Open AccessArticle
Chat GPT Performance in Multi-Disciplinary Boards—Should AI Be a Member of Cancer Boards?
by
Ibrahim Dogan, Mehmet Kadir Bartin, Ezgi Sonmez, Erdogan Seyran, Halil Alper Bozkurt, Mehmet Yuksek, Ezgi Dicle Serbes, Gunel Zalova and Sebahattin Celik
Healthcare 2025, 13(18), 2254; https://doi.org/10.3390/healthcare13182254 (registering DOI) - 9 Sep 2025
Abstract
Background: Multidisciplinary Tumor Councils (MDTs) are vital platforms that provide tailored treatment plans for cancer patients by combining expertise from various medical disciplines. Recently, Artificial Intelligence (AI) tools have been investigated as decision-support systems within these councils. Methods: In this prospective study, the
[...] Read more.
Background: Multidisciplinary Tumor Councils (MDTs) are vital platforms that provide tailored treatment plans for cancer patients by combining expertise from various medical disciplines. Recently, Artificial Intelligence (AI) tools have been investigated as decision-support systems within these councils. Methods: In this prospective study, the compatibility of AI (ChatGPT-4.0) with MDT decisions was evaluated in 100 cancer patients presented to the tumor council between November 2024 and January 2025. AI-generated treatment recommendations based on anonymized, detailed clinical summaries were compared with real-time MDT decisions. Cohen’s Kappa and Spearman correlation tests were used for statistical analysis. Results: Neoadjuvant treatment (45%) and surgery (36%) were the most frequent MDT decisions. AI recommended surgery (39%) and neoadjuvant treatment (37%) most frequently. A high concordance rate of 76.4% was observed between AI and MDT decisions (κ = 0.764 [95% CI; 0.658–0.870]
p < 0.001, ρ = 0.810 [95% CI; 0.729–0.868],
p < 0.001). Most inconsistencies arose in cases requiring individualized decisions, indicating AI’s current limitations in incorporating contextual clinical judgment. Conclusion: AI demonstrates substantial agreement with MDT decisions, particularly in cases adhering to standardized oncological guidelines. However, for AI integration into clinical workflows, it must evolve to interpret real-time patient data and function transparently within ethical and legal frameworks.
Full article
►▼
Show Figures
