Diagnostic Utility of Mycobacterium Tuberculosis Cell-free DNA

Tuberculosis (TB) remains a key infectious disease burden globally. According to the Global TB Report by the World Health Organization (WHO) in 2023, TB is the second leading infectious cause of death after COVID-19 in 2022. Extrapulmonary TB, amongst which tuberculous pleuritis (TBP) is one of the most common subtypes, poses additional obstacles to global TB control due to its difficulty in diagnosis. There is a large geographical variation in the incidence of TBP, with the proportion of TBP among all TB patients ranging from 2.2% to 31.4%. An increasing incidence of TBP was noted in two large retrospective Chinese cohort studies.

The paucibacillary nature of TBP leads to its diagnostic challenge. Despite being a gold standard diagnostic test, the diagnostic sensitivity of pleural fluid Mycobacterium tuberculosis (MTB) culture is suboptimal. It varies between 7.0% and 75.0%, depending on the culture medium and HIV status. Its long turnaround time of around 4 to 8 weeks also impairs its clinical practicality. Commercial PCR techniques, such as Xpert MTB/RIF or Xpert Ultra, are rapid and specific tests with positive results. However, their utilities are limited by suboptimal sensitivity. Pleural biopsy for histology and combined histology / culture might raise the sensitivity to 66% and 79%, but this procedure is more invasive than thoracentesis. Pleural fluid adenosine deaminase (ADA) is the most commonly studied biomarker, with diagnostic sensitivity and specificity of 0.93 and 0.90 at a cutoff of 40 U/L. Its cutoff range may vary with patient characteristics (age, comorbid illness). The difficulty in diagnosing TBP, therefore, leads to delayed initiation of anti-TB treatment. Empirical anti-TB treatment may be initiated based on compatible clinical presentations without confirmatory diagnostic microbiological results, bearing the risk of treatment toxicities (e.g. hepatotoxicity). This suggests an unmet clinical need for a better diagnostic tool for TBP.

Detecting MTB cell-free DNA (cfDNA) in the pleural fluid may solve the above problems. It is a potential diagnostic tool with minimal invasiveness and, most importantly, a direct confirmation of MTB in the pleural space. However, previous PCR-based methods reported suboptimal sensitivity for diagnosing TBP, of about 40-80%. The sensitivity may be limited by the low level of MTB cfDNA and the single-gene target (IS6110 or devR) in the detection. With the advancement of next-generation sequencing (NGS) technology, the detection of microbial cfDNA by metagenomic NGS (mNGS) has been applied in infectious disease diagnostics. Chang et al. have previously evaluated the use of mNGS in blood and urine samples for tuberculosis diagnostics. They revealed that such methodology is limited by the low MTB concentration and the background of contaminating non-tuberculous mycobacterial (NTM) DNA, which shares similar sequences to the pathogenic MTB.

Our group has recently developed a new laboratory assay measuring the MTB cfDNA levels. We hypothesise that the new MTB cfDNA assay has better diagnostic performance than conventional microbiological methods in discriminating pleural effusions due to TBP from non-TBP. Since the development cohort only contains a limited number of pleural fluid samples, a large-scale confirmatory study containing pleural effusions with a wide spectrum of causes is required to confirm its clinical utility before being introduced in clinical practice. This study has three major aims. First, the diagnostic accuracy of the new MTB cfDNA assay in diagnosing TBP will be determined using a large cohort containing pleural fluid samples of various causes from countries with different TB burdens. Second, the clinical and laboratory factors determining the pleural fluid MTB cfDNA level will be identified. Third, the ability of the assay to capture different anti-TB drug-resistance genes will be explored on pleural fluid from DR-TB endemic areas.