Evaluation of diagnostic efficacy of targeted next-generation sequenci

Introduction

Globally, community-acquired pneumonia (CAP) is a widespread and severe acute respiratory infection in children, significantly endangering their health and lives.¹ Identifying CAP early and administering appropriate antibiotics promptly are crucial for enhancing the prognosis in children.² Misdiagnosing CAP often leads to the use of unsuitable antibiotics, posing a significant issue in clinical settings.³ A major contributor to the growth of antimicrobial resistance is the misuse of these antimicrobial agents.⁴ Not only does this phenomenon complicate infection treatment, but it also results in a chain of negative effects, including rising healthcare costs and direct negative health impacts on children, which is a crucial issue.

The gold-standard method for diagnosing microbiological issues is through bacteria culture.⁵ The traditional fungal culture method, considered the gold standard, is used to diagnose fungal infections.⁶ These methods are limited by their inapplicability to noncultivable bacteria and their selectivity for easily cultivable bacteria.⁷ In the realm of molecular diagnostics for virus detection, nucleic acid amplification is regarded as the gold standard.⁸ Conventional nucleic acid extraction techniques are challenged by multi-step processes, significant time demands, and the need for substantial sample volumes. Despite their effectiveness, commercial kits are expensive. Studies indicate that even with combined techniques, these tests do not significantly increase pathogen detection in children with lower respiratory tract infections.^9–12 Metagenomic Next-Generation Sequencing (mNGS) is the application of NGS technology in clinical microbiological diagnostics to detect nucleic acids.¹³ Targeted next-generation sequencing (tNGS) has made significant strides in overcoming the limitations of mNGS, such as the high background of human host nucleic acids, challenges in discriminating colonization, and the associated high costs, by expanding its target spectrum from singleplex to high-multiplex assays.^14–16

Bronchoalveolar lavage has traditionally been viewed as the definitive method for diagnosing lower respiratory tract infections in children.^17,18 However, collecting BALF in children might require anesthesia and heavily depends on the physician’s expertise, which can hinder early and quick pathogen diagnosis. The procedure of nasopharyngeal (NP) swab sampling is simpler, and its use instead of BALF would considerably ease the diagnostic process.¹⁹ However, there are only a handful of case reports documenting the success of tNGS on NP swabs in diagnosing pediatric CAP. Thus, the goal of this research was to examine the diagnostic capability of tNGS in identifying pathogens from NP swabs of children with CAP.

Methods

Study Population

We retrospectively analyzed BALF specimens from 147 children with CAP at Tongling Maternal and Child Health Hospital between May 2023 and October 2024. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the Tongling Maternal and Child Health Hospital Ethics Committee (Ethics Approval No. TLFYBJY-2023026). Prior to the commencement of the study, written informed consent was obtained from the parents or legal guardians of all participating subjects, ensuring their comprehensive understanding of the study’s purpose, procedures, potential risks, and benefits.

Definitions

Pediatric CAP is an infection that starts outside the hospital, even if caused by pathogens with a specific incubation period that show symptoms after admission, according to the “Guidelines for the Diagnosis and Treatment of Pediatric Community-Acquired Pneumonia (2019 Edition)”.²⁰ The definition of severe CAP includes having at least one major criterion or three minor criteria, as stipulated by the Infectious Diseases Society of America/American Thoracic Society guidelines.²¹

Specimen Acquisition and Examinations

Nasal specimens were collected by inserting the swab tip until resistance was encountered, followed by gentle rotation of the swab several times.²² BALF specimens were obtained from all patients who underwent bronchoscopic examination, following standardized protocols.²³ Nasal and BALF samples were preserved in the biological specimen storage cabinets at Tongling Fourth People’s Hospital. The specimens were subsequently utilized for the detection of pathogens employing both tNGS technique and CMT. DNA/RNA extraction from these samples was performed using an automated nucleic acid extractor, the KingFisher Flex (Thermo Fisher Scientific Inc., MA, USA), following the manufacturer’s protocol. The preparation of polymerase chain reaction (PCR) libraries was carried out using the Respiration100TMPlus system (KingCreate, Guangzhou, China). The tNGS detection workflow adhered to the methods described in previous studies.

Statistical Analysis

Categorical data were depicted using percentages. Comparative analyses of categorical and numerical variables were performed using the Chi-square test and the Student’s t-test, respectively. Statistical significance was defined as a p-value less than 0.05. Interrater agreement between Nasal tNGS and BALF tNGS-detected pathogens was computed utilizing Cohen’s Kappa. The statistical analysis was conducted utilizing R software (version 4.4.1), and graphical representations were created with GraphPad software (version 8.1).

Results

Clinical Characteristics

Our research encompassed a total of 147 pediatric patients hospitalized due to CAP. Table 1 shows the clinical characteristics of the study population. The distribution of pediatric patients illustrated infants under 1 year of age accounted for 10 (6.80%) cases, toddlers between 1 and 3 years of age comprised 12 (8.16%) cases, children aged 3 to 5 years made up 43 (29.25%) cases, and children over 5 years of age constituted the majority with 82 (55.78%) cases. There are 83 (56.46%) males and 64 (43.54%) females in the study population. The median length of hospital stay was 6.00 days, with an interquartile range of 5.00 to 7.00 days. Regarding the severity of CAP, 81 (55.10%) patients were classified as having mild CAP, while 66 (44.90%) patients had severe CAP. In terms of signs and symptoms, the majority presented with wet cough (138 patients, 93.88%) and fever (138 patients, 93.88%). Other symptoms included dry cough in 9 (6.12%) patients, chest pain in 4 (2.72%) patients, shortness of breath or chest tightness in 22 (14.97%) patients, hemoptysis or bloody sputum in 9 (6.12%) patients, and fatigue in 17 (11.56%) patients. Radiological findings revealed multi-lobe involvement in 71 (48.30%) patients, multifocal consolidation in 46 (31.29%) patients, airway patency in 72 (48.98%) patients, atelectasis in 35 (23.81%) patients, pleural effusion in 13 (8.84%) patients, and interstitial edema in 5 (3.40%) patients.

Comparison of the Diagnostic Performances of CMT, Nasal tNGS, and BALF tNGS

In our study focused on pediatric patients with CAP, we utilized three distinct diagnostic approaches to detect a spectrum of pathogens (Figure 1). The CMT method identified 17 different pathogens. Nasal tNGS, on the other hand, detected a broader range of 25 distinct pathogens, and BALF tNGS identified 21 unique pathogens. There was an overlap, with 15 pathogens being detected by all three methods (Figure S1). CMT uniquely identified Chlamydia trachomatis, while nasal tNGS was the only method to detect Klebsiella pneumoniae, Human herpesvirus, and Moraxella catarrhalis, as detailed in Figure S1. In the study, CMT identified pathogens in 91 cases, which accounts for 61.90% of the participants. For those who underwent both nasal tNGS and BALF tNGS testing, pathogens were detected in 144 (98.00%) cases and 147 (100.00%) cases, respectively. The detection rates for both Nasal tNGS and BALF tNGS were significantly higher than those for CMT, with the differences being statistically significant (P<0.001) (Table S1).

Figure 1 Pathogen distribution among CMT, nasal tNGS, and BALF tNGS. Abbreviations: CMT, conventional microbiological tests; tNGS, targeted next-generation sequencing; BALF, bronchoalveolar lavage fluid.

Among these cases, the proportion of patients with a single pathogen detected by nasal tNGS and BALF tNGS was 35.6% and 54.4%, respectively. For patients with two pathogens, the proportions were 41.1% and 14.3%, respectively. For three pathogens, the proportions were 15.1% and 23.1%, and for four pathogens, they were 6.8% and 8.1% (Figure 2A). Among the single pathogen cases, Mycoplasma pneumoniae had the highest prevalence, with 71.1% detected by nasal tNGS and 81.3% detected by BALF tNGS (Figure 2B).

Figure 2 (A) The number of pathogens detected through nasal tNGS and BALF tNGS, (B) In the context of single-pathogen cases, the distribution of various pathogens detected by nasal tNGS and BALF tNGS. Abbreviations: tNGS, targeted next-generation sequencing; BALF, bronchoalveolar lavage fluid.

Concordance in Pathogen Detection Between NP Swabs and BALF Utilizing tNGS

Table 2 illustrates the comparative analysis of the detection of various pathogens using nasal tNGS and BALF tNGS. The two diagnostic methods exhibited almost perfect agreement in the detection of Human respiratory syncytial virus, with a Cohen’s kappa of 0.92 (95% confidence interval, CI, 0.81–1). Substantial agreement was observed between the tests for Human metapneumovirus (Cohen’s kappa, 0.83; 95% CI, 0.59–1), Aspergillus fumigatus (Cohen’s kappa, 0.66; 95% CI, 0.22–1), and Human parainfluenza virus (Cohen’s kappa, 0.61; 95% CI, 0.38–0.84). Moderate agreement was noted for Rhinovirus (Cohen’s kappa, 0.6; 95% CI, 0.42–0.78), Cytomegalovirus (Cohen’s kappa, 0.6; 95% CI, 0.28–0.92), Haemophilus influenzae (Cohen’s kappa, 0.56; 95% CI, 0.33–0.78), Epstein-Barr virus (Cohen’s kappa, 0.55; 95% CI, 0.32–0.79), and Human adenovirus (Cohen’s kappa, 0.53; 95% CI, 0.16–0.89). Fair agreement was found between the two tests for Candida albicans (Cohen’s kappa, 0.39; 95% CI, 0.12–0.65), Streptococcus pneumoniae (Cohen’s kappa, 0.38; 95% CI, 0.1–0.67), and Mycoplasma pneumoniae (Cohen’s kappa, 0.33; 95% CI, 0.21–0.46).

Table 2 Comparative Analysis of Detection and Concordance of Various Pathogens by Nasopharyngeal Swabs and BALF tNGS

Discussion

In children, respiratory illnesses are the most common, with pneumonia being a major reason for hospitalization, particularly in those under five, where it is a significant cause of illness and death.²⁴ CAP accounts for the majority of these cases, with varying incidence and mortality rates across countries, often higher in developing nations.²⁵ Therefore, rapid and accurate detection of pneumonia pathogens is of significant value in improving the prognosis of pediatric patients with pneumonia.

Specimens such as NP swabs, oropharyngeal swabs, sputum, peripheral blood, and BALF are typically used in testing for the etiology of pediatric pneumonia. Known for their simple collection and non-invasiveness, NP swabs have shown considerable diagnostic effectiveness in detection of Severe Acute Respiratory Syndrome Coronavirus 2 and influenza virus pneumonia.^26,27 BALF serves as an essential diagnostic tool, offering precise insights into the causative agents of pneumonia.²⁸ The invasive nature of BALF requires skilled practitioners, and it is often performed on children under anesthesia, complicating the early identification of disease causes.²⁹ Simultaneously, using empiric antibiotic therapy, which is both harmless and affordable, can be a practical alternative to invasive methods.³⁰

The human nasopharynx hosts a diverse microbial ecosystem of normal flora, which typically includes Streptococcus, Haemophilus, Staphylococcus, Moraxella, and Corynebacterium species.^31,32 The body’s defense against pathogenic and opportunistic infections is supported by these resident microbiota, which function as a vital microbial barrier.³³ NP swabs are commonly used for viral detection,³⁴ but there are few studies on their effectiveness in diagnosing pathogens in children with lower respiratory tract infections. The PERCH (Pneumonia Etiology Research for Child Health) study primarily utilized nasopharyngeal and oropharyngeal (NP-OP) swabs analyzed via multiplex PCR as a cornerstone for etiological inference.³⁵ Critically, PERCH demonstrated that while NP-OP detection alone has limitations—particularly for bacteria like Streptococcus pneumoniae and Haemophilus influenzae where colonization complicates interpretation. PERCH also highlighted the necessity of supplementary specimens to augment NP-OP findings for bacterial pathogens, as NP-OP data alone lacked sufficient discriminatory power for bacteria without density thresholds and adjustment for antibiotic exposure. Thus, while PERCH validated NP swabs as a pragmatic tool for viral identification in resource-limited settings, its design emphasized that robust etiology assessment requires contextualizing NP-OP data within a multi-specimen, analytically advanced framework to delineate pathogen-specific contributions accurately. Our study found that nasal tNGS had a significantly higher pathogen detection rate compared to conventional microbiological testing (CMT) methods (98.00% vs 61.90%, P<0.001). Using Cohen’s Kappa to assess the diagnostic agreement between nasal tNGS and BALF tNGS for various pathogens, we observed good concordance for Human metapneumovirus, Aspergillus fumigatus, Human parainfluenza virus, Rhinovirus, Cytomegalovirus, Haemophilus influenzae, Epstein-Barr virus, Human adenovirus, Candida albicans, Streptococcus pneumoniae, and Mycoplasma pneumoniae.

Limitations

As this is a retrospective study, we are currently unable to obtain data on patient outcomes and secondhand smoke exposure. This study lacks formal turnaround time assessment and antimicrobial resistance gene detection, particularly concerning given the issue of antibiotic misuse in pediatric community-acquired pneumonia. Future investigations should evaluate TAT within real-world clinical workflows and incorporate genomic resistance markers to enhance targeted therapeutic strategies.

Conclusions

NP swab-based tNGS emerges as a highly effective diagnostic tool for pediatric CAP, demonstrating superior detection rates and strong concordance with BALF tNGS, particularly in identifying viral pathogens and providing early etiological diagnosis. This approach holds promise for streamlining the diagnostic process and informing timely treatment decisions in pediatric pneumonia management.