Predicting and preventing surgical wound infections after brain or spinal tumor surgery

A Prospective, Single-Center Clinical Validation Study: Machine Learning Model-Based Risk Stratification Intervention for Reducing Surgical Site Infection After Central Nervous System Tumor Surgery

This trial will test whether a machine-learning tool that flags people having brain or spine tumor surgery for extra infection-prevention care can lower their risk of surgical wound infections.

Quick facts

• Phase: Not applicable
• Study type: Interventional
• Enrollment: 500 (estimated)
• Sex: All
• Sponsor: Cancer Institute and Hospital, Chinese Academy of Medical Sciences (Academic / other)
• Locations: 1 site (Beijing, PUMC)
• Trial ID: NCT07378683 on ClinicalTrials.gov

What this trial studies

This interventional study uses an intelligent prediction model to generate individualized risk scores for patients undergoing elective craniotomy or spinal tumor resection and applies an enhanced bundle of preventive measures for those identified as high risk while others receive standard postoperative care. The model is built with machine learning and explainable AI methods (including SHAP) to make risk drivers transparent to clinicians. Eligible adults with pathologically confirmed primary or metastatic CNS tumors will be enrolled at a single center and followed for postoperative surgical site infections. The primary goal is to compare infection rates between model-guided enhanced prevention and standard care.

Who should consider this trial

Good fit: Adults (age ≥18) with pathologically confirmed primary or metastatic brain or spinal tumors scheduled for elective tumor resection who expect to survive more than three months and can consent are the intended participants.

Not a fit: Patients with an active infection before surgery, recent therapeutic antibiotic use, severe immunosuppression, pregnancy or lactation, or contraindications to the planned antibiotics are unlikely to benefit or be eligible for the intervention.

Why it matters

Potential benefit: If successful, using the model to target extra preventive measures could reduce postoperative wound infections and shorten recovery after brain or spine tumor surgery.

How similar studies have performed: The investigator group has previously developed a highly accurate retrospective prediction model, but prospective use of such models to guide preventive interventions in CNS tumor surgery is still largely untested.

Eligibility criteria

Inclusion Criteria:

1. Patients with a pathologically confirmed primary or metastatic central nervous system (brain or spinal) tumor.
2. Scheduled for elective craniotomy or spinal tumor resection surgery.
3. Age ≥ 18 years.
4. Expected survival \> 3 months, and able/willing to comply with postoperative follow-up.
5. Voluntary participation and provision of written informed consent.

Exclusion Criteria:

1. Presence of active systemic or local surgical site infection before surgery.
2. Use of therapeutic antibiotics for any reason within 72 hours prior to surgery.
3. Concurrent severe immunosuppressive conditions, or chronic use of high-dose immunosuppressants.
4. Pregnancy or lactation.
5. Known contraindications or severe allergy to the antibiotics planned for use in the study.
6. Any other condition that, in the investigator's judgment, may affect study participation or outcome assessment.

Where this trial is running

Beijing, PUMC

• National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College — Beijing, Pumc, China (Recruiting)

Study contacts

• Study coordinator: Ming Yang, MD

Conditions Surgical Site Infection; Central Nervous System Tumor; surgical site infection; central nervous system tumor; machine learning; explainable artificial intelligence; SHAP; clinical decision support system