The research, published in Frontiers in Cellular and Infection Microbiology, was led by a team at Chifeng Municipal Hospital in Inner Mongolia. The investigators enrolled 941 patients with suspected lower respiratory tract infections who all underwent flexible bronchoscopy, a procedure in which sterile saline is instilled into the radiographically affected parts of the lung and then suctioned back out as bronchoalveolar lavage fluid, or BALF. This fluid carries microbes from deep within the airways, but it also carries an enormous load of human genetic material from leukocytes and airway epithelial cells, a background that has long plagued sequencing-based diagnostics by diluting sparse microbial signals.

Patients were divided into two non-randomized cohorts. One group of 479 patients received metagenomic next-generation sequencing, or mNGS, in which all DNA in the sample is fragmented, sequenced without bias, and then mapped computationally against a comprehensive microbial database after human sequences are filtered out. The other 462 patients received targeted next-generation sequencing, or tNGS, which uses multiplex PCR primers to selectively enrich more than 500 common respiratory pathogens, including bacteria, fungi, viruses, and atypical organisms, before sequencing. Crucially, the tNGS workflow captured both DNA and RNA, reverse-transcribing viral RNA into cDNA, whereas the mNGS assay in this study was DNA-only, a technical asymmetry that the authors acknowledge likely underestimated RNA viruses in the metagenomic cohort.

Measured against a composite final clinical diagnosis adjudicated by a multidisciplinary panel of senior pulmonologists and microbiologists, both platforms clearly outperformed conventional laboratory testing. Metagenomic sequencing achieved an overall sensitivity of 90.6 percent, while the targeted panel reached 73.4 percent. Specificity was low for both assays, at 21.4 percent for mNGS and 14.3 percent for tNGS, a familiar problem in respiratory specimens where commensal flora and airway colonizers blur the line between infection and mere microbial presence. Yet the positive predictive values were strikingly high, 97.4 percent and 96.5 percent respectively, meaning that when either platform called a pathogen positive, clinicians could trust that finding enough to act on it.

The most dramatic results concerned the pathogens that traditional diagnostics habitually miss. Targeted sequencing identified 54 cases of Mycobacterium tuberculosis, the bacterium responsible for tuberculosis, in a cohort where routine acid-fast bacilli smears and standard nucleic acid amplification tests had missed 42 of those positives, a conventional false-negative rate of 77.8 percent. Mycobacterial culture, the historical gold standard, can take weeks to yield a result, so the speed advantage alone is clinically significant. The same pattern emerged for Pneumocystis jirovecii, an opportunistic fungal pathogen that can cause life-threatening pneumonia in immunocompromised patients, where conventional silver staining produced false negatives in 80 percent of the cases evaluated while tNGS rapidly established the correct diagnosis.

Perhaps the most eye-opening finding was the sheer prevalence of polymicrobial infection. In the tNGS cohort, 93.5 percent of positive cases involved more than one pathogen, with single-pathogen infections accounting for only 6.5 percent. Cross-kingdom co-infections, mixing bacteria, viruses, and fungi simultaneously, made up 32.6 percent of positives. The most frequent pairing was Streptococcus pneumoniae with Human herpesvirus 7, seen in 30 patients, followed by bacterial double infections such as Klebsiella pneumoniae with Streptococcus pneumoniae in 25 cases. Combinations of opportunistic fungi with bacteria were also common, including Pneumocystis jirovecii alongside Pseudomonas aeruginosa or Klebsiella. Conventional tests, which typically hunt for a single pathogen or depend on highly specific culture conditions, are structurally ill-equipped to capture this complexity, and the study suggests the true etiological landscape of severe pneumonia is far more crowded than routine diagnostics imply.

Sequencing results also changed how patients were treated. In the mNGS cohort, where 461 of 479 patients had positive results, attending physicians modified initial empirical regimens in just over half of the cases, escalating therapy for 184 patients, or 39.9 percent, with broader agents such as carbapenems, glycopeptides, or targeted antifungals. In the tNGS group, 341 patients tested positive, and 77 of them, or 22.6 percent, required treatment escalation because their initial empirical antibiotics retrospectively proved inadequate against the pathogens identified. The clinical stakes of these adjustments are considerable, since severe pneumonia patients are usually started on broad-spectrum antibiotics empirically, and any unexpected, resistant, or fastidious organism hiding beneath that coverage can drive the illness into a fatal trajectory before culture results ever return.

Encouragingly, the patients whose therapy was escalated on the basis of sequencing did about as well as those who had received appropriate antibiotics from the start. Among the 77 tNGS-guided escalation patients, 66.2 percent showed clinical and radiological improvement, statistically indistinguishable from the 66.7 percent improvement rate in a comparison group of 249 patients whose initial empirical regimens were confirmed by the genomic data as adequate and therefore left unchanged, with a P value of 0.946. The authors interpret this as evidence that sequencing-guided escalation effectively rescued potentially failing empirical therapies, allowing patients whose initial treatment was wrong to catch up with those who happened to be treated correctly from the outset.

The study’s authors are careful to note that these outcome data come with important caveats. Because the design was retrospective, patients were not randomly assigned to the two platforms, and the diagnostic committee that set the reference standard had access to the sequencing results, raising the possibility of incorporation bias. Missing baseline variables, including formal severity scores such as APACHE II and SOFA, prevented robust multivariate regression, so the clinical improvements observed remain observational associations rather than proven causal effects of sequencing itself. Recovery from severe pneumonia depends on many factors beyond diagnosis, and the authors explicitly call for prospective, multicenter, paired-testing studies with long-term follow-up to isolate sequencing’s true impact on survival.

Practical considerations also temper enthusiasm for routine deployment. Metagenomic sequencing remains expensive, demands specialized bioinformatic infrastructure, and can take longer to return results than rapid point-of-care tests. Targeted panels reduce cost and, by acting as a molecular filter against human genomic background, appear to sharpen sensitivity for low-titer and fastidious organisms, but they can only see the roughly 500 pathogens they are designed to detect. Still, the Chifeng team’s conclusion is that both platforms offer substantial value, with tNGS standing out for elusive and low-abundance pathogens, and that sequencing-guided diagnostics are already steering antimicrobial decisions in ways associated with favorable clinical progress. As genomic testing costs continue to fall, the study offers a concrete glimpse of how precision diagnostics may reshape the management of severe respiratory infections, one lavage sample at a time.

Subject of Research: Comparison of metagenomic and targeted next-generation sequencing for diagnosing lower respiratory tract infections

Article Title: Comparative evaluation of metagenomic and targeted next-generation sequencing for pathogen detection in bronchoalveolar lavage fluid from patients with lower respiratory tract infections: a retrospective cohort study

Article References: Zhao, R., Na, H., Miao, W., Liu, L., Zhang, J., Miao, Y., Zhao, Y., & Li, X. (2026). Comparative evaluation of metagenomic and targeted next-generation sequencing for pathogen detection in bronchoalveolar lavage fluid from patients with lower respiratory tract infections: a retrospective cohort study. Frontiers in Cellular and Infection Microbiology, 16, Article 1877320. https://doi.org/10.3389/fcimb.2026.1877320

Image Credits: AI Generated

DOI: 10.3389/fcimb.2026.1877320

Keywords: lower respiratory tract infections, metagenomic next-generation sequencing, targeted next-generation sequencing, bronchoalveolar lavage fluid, antimicrobial stewardship, Mycobacterium tuberculosis, Pneumocystis jirovecii, polymicrobial infection, diagnostic sensitivity, precision medicine, clinical microbiology, pneumonia