Transient and permanent hydrocephalus following resection of brain metastases located in the posterior fossa: incidence, risk factors and the necessity of perioperative external ventricular drainage placement.

PURPOSE: Prophylactic insertion of an external ventricular drainage (EVD) prior to the resection of posterior fossa metastases (PFMs) is a common approach to address postoperative transient and permanent hydrocephalus. However, predicting surgery-related hydrocephalus in the preoperative phase continues to be a challenge. This study aims to analyze the incidence, preoperatively collectable risk factors and necessity of perioperative external ventricular drainage placement after posterior fossa metastasis surgery.

METHODS

All patients undergoing surgery for PFMs at the authors' neuro-oncological center between 2015 and 2021 were identified and assessed for postoperative hydrocephalus occurrence. Tumour volume, edema volume, and 4th ventricle volume were assessed on preoperative magnetic resonance imaging scans using the IntelliSpace Portal 5.0. A multivariable logistic regression analysis was performed to identify possible predictors for postoperative hydrocephalus occurrence.

RESULTS

Postoperative hydrocephalus occurred in 14 of the 130 identified PFM patients (11%). Multivariable analysis and receiver operating characteristic (ROC) analysis revealed a 4th -ventricle-to-tumor-volume ratio ≤ 0.02 (OR 33.1, 95% CI 3.8-284.3, p = 0.001), an edema-to- tumor-volume ratio ≤ 0.85 (OR 10.6, 95% CI 2.4-47.4, p = 0.002), an imaging-morphological contact to the 4th ventricle (OR 5, 95% CI 1.4-18, p = 0.013), and multiple intracranial metastases (OR 2.4, 95% CI 1-5.9, p = 0.045) as independent predictors for surgery-related postoperative hydrocephalus occurrence.

CONCLUSION

The present study identifies preoperatively detectable risk factors for the occurrence of postoperative hydrocephalus following surgery for PFMs. These findings may provide guidance in clinical decision-making regarding prophylactic EVD placement.