Anticipating complications in stereotactic brain biopsy: a predictive approach

Stereotactic brain biopsy is a critical procedure in neurosurgery, particularly for the diagnosis of brain tumors and cryptogenic neurological diseases. Despite its safety profile, biopsy procedures carry a risk of complications. This study aimed to identify predictors of symptomatic complications in a large cohort and develop a risk prediction score. This retrospective single-center study examined data from 2,338 stereotactic brain biopsies performed over 15 years. The primary outcomes included complication rates, severity, timing, and management. Factors such as patient demographics, medical history, lesion characteristics, and biopsy procedures were analyzed. Predictive models were created using least absolute shrinkage and selection operator (LASSO) regression to select key variables with cross-validation and a random forest algorithm for further refinement. Owing to insufficient predictive performance for clinical use, we used variables selected by LASSO regression to construct an analytical multivariate model. Symptomatic complications occurred in 3.9% (95% confidence interval (CI) 3.1-4.7) of cases, with 0.8% being fatal. Of the symptomatic complications, 46.2% occurred within the first hour following biopsy and 71.4% within two hours. Key predictive factors included biopsy repetition (odds ratio, 3.3; 95%CI [1.1-9.6], p = 0.050), advanced age (1.2 [1.02-1.4], p = 0.048), lesion location (brainstem (4.1 [1.6-10.4], p = 0.004), pineal region (16.2 [3.0-89.4], p = 0.001), deep brain (1.8 [1.1-2.9], p = 0.016)), and toxoplasmosis (4.9 [1.4-17.7], p = 0.038). The best predictive model achieved an area under the curve (AUC) of only 0.64 and the random forest models had poorer discriminative accuracy (AUC < 0.6). Symptomatic complications following stereotactic brain biopsy are rare, but are associated with specific patient profiles. Although predictive modeling provided moderate accuracy, further refinement is necessary for reliable risk stratification. Awareness of high-risk patient characteristics and rigorous procedural planning are essential for minimizing complications. Future studies should explore advanced predictive methods and refine the risk assessment tools to improve patient outcomes.