Neuroradiological Findings in Central Nervous System Tuberculosis: A Systematic Review and Meta-analysis Comparing Adult and Pediatric Populations

BACKGROUND: Neurotuberculosis is a severe form of extrapulmonary tuberculosis causing significant morbidity and mortality. PURPOSE: To compare the imaging phenotypes of adult and pediatric populations with CNS tuberculosis through a quantitative meta-analysis, and to characterize the broader spectrum of neurotuberculosis radiological findings through a systematic review emphasizing standard and advanced imaging techniques. DATA SOURCES: A systematic search was conducted across Embase, LILACS, PubMed, and Scopus databases. STUDY SELECTION: Following PRISMA guidelines, observational studies detailing CT and MRI findings in CNS tuberculosis were included. DATA ANALYSIS: Screening and data extraction were conducted using Rayyan and R software. Random-effects meta-analyses calculated pooled prevalence estimates (95% CI) stratified by age (adult vs. pediatric). DATA SYNTHESIS: From 91 studies (5,955 patients) in the systematic review, 32 low-risk-of-bias studies (1,675 patients: 24 adult, 1,415 patients; 8 pediatric, 260 patients) were meta-analyzed. Weighted mean age (by study sample size) was approximately 32.0 years in adults (23/24 studies reporting age; N=1,301) and 6.0 years in children (8/8 studies; N=260). Pooled prevalence estimates were numerically higher in pediatric than adult cohorts for meningovascular complications, including leptomeningeal enhancement (89% vs. 61%), hydrocephalus (81% vs. 36%), and brain infarctions (65% vs. 33%, with greater proportional cortical involvement). Tuberculomas showed similar prevalence (∼60%, predominantly multiple). Advanced MRI modalities (DWI, 1H-MRS, vessel wall imaging) provided detailed lesion characterization. Spinal involvement was evaluated almost exclusively in adults, precluding a quantitative comparative meta-analysis due to sparse pediatric spinal data. LIMITATIONS: High heterogeneity across primary studies, geographic concentration in endemic regions, and insufficient pediatric spinal data. CONCLUSIONS: Neurotuberculosis imaging phenotypes differed numerically by age group, with pediatric cases showing a higher pooled burden of meningovascular complications. While intracranial age-stratified differences were established, comparative spinal analysis was limited by sparse pediatric data.