Systematic review of transcranial and endoscopic endonasal approaches for craniopharyngiomas in children: is there an evolution?

OBJECTIVE The optimal surgical approach for pediatric craniopharyngiomas (CPs) remains a matter of debate, with selection bias classically precluding a fair comparison of outcomes between the transcranial approach (TCA) and endoscopic endonasal approach (EEA). The purpose of this systematic review was to analyze the current role of EEA in the treatment of pediatric CPs and to determine whether, upon expansion of its indications, a comparison with TCA is valid. METHODS A systematic review of English-language articles published between February 2010 and June 2022 was performed to identify studies in the MEDLINE (PubMed) and Embase databases reporting on the resection of pediatric CPs. Included were articles reporting on pediatric CPs removed through TCA or EEA. Case reports, review articles, and earlier or less comprehensive series by the same center were excluded. Baseline characteristics and outcomes were analyzed. Prediction intervals (PIs), heterogeneity (Q, I-2, and tau(2) statistics), and publication bias (funnel plot analysis) were assessed. RESULTS A total of 835 patients underwent TCA (18 articles) and 403 patients underwent EEA (19 articles). Preoperatively, the mean patient age (p = 0.055, PI = 5.05-15.11), visual impairment (p = 0.08, PI = 19.1-90.5, I-2 = 80%), and hypothalamic syndrome (p = 0.17, PI = 6.5-52.2, I-2 = 62%) did not significantly differ between the EEA and TCA groups. Endocrine deficit (anterior pituitary deficit [p < 0.001, PI = 16.5-92.9, I-2 = 81%] and diabetes insipidus [p < 0.001, PI = 6.3-60.6, I-2 = 43%]) was more frequent in the EEA group. Hydrocephalus and signs/symptoms of raised intracranial pressure were significantly higher (p < 0.001, PI = 5.2-73.3, I-2 = 70% vs p < 0.001, PI = 4.6-73, I-2 = 62%, respectively) in the TCA group. Recurrent lesions (p = 0.52, PI = 2.7-87.3, I-2 = 13%), tumor size (p = 0.25, PI = 22.1-56.8), third ventricle involvement (p = 0.053, PI = 10.9-81.3, I-2 = 69%), and hypothalamic involvement (p = 0.06, PI = 8.5-83.6, I-2 = 79%) did not differ significantly between the approaches. EEA was preferred (p = 0.006, PI = 26.8-70.8, I-2 = 40%) for sellar-suprasellar CPs, whereas TCA was preferred for purely suprasellar CPs (p = 0.007, PI = 13.5-81.1, I-2 = 61%). There was no difference between the approaches for purely intrasellar lesions (p = 0.94, PI = 0-62.7, I-2 = 26%). The breadth of PIs, I-2 values, and analysis of publication bias showed substantial variability among the pooled data, hindering the possibility of outcome meta-analyses. CONCLUSIONS With the adoption of extended approaches, the use of EEA became appropriate for a wider spectrum of pediatric CPs, with associated excellent outcomes. Although a fair comparison between outcomes in the EEA and TCA groups was hindered because of the differences in patient populations and tumor subtypes, given the increased versatility of EEA and improved expertise in its use, surgeons can now select the optimal surgical approach based on the unique benefits and drawbacks of each pediatric CP.