Commentary: Metagenomic next-generation sequencing for Mycobacterium tuberculosis complex detection: a meta-analysis

We read with interest the meta-analysis by Yulian Li et al on the metagenomic next-generation sequencing (mNGS) for Mycobacterium tuberculosis complex detection (1). The use of mNGS for pathogen detection provides microbial profiling information and can also be used to customize the detection of drug resistance, potentially revolutionizing TB diagnosis. Despite this, many academic groups or health organizations, including the World Health Organization, do not recommend mNGS as an initial diagnostic tool for TB. Possible reasons include the high cost of mNGS and the possibility of false positive results (2).From this analysis, the authors concluded that the specificity of mNGS for TB diagnosis is 100% (95% confidence interval 0.99-1.00). The result is too perfect. From a diagnostic technique point of view, we believe that this high specificity is not very reliable and is more likely to be the result of publication bias. The basic principle of mNGS is also birdshot sequencing, and the primer design is not particularly improved compared to other molecular diagnostic techniques. Compared to other validated assays (e.g., Xpert MTB/RIF), mNGS does not technically reduce the possibility of false positives and may introduce more false positives due to contamination because of the semi-open assay environment. (3,4). The absence of randomized blinded trials in the cited studies also contributed to the unreliable results.In addition, it is very important to note that the analysis says that composite diagnostic criteria were almost always used in the cited studies. Some of the studies included mNGS itself in their diagnostic criteria, which is obviously not very appropriate for use in the evaluation of this study (5); another part did not describe whether or not mNGS was included. we believe that even if composite diagnostic criteria are used to do the analysis, a comparison with bacteriologic testing should be provided at the same time.In addition, it should be noted that the authors mention several times in both the INTRODUCTION section and the CONCLUSION section that mNGS has important value for areas with high TB prevalence, but do not evaluate the strengths and weaknesses of the method compared to other molecular methods. In terms of sensitivity, it is not superior to existing methods (such as the Xpert Ultra) (6), and the cost is much higher. The authors cite "a 10% prevalence rate" for the COVID-19 population in Africa, whereas the references are all for the Chinese population, and this misplaced argument is also unreasonable.The mNGS has the property of being a high-throughput assay and may therefore be more suitable for drug resistance detection of M. tuberculosis or for differential diagnosis of unexplained infections than for the initial diagnosis of TB. The World Health Organization has also recently actively evaluated the value of targeted NGS for Mycobacterium I n r e v i e w tuberculosis drug resistance detection (7), and again has not recommended its use for the initial diagnosis of tuberculosis. Based on currently available evidence, we believe that further evaluation of the suitability of mNGS for the diagnosis of TB is needed.