Nanopore sequencing outperforms culture in diagnosing NTM lung disease, retrospective cohort finds

Diagnosing pulmonary disease caused by nontuberculous mycobacteria has long frustrated clinicians, because the chronic cough, sputum production, fatigue and weight loss that mark the condition overlap with tuberculosis, chronic obstructive pulmonary disease, recurrent pneumonia and chronic bronchitis. Confirmatory culture, the traditional gold standard, can take weeks to yield a result and performs poorly in patients with low bacterial burdens, delaying treatment. A retrospective cohort study now reports that a rapid sequencing approach called nanopore targeted sequencing identified far more cases than conventional methods in a large Chinese patient group, while also exposing a dense web of co-infecting pathogens that routine diagnostics tend to miss.

Researchers at Nanjing Second Hospital analyzed 350 patients suspected of having pulmonary nontuberculous mycobacterial infection between June 2023 and March 2025. Of these, 293 were diagnosed with nontuberculous mycobacterial pulmonary disease according to 2020 Chinese Medical Association and 2021 ATS/IDSA guidelines, which require compatible respiratory symptoms, characteristic chest imaging and microbiological evidence such as the same species recovered from at least two sputum samples, one bronchial wash or lavage, or a lung biopsy. The remaining 57 patients formed a non-NTM comparison group. For patients who were culture-negative, diagnosis relied on a multidisciplinary review by respiratory physicians, radiologists and microbiologists rather than sequencing results alone, with final confirmation also weighing each patient’s response to targeted anti-NTM treatment.

The sequencing assay, marketed as the TBseq test, runs on Oxford Nanopore MinION or GridION instruments and targets the 16S rRNA and hsp65 gene regions to identify mycobacterial species. The panel covers the Mycobacterium tuberculosis complex and more than 30 clinically relevant nontuberculous species, alongside common bacterial, viral and fungal pathogens and a panel of antibiotic resistance genes. Genomic DNA was extracted from bronchoalveolar lavage fluid, sputum and tissue specimens using a standardized kit protocol, libraries were built with a multiplex targeted amplification approach, and sequencing proceeded on R9.4.1 flow cells with automated analysis software for result interpretation.

Compared against acid-fast bacillus smear, mycobacterial culture on liquid and solid media, and species identification by MALDI-TOF mass spectrometry with Sanger sequencing, the sequencing method achieved the highest sensitivity at 95.9 percent, with a 95 percent confidence interval of 93.6 to 98.2, and the largest area under the ROC curve at 0.892. Its specificity of 82.5 percent was slightly below the 84.2 percent of mycobacterial culture. Culture, despite its high specificity, detected only 50.5 percent of cases, while the acid-fast smear managed just 25.6 percent sensitivity. A Venn diagram analysis confirmed that sequencing flagged the greatest number of positive cases, including many that other methods missed. Among all NTM-positive patients, sequencing was positive in 281 cases, culture in 148 and species identification in 98; sequencing alone accounted for 100 of the 105 patients identified by a single method.

The authors caution that the reported specificity may be inflated. Because the reference standard for disease was a composite of clinical, radiological and microbiological criteria that partially incorporated sequencing results in culture-negative patients, incorporation bias could have made the test appear more specific than it truly is. No universal read-count or abundance thresholds exist to separate true infection from colonization or background noise, and the study’s differentiation relied on semi-quantitative consideration of read abundance and rank within the composite standard. Future work with independent reference standards and quantitative PCR or RNA-based sequencing is needed to establish clinically actionable cutoffs.

The patient characteristics themselves carried signals. The NTM group was significantly older, with a median age of 61.0 years versus 43.5 in the comparison group, and showed significantly lower CD3 and CD4 T-cell counts and a reduced CD4/CD8 ratio, alongside a longer disease course. Body mass index was comparable between groups. Cough with sputum was the most frequent symptom, and imaging showed patchy parenchymal and nodular shadows, with lesions concentrated in the right middle lobe, the lingular segment and the basal segments of both lower lobes. Hypertension affected 16.0 percent of the NTM patients and a history of tumor 7.8 percent. Night sweats, a classic tuberculosis symptom, were uncommon in both groups.

Species distribution leaned heavily toward Mycobacterium intracellulare, with 157 isolates, followed by M. abscessus with 48 and M. kansasii with 20. Less common organisms included M. avium, M. gordonae, M. chelonae, M. mucogenicum and M. fortuitum, plus a tail of rare species such as M. xenopi, M. chimaera and M. europaeum. Accurate speciation matters clinically because treatment regimens differ sharply by organism; M. abscessus is notoriously drug-resistant and demands a macrolide-based multidrug approach, while newer options such as inhaled amikacin for MAC and dual beta-lactam or oxazolidinone regimens for M. abscessus depend on early, correct species identification.

Perhaps the most striking result was the co-infection burden. Among the 293 NTM patients, 164, or 56.0 percent, carried one or more additional pathogens. Ninety-four cases involved a single co-pathogen, 45 involved two, 17 involved three, and a few involved up to six. Mycobacterium tuberculosis, detected in 32 patients, and human herpesvirus, in 19, were the most frequent companions, followed by Aspergillus in 12, Candida in 11 and Pseudomonas aeruginosa in 3. Specific combinations also emerged, with Candida plus herpesvirus in six cases, tuberculosis plus Nocardia farcinica in four, and tuberculosis plus herpesvirus in four. The authors note that the clinical significance of these findings remains to be fully elucidated and that detecting multiple organisms does not automatically mean each is causing active disease.

When the researchers split the cohort by culture status, 138 patients were sequencing-positive but culture-negative and 143 were positive by both. The culture-negative group was younger, with a median age of 60.0 versus 64.0 years, had a slightly higher BMI, and was more often diagnosed through lung pathological tissue rather than lavage fluid, suggesting lower microbial burden or more localized disease. Radiologically, these patients had less extensive involvement, with 21.0 percent showing five-lobe disease versus 50.3 percent of the culture-positive group. Species composition stayed largely consistent across both groups, with M. intracellulare predominating in each, which the authors interpret as evidence that sequencing detects authentic pathogens rather than nonspecific signals in culture-negative cases.

The study has clear boundaries. It is single-center and retrospective, the targeted panel cannot detect rare or novel pathogens outside its predefined range, and clinical outcomes and phenotypic drug susceptibility correlations were not fully analyzed. Cost-effectiveness comparisons with standard workflows remain to be done. Still, the authors conclude that nanopore targeted sequencing, by delivering results within a single work shift and simultaneously identifying species, co-infections and resistance markers, may meaningfully improve NTM diagnosis when integrated with clinical symptoms, radiographic findings and host immune status rather than used in isolation.

Subject of Research: Diagnostic performance of nanopore targeted sequencing for pulmonary nontuberculous mycobacterial disease in a retrospective cohort

Article Title: Clinical features of nontuberculous mycobacteria and diagnosis with nanopore targeted sequencing: a retrospective cohort study

Article References: Liu, X., Wang, T., Guo, Y., Zeng, Y., & Gao, W. (2026). Clinical features of nontuberculous mycobacteria and diagnosis with nanopore targeted sequencing: a retrospective cohort study. Annals of Clinical Microbiology and Antimicrobials, 25(1), Article 37. https://doi.org/10.1186/s12941-026-00876-1

Image Credits: AI Generated

DOI: 10.1186/s12941-026-00876-1

Keywords: nontuberculous mycobacteria, nanopore targeted sequencing, pulmonary NTM disease, diagnostic sensitivity, mycobacterial culture, co-infections, Mycobacterium intracellulare, Mycobacterium abscessus, bronchoalveolar lavage, retrospective cohort, T-cell immunodeficiency, antimicrobial resistance genes