Despite the utilization of existing antifungal drugs, invasive fungal diseases (IFDs) are estimated to result in approximately 6.5 million incidence and 3.8 million deaths worldwide annually (Denning, 2024). Among the IFDs, invasive pulmonary fungal infections (IPFIs) are frequently observed in clinical settings, notably among individuals suffering from lung cancer, diabetes, tuberculosis, compromised immune systems, and those on prolonged antibacterial medication. In recent years, the prevalence and mortality of IPFI are generally increasing (Parums, 2022). Hence, rapid identification of pathogens is of great importance for IPFI.

At present, the conventional techniques for diagnosing IPFI encompass smears, cultures, antigen-antibody assays, and molecular biology tests. However, these existing methods are challenging to meet clinical needs due to low positivity rates and because they are time-consuming and need prior assumptions. Patients who test negative by conventional methods are treated with empirical antibiotics, which may lead to potential reinfection and adverse reactions (Huang et al., 2020). In order to meet the current need for prompt identification of pathogens and initiate timely and appropriate treatment, next-generation sequencing (NGS) has seen significant advancements this year, which becomes a front-line diagnostic in identifying rare pathogens and in the assessment of patients who may be suffering from severe infections (Wilson et al., 2014; Miller et al., 2019; Zhu et al., 2020; Zheng et al., 2021).

The common applications of NGS in diagnostic laboratories include metagenomic NGS (mNGS) and targeted NGS (tNGS). mNGS allows for the comprehensive detection of a wide range of pathogens, including viruses, bacteria, fungi, and parasites. However, the simultaneous detection of both DNA and RNA processes is economically challenging (Li et al., 2022). tNGS based on multiplex PCR amplification or probe capture enriches nucleic acid of known pathogen (Zhao et al., 2021; Singh, 2022). As a more cost-effective and high sensitivity assay, tNGS is gaining attention in clinical infectious management (Chen et al., 2024). To date, there has been a lack of comparative analysis between mNGS and tNGS in IPFI. Therefore, we conducted this retrospective study among patients with IPFI to evaluate the diagnostic efficacy of mNGS, tNGS, and conventional microbiological tests (CMTs) to assess the practical application of mNGS and tNGS in the clinical management for managing invasive IPFIs.

Patients presenting with IPFI often exhibit a range of underlying health conditions, posing a challenge in clinical management. The options for anti-fungal medications are limited, typically being both expensive and require prolonged treatment duration. Additionally, anti-fungal drugs often present significant adverse reactions and may be poorly tolerated by patients. Therefore, achieving an accurate diagnosis of IPFI is critically important.

The emergence of NGS has brought great progress to the diagnosis of infectious diseases. In recent years, mNGS has been widely accepted and used in clinical practice for early treatment or differential diagnosis of critically ill patients (Chiu and Miller, 2019; Greninger and Naccache, 2019; Chen et al., 2020). Our previous study showed that mNGS possesses greater sensitivity and a higher rate of clinical acceptance compared to culture when used as a reference for clinical diagnosis (Li et al., 2024), which was widely reported and supported by several studies. However, mNGS still has some limitations: easily influenced by human genes, expensive, and lack of standardization in experimental procedures and sequencing report interpretation. In contrast, tNGS involves the use of a panel of specific sequences from pre-selected pathogens, which offers higher specificity and cost-effectiveness and can eliminate the interference of human-derived genes (Wylie et al., 2015).

Upon reviewing relevant literature, it is found that there were few published studies compared the performance of tNGS and mNGS in lower respiratory tract infections (Huang et al., 2024; Sun L. et al., 2024; Wei et al., 2024). However, assessments of the clinical diagnostic value of tNGS in IPFI are limited. In the present study, two promising NGS-based detection methods, mNGS and tNGS, were compared in fungal diagnosis effectiveness in BALF. The tNGS panel used in this study was a primer or probe specifically constructed for 198 pathogens.

In our study, according to composite clinical judgement, 61 patients with IPFI and 54 non-IPFI cases were identified. We then compared the diagnostic efficiency and consistency between traditional culture method, mNGS, and tNGS, to assess the clinical validity of NGS in detecting fungi. In general, the diagnostic efficacy of mNGS and tNGS was comparable in identified IPFI without statistically different (P > 0.05), which aligns with the findings from previous studies (Li et al., 2022; Huang et al., 2024). Meanwhile, mNGS and tNGS, with their unbiased detection advantages, can identify pathogens that are difficult to detect by traditional culture methods, such as Pneumocystis jirovecii, Cryptococcus neoformans, viruses, and non-tuberculous mycobacteria. Our findings indicated that mNGS had equivalent sensitivity to tNGS, but with slightly higher specificity, PPV, and NPV compared to those of tNGS. However, tNGS identified more A. terreus complex, A. niger complex, and C. parapsilosis than mNGS. Moreover, in our study, the sensitivity and NPV of mNGS and tNGS were markedly higher than those of the combined CMTs, which include culture, G test, and IF, as well as traditional pathogen detection methods alone (P < 0.05). The combination of CMTs with mNGS identified all patients with IPFI with a NPV of 100%, whereas the combination of CMTs with tNGS missed one patient with IPFI (case 31). A. fumigatus was detected by mNGS (number of sequences: 6) in the 31st patient with IPFI with GM test being positive and with culture being negative. Interestingly, P. jirovecii, an opportunistic fungus often seen in immunocompromised patients with mixed infections, was the most frequently detected in both NGS-based assays in our study. This prevalence may be associated with the older age of the patients and the presence of underlying diseases. Coinfections with P. jirovecii, cytomegalovirus (CMV), and Herpesviruses such as Epstein-Barr virus (EBV) have been widely reported, and our findings are in line with previous studies (Rucar et al., 2024).

The detection efficiency of two-NGS based assays is excellent; however, there are also cases that are not consistent with the positive culture results. mNGS failed to detect fungi in three patients (cases 18, 28, and 37) of C. albicans, A. fumigatus, and A. niger complex, respectively, whereas culture and tNGS were positive. The three cases were confirmed as true positives by traditional culture and tNGS. However, the low sequence counts detected by tNGS, which barely met the reporting threshold, suggested low pathogen loads. Therefore, we suppose the false negative values of mNGS were likely due to these low pathogen loads. In addition, tNGS missed C. orthopsilosis and P. kudriavzevii, whereas culture and mNGS were positive in case 48 and case 59, respectively. These two cases were confirmed true positive as well. Since the differences have no pattern and are not caused by the differences in the detection ranges of mNGS and tNGS, these differences may be related to reasons such as degradation of samples due to storage or wet experimental performance.

Except for fungi, the detection performance of the two NGS-based assays for bacteria and virus was similar as well, which is comparable to that in the previous studies (Li et al., 2022; Sun W. et al., 2024). As the development of tNGS products was targeting specific infectious pathogens, our BALF tNGS panel mainly focuses on lower respiratory tract infection pathogens. For bacteria, due to the broader coverage of mNGS, the detection rate of mNGS is higher than that of tNGS in our study. mNGS detected a higher number of E. faecium, C. striatum, R. mannitolilytica, and B. ambifaria compared to tNGS, whereas tNGS identified more S. anginosus group. E. faecium and C. striatum are the components of the normal flora of skin and mucous membranes and are often regarded as opportunistic pathogens, particularly in individuals with compromised immune function. In such cases, they may lead to severe infections, thereby possessing significant clinical implications. Although the significance of R. mannitolilytica and B. ambifaria in lower respiratory tract infections is limited, a rare case of spheroid pneumonia caused by R. mannitolilytica was diagnosed through mNGS as reported (Ma et al., 2023). For viruses, tNGS has a greater advantage than mNGS in the detection of RNA viruses, such as SARS-CoV-2, Rhinovirus A, human respiratory syncytial virus, human coronavirus HKU1, and OC43, which could not be detected by the mNGS DNA process. However, the detection rate of several DNA virus such as CMV was also slightly lower than that of mNGS in our study (mNGS, n = 15; and tNGS, n = 5). The missed detection of CMV by tNGS was mainly due to the filtering of the sequence number that did not meet the reporting standard.

In addition, the ability of tNGS and mNGS to simultaneously identify bacteria, fungi, and viruses was also illustrated in our study. By comparing the results of culture and these two NGS-based technologies, the coinfection rate increased from 7.83% to 47.83% and 56.5% with significant difference for tNGS and mNGS, respectively. However, there was no significant difference observed between mNGS and tNGS, suggesting that the detection capabilities of both methods in cases of mixed infection are largely comparable as well.

As advanced diagnostic tools, mNGS and tNGS often elicit high expectations in clinical practice. However, there are several issues that need attention in the result interpretation. mNGS and tNGS have differences in the presentation of the final sequence number. Since tNGS was based on ultra-multiplex PCR, which specifically increases the number and proportion of nucleic acid sequences of the target pathogen, therefore, the sequence number of tNGS in the final report tends to be higher than that of mNGS. Therefore, the results reporting rules of each pathogen of these two methods need to be combined with clinical practice to avoid false negatives or false positives that may occur in bioinformatics analysis. Accordingly, it is essential to engage in comprehensive communication with clinicians regarding the fact that the sequence counts obtained from tNGS and mNGS should not be directly compared horizontally. Moreover, because NGS-based technologies could not differentiate between colonizing and pathogenic pathogens, it is essential to consider the pathogenic significance when detecting colonizing bacteria with low sequence, such as Candida albicans, in BALF.

Our study had several limitations. First, it was a single-center retrospective study, which may limit the generalizability of the results and introduce potential biases related to the specific patient population. Second, the sample size was relatively small, and several cases lacked paired traditional results including G test and IF, owing to the restriction of enrolled cases.

Collectively, in our cohort of 115 patients, tNGS showed comparable diagnostic value to mNGS and was significantly superior to CMT. By focusing on specific pathogen-related gene regions, tNGS not only reduces the volume of sequencing data but also lowers costs and enhances cost-effectiveness for detecting common pathogens. In our lab, tNGS completes the process from sample processing to sequencing in just 15–16 h, compared to mNGS of 20–24 h, facilitating rapid clinical diagnosis. Additionally, for non-critical patients, targeted detection by tNGS can resolve most issues and shows higher sensitivity in detecting low-abundance pathogens, making it highly applicable in clinical settings. Given these practical advantages of tNGS in clinical microbiology, patients with lower respiratory tract infections can be referred for tNGS alongside traditional pathogen detection methods to identify common pathogens. Conversely, mNGS is more appropriate for critical illnesses, particularly in cases of unexplained infections. However, due to the variations in wet and dry experimental procedures across different laboratories, both NGS-based methods currently lack widely accepted standards and quantitative thresholds; therefore, NGS is not intended to supplant traditional diagnostic methods. Instead, NGS-based methodologies should be integrated flexibly with conventional techniques to enhance clinical diagnostic accuracy and inform treatment strategies effectively.