Subclinical TB infections, known as early-stage active TB disease, are easily ignored, due to a lack of TB symptoms. In order to rule out active TB disease in such cases, radiologic and immunologic assays of sputa or BALF could be used to help detect active disease. However, in cases where patients cannot produce sputa or receive smear-negative results, clinicians are generally reluctant to administer early empirical anti-TB treatments based solely on radiologic findings (e.g., miliary patterns on X-rays) and results of immunologic assays. Therefore, other diagnostic techniques are needed to correctly diagnose TB in sputum-scarce or smear-negative patients with suspected PTB.
When nanopore sequencing assay technologies first became available in the marketplace, they were mainly used for genome sequencing [10]. With the advancement of sequencing chemistries and computational capacity, nanopore sequencing assay technologies have matured into clinical applications in recent years [11]. One such application that is currently used in clinical infectious disease settings, nanopore sequencing, is used most often to diagnose patients experiencing fever without detectable localized infections who lack M. tuberculosis culture-positive results [12,13]. In addition, nanopore sequencing assays can support rapid diagnoses of infections caused by slow-growing microorganisms, such as mycobacteria that cause TB and non-tuberculous mycobacterial (NTM) infections [14-19]. Indeed, correct diagnosis of TB versus NTM through rapid identification of mycobacteria at the species level is critically important prior to administration of specific and effective medications in order to maximize treatment outcomes. Moreover, accurate early diagnosis can avoid unnecessary testing and side effects associated with unnecessary medications, thus reducing overall treatment costs. Toward this end, sequence-based assays that only require at most a few days for completion have transformed clinical TB diagnosis, even when samples must be sent to private laboratories for testing. Therefore, use of sequence-based methodologies has undeniably greatly impacted clinical management of mycobacterial infections [2].
The Oxford Nanopore Technologies MinION device is gaining popularity as a platform used by routine clinical microbiology testing laboratories to perform nanopore sequencing assays. When nanopore sequencing assays were first used for laboratory diagnosis of infectious diseases, usually more than 10 samples were processed in one sequencing run in order to reduce the sequencing cost per sample [2]. However, the recent development of the Oxford Nanopore Technologies MinION device has expedited laboratory adoption of diagnostic nanopore sequencing assays, due to its low equipment cost, short turn-around-time, and small physical size for increased portability [5,6]. When this platform first became commercially available in 2015, its sequencing error rate was still very high [20]. Nevertheless, after several rounds of improvement, the sequencing error rate reached an acceptable range [21,22] that led to increased demand for the device. In turn, increased demand has fueled further improvements that have led to enhanced scalability, ease of use, and flexibility that have enabled this technology to better serve clinical microbiology laboratories handling diverse specimen volumes. Taken together, these advantages highlight the potential of the Oxford Nanopore Technologies MinION device to serve as an effective platform that will greatly transform clinical microbiological testing [2].
The rapid development of next-generation sequencing has stimulated detection of bacteria, fungi, viruses, and parasitic organisms through nontargeted DNA/RNA sequencing that has facilitated quick identification of pathogens to support early and accurate infectious disease diagnosis. For example, Huang et al. employed a sequence-based assay to successfully detect human pathogens in 94.49% of samples from patients with pulmonary infections who had tested negative for pathogens using traditional pathogen detection methods. Their results thus demonstrated that sequence-based assay accuracy and sensitivity exceeded corresponding standard pathogen detection assay performance indicators [23]. Similarly, Wang et al. found that their sequence-based assay was more sensitive than traditional methods for detection of mixed pulmonary infections [24], whereas Chen et al. obtained sequence-based test results for BALF specimens collected from patients with severe PTB disease that closely aligned with standard culture method-based results [25]. Here we report for the first time successful nanopore sequencing assay-based detection of M. tuberculosis in BALF collected from sputum-negative TB patients and demonstrate its value for use in diagnosing suspected PTB cases. Ultimately, the nanopore sequencing method provided superior sensitivity (75.86%) as compared to culture and Xpert assay sensitivities of 48.28% and 41.38%, respectively. Thus, these results demonstrated greater sensitivity of this method as compared to sensitivities of conventional culture-based and Xpert tests, although its specificity was comparable to that obtained from the conventional tests (e.g., Xpert). Finally, the area under the curve (AUC) value obtained for the nanopore sequencing assay exceeded AUC values obtained for the culture-based assay and Xpert, thus indicating that the nanopore sequencing assay provided superior diagnostic performance when used to test suspected PTB cases. As a final consideration, in cases where one sampling technique is contraindicated, our results revealed that similar results were obtained for sputum and BALF samples, thus demonstrating that clinicians can choose between the two sampling techniques without hesitation. Taken together, these results collectively suggest that nanopore sequencing should be useful as an additional testing method to support improved diagnostic detection of PTB cases.
Although a standard culture-based mycobacterial detection assay was evaluated in this study, this method could not discriminate between TB and NTM organisms as well as Xpert and nanopore sequencing methods, since the culture-based assay produced culture-positive results for both mycobacterial types, as illustrated by the fact that although culture assay specificity was suitable for discriminating between mycobacterium and non-mycobacterium groups, culture assay results obtained for 3 of 29 control samples with NTM disease were-positive for M. tuberculosis. Thus, the low specificity of the culture method makes it of limited value when used for TB diagnosis such that the high number of false-positive results obtained using the culture-based method indicate that results obtained using this test should be carefully interpreted when used for TB diagnosis in areas where NTM is epidemic.
Xpert also has limitations in that this assay only targets rpoB as a specific M. tuberculosis complex-associated sequence, which is present as only a single copy per genome. By contrast, the nanopore sequencing assay targets the IS6110 insertion sequence, which is not only specific for the M. tuberculosis complex, but is present at 10 to 12 copies in genomes of various M. tuberculosis strains. Importantly, this feature of nanopore sequencing assays makes them more sensitive than culture-based and Xpert assays for use in detecting diverse clinical M. tuberculosis isolates exhibiting variable tissue dissemination patterns and pathogenic properties.
Our study had several limitations. First, the study was based on a relatively small sample size, which might have introduced bias into our results that should be addressed through future studies to evaluate the diagnostic efficacy of the nanopore sequencing assay. Second, since our research focused on BALF specimens, the collection of lavage specimens containing numbers of pathogenic bacteria below the detection level of the test may have resulted in false-negative results as a reminder that the nanopore sequencing assay is still an auxiliary tool that should be used in combination with clinical features, radiological imaging findings, and laboratory test results to diagnose suspected PTB cases. Third, we could not rule out that false-negative and false-positive nanopore sequencing assay results occurred due to (1) a sequence depth that was too low; (2) high host genome background noise and low microbial pathogen biomass; (3) administration of antibiotics to patients prior to testing; and (4) contamination of samples with genomes of environmental microbes or human flora [23]. Fourth, the absence of certain control diseases, such as lymphomas and rheumatoid arthritis (rare diseases in patients of Beijing Chest Hospital), may have biased our results.
In summary, our findings demonstrate that nanopore sequencing assay could permit improved detection of M. tuberculosis in BALF samples in sputum-scarce or smear-negative cases with suspected PTB. Further investigations are needed to confirm our findings based on larger and more diverse patient populations.