Comparison of various predictive tools in predicting risk of cerebrospinal fluid diversion post-resection of posterior fossa tumors: A systematic review.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
BACKGROUND & OBJECTIVE: Posterior fossa tumors (PFTs) frequently cause hydrocephalus (HCP), requiring permanent cerebrospinal fluid (CSF) diversion post resection in both pediatric and adult patients. We aimed to compare the performance of various tools in predicting the risk of postoperative hydrocephalus and improving prediction for better neurosurgical decision-making. METHODOLOGY: A comprehensive literature search was conducted across Google Scholar and PubMed database adhering to Preferred Reporting Items for Systematic Review and Meta-analyses (PRISMA) guidelines using keywords such as posterior fossa tumors, hydrocephalus, CSF diversion, and predictive models. A total of ten original articles with a sample size of 1597 from 2021 to 2025 were selected for data extraction. Study quality evaluation was executed via PROBAST tool. RESULTS: The cumulative mean ages were 7.18±1.64 years for pediatric patients and 53.07±0.81 years for adults. Pediatric patients accounted for 67.31% (1075) patients while adults accounted for 19.41% (402) patients. Preoperative hydrocephalus was present in 52.4% (833) patients, out of which 22.9% (367) required CSF diversion post resection. Pooled post-operative shunt rates revealed higher shunt rate 53.9% (715) in patients with preoperative hydrocephalus than those without 13% (80). Logistic regression was used in 88.8% (8) of the identified models while AI based model demonstrated best performance (AUC = 0.938). CONCLUSION: Of all the predictive models developed, till now, to predict the need for CSF diversion after PFT resection, artificial intelligence-based model shows superior accuracy for improving hydrocephalus risk prognostication. Apart from clinical, demographic, surgery-related and radiological predictors incorporated in conventional predictive models, the artificial-intelligence based model improves risk prediction by utilizing complex patterns in intraoperative and postoperative imaging of patients.