s. l.], v. 118, n. 4, p. 745–753, abr. 2022. Disponível em: https://www.scielo.br/j/abc/a/4x8TTTyVbDmhdpsQWYdpFKK/. Acesso em: 17 jun. 2025.
Martins, Victor Balceiro Legname; THOM, Laysla Rangel Freitas; GONÇALVES, Júlia Mayse Soares; DESTEFANI, Afrânio Côgo. Avanços da inteligência artificial na cardiologia: uma revisão abrangente. Revista Ibero-Americana de Humanidades, Ciências e Educação – REASE, São Paulo, v. 10, n. 4, p. 2442–2453, abr. 2024. Disponível em: https://periodicorease.pro.br/rease/article/view/13643/6689. Acesso em: 25 abr. 2025.
SANTOS, Maria Aparecida; OLIVEIRA, João Carlos. O impacto da inteligência artificial na educação contemporânea. Revista Apoena, Salvador, v. 5, n. 2, p. 45–60, jul./dez. 2023. Disponível em: https://publicacoes.unijorge.com.br/apoena/article/view/396/364. Acesso em: 25 abr. 2025.
Genaro LE, Marconato JV, Pinotti FE, Valsecki Júnior A, Adas Saliba T, Lopez Rosell F. Discurso do sujeito coletivo: bases conceituais, enfoques e aplicações em pesquisas qualitativas. Rev. Pesq. Qual. [Internet]. 14º de outubro de 2024 [citado 12º de dezembro de 2025];12(32):553-69. Disponível em: https://editora.sepq.org.br/rpq/article/view/739
Bancovsky E. Acurácia do algoritmo de inteligência artificial para auditoria de qualidade de exames de ultrassonografia abdominal [tese]. São Paulo: Universidade Federal de São Paulo, Escola Paulista de Medicina; 2022. 54 f.
Melo GBP, Araujo IBB, Guedes GP, Alves RN, Requeijo MJR. Um comparativo entre a inteligência artificial e os médicos radiologistas no campo da interpretação de exames de imagens. Research, Society and Development. 2024;13(11):e125131147411.DOI:10.33448/rsd-v13i1.47411.
Roski J, Chapman W, Heffner J, Patel R, Weinstein J, Zink A. How artificial intelligence is changing health and health care. In: Whicher D, Ahmed M, Israni ST, et al., editors. Artificial Intelligence in Health Care: The Hope, the Hype, the Promise, the Peril [Internet]. Washington (DC): National Academies Press (US); 2023 Aug 2 [cited 2025 Dec 12]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK605950/
Harskamp, R. E.; De Clercq, L. Performance of ChatGPT as an AI-assisted decision support tool in medicine: a proof-of-concept study for interpreting symptoms and management of common cardiac conditions (AMSTELHEART-2). Acta Cardiologica, v. 79, n. 3, p. 358-366, 2024. DOI: https://doi.org/10.1080/00015385.2024.2303528.
Yavuz, Y. E.; Kahraman, F. Evaluation of the prediagnosis and management of ChatGPT-4.0 in clinical cases in cardiology. Future Cardiology, v. 20, n. 4, p. 197-207, 2024. DOI: https://doi.org/10.1080/14796678.2024.2348898.
Baladrón, C. et al. Assessing the accuracy of ChatGPT as a decision support tool in cardiology. Revista Española de Cardiología (English Edition), v. 77, n. 5, p. 433-435, 2024. DOI: https://doi.org/10.1016/j.rec.2023.11.011.
Fernández-Cisnal, A. et al. Performance of an artificial intelligence chatbot with web search capability in cardiology-related assistance: a simulation study. Revista Española de Cardiología (English Edition), v. 76, n. 12, p. 1065-1067, 2023. DOI: https://doi.org/10.1016/j.rec.2023.06.008.
Monroe, C. L. et al. Evaluation of responses to cardiac imaging questions by the artificial intelligence large language model ChatGPT. Clinical Imaging, v. 112, p. 110193, 2024. DOI: https://doi.org/10.1016/j.clinimag.2024.110193.
Mola, S.; Yildirim, A.; Gül, E. B. Artificial intelligence in cardiac treatment decision-making: an evaluation of the performance of ChatGPT versus the heart team in coronary revascularization. Reviews in Cardiovascular Medicine, v. 26, n. 8, p. 38705, 2025. DOI: https://doi.org/10.31083/RCM38705.
Sudri, K. et al. Enhancing coronary revascularization decisions: the promising role of large language models as a decision-support tool for multidisciplinary heart team. Circulation: Cardiovascular Interventions, v. 17, n. 11, p. e014201, 2024. DOI: https://doi.org/10.1161/CIRCINTERVENTIONS.124.014201.
Jiang, S. et al. Knowledge, attitudes, and practices of cardiovascular health care personnel regarding coronary CTA and AI-assisted diagnosis: a cross-sectional study. Journal of Global Health, v. 15, p. 04103, 2025. DOI: https://doi.org/10.7189/jogh.15.04103.
Ozkan, E. et al. Global health care professionals’ perceptions of large language model use in practice: cross-sectional survey study. JMIR Medical Education, v. 11, p. e58801, 2025. DOI: https://doi.org/10.2196/58801.
Algoraini, Y. et al. Perceptions of large language models in medical education and clinical practice among pediatric emergency physicians in Saudi Arabia: a multiregional cross-sectional study. Frontiers in Public Health, v. 13, p. 1634638, 2025. DOI: https://doi.org/10.3389/fpubh.2025.1634638.
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Josivan Soares Alves Júnior, Thayse Mota Alves, Debora Regina Alves Raposo, Alex Junior Vieira Sousa, Josué Luis Pereira Negreiros, Larissa Gomes Freire, Evely Laís Valença Melo, Cosme Michael Santos Farias
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Submission of a paper to Journal of Health Informatics is understood to imply that it is not being considered for publication elsewhere and that the author(s) permission to publish his/her (their) article(s) in this Journal implies the exclusive authorization of the publishers to deal with all issues concerning the copyright therein. Upon the submission of an article, authors will be asked to sign a Copyright Notice. Acceptance of the agreement will ensure the widest possible dissemination of information. An e-mail will be sent to the corresponding author confirming receipt of the manuscript and acceptance of the agreement.
