A systematic review and meta-analysis of the diagnostic accuracy of ultrasound in the diagnosis of paediatric tuberculosis.

Abstract

Diagnosing tuberculosis (TB) in children is challenging due to non-specific symptoms, paucibacillary disease, and difficulty producing sputum. Chest X-rays (CXRs), though widely used, are often inaccessible in low-resource settings and involve radiation. Ultrasound (US) is a radiation-free, portable, and potentially low-cost alternative that can detect pulmonary and extrapulmonary TB features. However, its diagnostic accuracy in paediatric TB remains unclear. This systematic review and meta-analysis assessed US diagnostic performance for paediatric TB across anatomical sites. Following PRISMA-DTA guidelines, we searched five databases through May 2025. Studies were included if they involved children under 15 with presumptive TB and reported US diagnostic accuracy data. Data extraction, quality assessment (QUADAS-2), and meta-analyses using a bivariate random-effects model were conducted. Graham's TB classification served as the reference standard. CXR was used as a comparator where available, with agreement assessed via Cohen's kappa. Of 17 019 records, 7 studies involving 945 children met inclusion criteria. Pooled US sensitivity was 52% (95% CI: 46-58%), and specificity was 76% (95% CI: 67-83%). US showed high specificity but low sensitivity across most features, including abdominal lymphadenopathy and pericardial effusion; pleural effusion had slightly higher sensitivity (18%). Agreement with CXR was moderate (kappa 0.24-0.42). Variability in US protocols, operator skills, and reference standards limited generalizability. Only one study had low risk of bias across all QUADAS-2 domains. US is a promising adjunct for paediatric TB diagnosis in resource-limited settings, but standardization and validation are needed to improve its standalone utility.