Using Artificial Intelligence to Guide Fluid Therapy During Major Cancer Surgery: A Randomized Controlled Trial (FOCUS-AFM)

National Cancer Institute, Naples

Status: Not yet enrolling

Conditions: Neoplasm; Hemodynamic (MAP) Stability; Fluid Therapy DURING SURGERY; Cancer Surgery; Cancer; Artificial Intelligence (AI)

Treatments: Other: Standard fluid management; Other: Artificial Intelligence-Assisted Fluid Management

Study type: Interventional

Funder types: Other

Identifiers

NCT07314853

FOCUS AFM

Details and patient eligibility

About

The goal of this clinical trial is to learn if using artificial intelligence to guide intravenous fluid therapy during major cancer surgery can help keep blood pressure more stable compared with usual care in adult patients undergoing major cancer surgery. The main questions it aims to answer are:

• Does artificial intelligence-guided fluid therapy reduce hypotensive events during surgery?
• Does this approach improve recovery and reduce complications after major cancer surgery?

Researchers will compare artificial intelligence-guided fluid therapy with standard fluid management to see if the artificial intelligence-guided approach provides better support during surgery.

Participants will:

• Undergo major cancer surgery under general anesthesia
• Receive either artificial intelligence-guided fluid management or standard fluid management during surgery
• Be monitored during and after surgery as part of routine clinical care
• Be followed after surgery to assess recovery and possible complications

Full description

This multicentre randomized controlled clinical trial evaluates the impact of artificial intelligence-guided intraoperative intravenous fluid therapy on hemodynamic management and postoperative outcomes in adult patients undergoing major abdominal cancer surgery.

Optimal intraoperative fluid therapy is a critical component of anesthetic management during major oncologic surgery. Both inadequate and excessive fluid administration may contribute to hemodynamic instability and postoperative complications. Episodes of intraoperative hypotension have been consistently associated with increased postoperative morbidi ty and mortality, particularly in high-risk surgical populations. Although goal-directed strategies for intraoperative fluid therapy have been proposed, their implementation in routine clinical practice remains heterogeneous and highly operator-dependent.

Advances in artificial intelligence have enabled the development of decision support systems capable of integrating continuous hemodynamic data derived from standard intraoperative monitoring. These systems are designed to assist clinicians by analyzing multiple physiologic variables in real time and providing recommendations for intravenous fluid administration aimed at supporting circulatory stability, while preserving full clinician control over treatment decisions.

In this trial, participants undergoing major abdominal cancer surgery under general anesthesia are randomly assigned to receive either artificial intelligence-guided intravenous fluid therapy or standard intravenous fluid management according to routine clinical practice. Randomization is centralized and stratified by relevant procedural factors. In the intervention group, intravenous fluid administration is supported by an artificial intelligence-based decision support system that continuously analyzes intraoperative hemodynamic data and generates recommendations for fluid challenges. Clinicians are strongly encouraged to follow the system recommendations; however, they retain full responsibility and may accept or override these recommendations based on their clinical judgment. In the control group, intraoperative fluid therapy is managed according to usual clinical practice without artificial intelligence guidance.

Standard perioperative monitoring is applied in both study groups, including continuous invasive arterial blood pressure monitoring. Intraoperative hemodynamic variables, fluid administration, and use of vasoactive medications are recorded prospectively using electronic anesthesia records and monitoring system outputs. Postoperative clinical data are collected during routine inpatient care and scheduled follow-up.

The study focuses on the intraoperative period as a key window during which hemodynamic management may influence postoperative recovery and longer-term outcomes. By evaluating an artificial intelligence-based decision support approach in a randomized multicentre setting, this trial aims to generate evidence on whether technology-assisted intravenous fluid therapy can improve intraoperative management and support better clinical outcomes in patients undergoing major cancer surgery.

Enrollment: 176 estimated patients

Sex: All

Ages: 18+ years old

Volunteers: No Healthy Volunteers

Inclusion criteria

• Age ≥ 18 years
• American Society of Anesthesiologists (ASA) Physical Status II-IV
• Undergoing major abdominal oncologic surgery (open or laparoscopic) with an expected duration >2 hours
• Planned invasive arterial pressure monitoring
• Ability to understand and sign informed consent

Exclusion criteria

• Significant arrhythmias (e.g., persistent atrial fibrillation)
• Severe aortic stenosis
• Emergency surgery
• Sepsis
• End-stage renal disease on dialysis
• Pregnancy
• Impossibility to cannulate the radial artery
• Refusal to participate or refusal of data processing consent

Trial design

Primary purpose: Treatment

Allocation: Randomized

Interventional model: Parallel Assignment

Masking: None (Open label)

---

176 participants in 2 patient groups

Standard Fluid Management

Active Comparator group

Description:

Participants receive intraoperative intravenous fluid therapy managed according to usual clinical practice, without artificial intelligence guidance. Fluid administration is determined by the attending clinician based on standard monitoring and clinical judgment.

Treatment:

Other: Standard fluid management

AI-Assisted Fluid Management

Experimental group

Description:

Participants assigned to this arm receive intraoperative intravenous fluid therapy supported by an artificial intelligence-based clinical decision support system. The system analyzes real-time hemodynamic data and provides recommendations for fluid administration. Clinicians are strongly encouraged to follow these recommendations but may accept or override them based on clinical judgment.

Treatment:

Other: Artificial Intelligence-Assisted Fluid Management