Implementation and learning curve in AI-assisted fluid management during abdominal oncologic surgery: a retrospective...

Area of Science:

• Anesthesiology
• Artificial Intelligence in Medicine
• Surgical Oncology

Background:

• Intraoperative fluid management in major abdominal oncologic surgery is complex and operator-dependent.
• Assisted Fluid Management (AFM) is an AI decision support system using real-time Stroke Volume (SV) analysis.
• Limited data exist on AFM adoption and clinician interaction evolution in clinical practice.

Purpose of the Study:

• To evaluate the adoption and learning curve of an AI-based Assisted Fluid Management (AFM) system in major abdominal oncologic surgery.
• To assess changes in clinician fluid management behavior and the physiological effectiveness of fluid challenges over time.
• To determine the impact of AFM integration on Stroke Volume (SV) response.

Main Methods:

• Retrospective observational study using a prospectively maintained institutional database.
• Included adult patients undergoing major abdominal oncologic surgery with intraoperative AFM monitoring.
• Compared two time periods post-AFM implementation, analyzing fluid challenge initiation, hemodynamic effectiveness (SV response), and bolus characteristics.

Main Results:

• Analysis of 404 fluid challenges in 59 patients.
• Significant decrease in clinician-initiated boluses and increase in AFM-suggested fluid challenges over time (p < 0.001).
• AFM-suggested boluses showed improved effectiveness and ΔSV over time (p < 0.05), indicating a learning curve effect.

Conclusions:

• Progressive integration of AFM into anesthetic practice correlates with behavioral changes and improved physiological effectiveness of fluid management.
• AI-based decision support systems can enhance consistency and physiological targeting in fluid management.
• Findings support further prospective studies on AFM's impact on clinical outcomes.