Abstract BACKGROUND Brainstem tumors represent ~10% of pediatric brain tumors, ~80% of these are diffuse intrapontine gliomas (DIPG, primarily WHO Grade 4 midline glioma, H3 K27M-mutant). Given invariably poor prognosis in these cases, there continues to be immense variation (across the US and internationally) in performing biopsy of these lesions as opposed to making a presumptive diagnosis based on MRI. Several multi-center studies in recent years have aimed to show the safety profile of biopsy but an updated meta-analysis is lacking. METHODS We found 29 studies of pediatric brainstem biopsy in the last 20 years (2003-2023), representing data from 1002 children. We applied meta-analysis of proportions using a random-effects model to generate point estimates, confidence intervals, and measures of heterogeneity. RESULTS 87% of procedures were stereotactic needle biopsies (of these, 62% with a frame, 14% without frame, and 24% robotic.) Biopsy yield (i.e. the rate at which biopsy resulted in a tissue diagnosis) was 94% (95% CI 92-97). Temporary complications were seen in 6% (95 CI 4-8), most commonly including 1) radiographic evidence of bleeding or air without an associated fixed neurological deficit, 2) cranial nerve dysfunction, and 3) delayed awakening. Permanent complications (excluding death) were seen in 1% (95% CI 0.5-2), most commonly including cranial nerve dysfunction and limb weakness. 5 deaths were reported in the entire pooled cohort of 1002 children (0.5%). Gliomas represented 92% of tumors (95% CI 89-95). Of the gliomas for which WHO grading was provided, 68% were high grade (III/IV). CONCLUSION When counseling families on the merits of brainstem biopsy in children, it is reasonable to state that permanent morbidity is rare (< 2%). If biopsy is performed specifically to facilitate enrollment in clinical trials requiring a molecular diagnosis, the risks of biopsy outlined here should be weighed against potential benefits of trial enrollment.
Read full abstract