To compare MDT-based decision-making with an AI-assisted model in the postoperative management of patients undergoing surgery for pancreatic cancer, focusing on efficiency and workload reduction.
Key Findings:
The overall concordance rate between AI-generated recommendations and MDT decisions was 80%, with a Cohen's kappa coefficient of 0.625, indicating moderate agreement beyond chance. These findings suggest a promising alignment between AI and MDT decisions, warranting further exploration.
Interpretation:
AI-assisted decision-support systems may approximate MDT recommendations in postoperative pancreatic cancer management, but discrepancies highlight the importance of expert clinical judgment and contextual interpretation.
Limitations:
The study is based on a limited sample size of 15 patients for comparison, which may affect the generalizability of the findings. Additionally, the AI model was not fine-tuned or retrained, potentially impacting its performance.
Conclusion:
AI-based models may serve as supportive tools in postoperative clinical decision-making by potentially reducing workload and time burden, but should complement rather than replace multidisciplinary expert evaluation, necessitating further validation.
