Challenges and advancements in the diagnosis of central nervous system tuberculosis

Central nervous system (CNS) tuberculosis is one of the most severe forms of extrapulmonary TB, often leading to substantial morbidity and mortality owing to its complex clinical presentation and frequent diagnostic delays. Despite global TB control efforts, CNS TB continues to present unique difficulties owing to its nonspecific clinical manifestations, the paucibacillary nature of cerebrospinal fluid (CSF), and the limitations of conventional diagnostic methods. The diagnostic dilemma is further compounded in resource-limited settings where advanced molecular tools and neuroimaging may not be readily available. Timely and accurate diagnosis remains a cornerstone in reducing mortality and neurological sequelae. Clinically, CNS TB can present as tuberculous meningitis (TBM), tuberculoma, abscess, or spinal TB, each with overlapping features with other infectious and non-infectious conditions. This spectrum makes early clinical diagnosis particularly elusive. Traditional diagnostic methods such as CSF smear for acid-fast bacilli (AFB), culture, and routine biochemistry often fail to provide definitive answers. AFB smear sensitivity is dismal (less than 10 %), and culture, although considered a reference standard, takes weeks, during which the disease may progress rapidly. In recent years, advances in molecular diagnostics have brought hope. Nucleic acid amplification tests (NAATs), including GeneXpert MTB/RIF and the more sensitive GeneXpert Ultra, have enhanced detection compared to smear microscopy; however, their diagnostic yield in cerebrospinal fluid (CSF) remains suboptimal due to the typically low bacillary burden in CNS tuberculosis. WHO's 2021 Module 3 guidelines emphasize that despite these advances, NAAT sensitivity for CNS TB remains suboptimal and cannot replace clinical judgment and imaging. CSF Xpert positivity in TBM varies between 50 and 70 %, with a negative result not excluding disease. LAMP (loop-mediated isothermal amplification) and multiplex PCR assays are also emerging but lack standardization for CNS specimens. Neuroimaging techniques like MRI, MR spectroscopy, and diffusion tensor imaging (DTI) play an essential adjunctive role in diagnosis, particularly when CSF results are inconclusive. However, findings are often non-specific and require integration with clinical and laboratory data. Advanced imaging and artificial intelligence (AI)-based radiomic analysis are being explored for pattern recognition in neuro-TB, showing promise but are not yet widely applicable in clinical practice. Biomarkers like adenosine deaminase (ADA) along with interferon gamma, neopterin, and IP-10 in CSF have been studied, but none have yet replaced microbiological confirmation. Furthermore, diagnosis in specific populations-such as extremes of age, people living with HIV, and elderly individuals-presents additional complexities due to atypical presentations and altered immune responses. This review comprehensively discusses the multifaceted challenges in diagnosing CNS TB and the emerging diagnostic tools and strategies. It evaluates the sensitivity, specificity, and utility of conventional and advanced modalities and highlights evidence from WHO guidelines and recent literature. Addressing diagnostic barriers and integrating novel technologies into practice, particularly in high-burden countries, is essential to improving outcomes.