Meta‐Analysis of the Application Value of Metagenomic Next‐Generation Sequencing Technology in the Diagnosis of Infectious Diseases of the Spine

ABSTRACT

Objective

To evaluate the diagnostic efficacy of metagenomic next‐generation sequencing (mNGS) in infectious diseases of the spine (IDS).

Methods

Systematic literature on the application of mNGS in the diagnosis of IDS was retrieved by two independent researchers from databases including Pubmed, China National Knowledge Infrastructure (CNKI), Wanfang, and VIP from the inception to 30 November 2024. Meta‐analysis was conducted using Meta‐Disc 1.4 and Stata 18.0 software.

Results

The initial search identified 314 articles. After applying predefined inclusion and exclusion criteria, 15 studies were included, encompassing 1236 patients, of which 835 had confirmed IDS. Meta‐analysis revealed that the pooled sensitivity and specificity of mNGS for IDS diagnosis were 0.95 (95% CI: 0.88–0.98) and 0.60 (95% CI: 0.48–0.71), respectively. The positive likelihood ratio was 2.3 (95% CI: 1.7–3.2), and the negative likelihood ratio was 0.09 (95% CI: 0.04–0.22). The pooled diagnostic odds ratio was 26 (95% CI: 9–75), with an area under the summary receiver operating characteristic curve of 0.85 (95% CI: 0.82–0.88).

Conclusion

The primary diagnostic value of mNGS lies in its ability to serve as a rapid screening tool for disease exclusion. However, for diagnosing IDS, it is essential to integrate other clinical indicators for a comprehensive assessment to confirm the diagnosis.

To evaluate the diagnostic efficacy of metagenomic next‐generation sequencing (mNGS) in infectious diseases of the spine (IDS). Systematic literature on the application of mNGS in the diagnosis of IDS was retrieved by two independent researchers from databases including Pubmed, CNKI, Wanfang, and VIP from the inception to 30 November 2024. The application of mNGS in IDS diagnosis has high diagnostic value.

1. Introduction

Infectious diseases of the spine (IDS) are a group of infectious diseases of different parts of the spine (vertebral body, intervertebral disc, appendages, spinal canal and adjacent paravertebral tissues) caused by different pathogenic microorganisms [1, 2]. In developed countries, the incidence of IDS ranges from 1 to 2.5 per 100 000 people, with a mortality rate of 2% to 4% [2]. The incidence of IDS has been on the rise annually due to an aging population, advances in diagnostic techniques and an increase in invasive surgical procedures [3]. The insidious onset, atypical clinical presentation, and lack of specificity of imaging and laboratory tests often complicate the diagnosis of IDS, leading to potential misdiagnosis and delayed or inappropriate treatment [4, 5]. Early identification of the causative pathogen is crucial for effective management of IDS [6], but traditional bacterial culture methods are time‐consuming, and exhibit low sensitivity [7].

Next‐generation sequencing (NGS), also known as high‐throughput or massively parallel sequencing, is a category of technologies that enable the simultaneous and independent sequencing of thousands to billions of DNA fragments. The applications of NGS in clinical microbiological testing are numerous and encompass metagenomic NGS (mNGS), which offers an unbiased method for detecting pathogens [8]. mNGS allows for the rapid sequencing of nucleic acids in samples, (including those from humans and pathogenic microorganisms) and enables comparison with human genome sequences and pathogenic microbial genome sequences to identify the species and proportions of microorganisms present. It provides an optimal approach for genomic analysis of all microorganisms in a sample, not just those that can be cultured [9]. This method facilitates the swift and objective detection of various pathogens, such as viruses, bacteria, fungi, and parasites, in clinical samples.

There has been a gradual increase in the number of studies using mNGS technology to diagnose IDS, but the diagnostic value is still controversial due to the small sample size of individual studies and the wide variation in reported sensitivities. Therefore, this study conducted a comprehensive review of the existing literature on the application of mNGS in the diagnosis of IDS, both nationally and internationally, and used meta‐analysis to evaluate the diagnostic value of mNGS technology in IDS.

2. Materials and Methods

2.1. Search Strategy

Systematic literature on the application of mNGS in the diagnosis of IDS was retrieved from databases including PubMed, China National Knowledge Infrastructure (CNKI), Wanfang, and VIP databases from inception to 30 November 2024 in either English or Chinese language publications. The selection of relevant literature was conducted in two phases. Firstly, domestic and international publications that applied mNGS technology to IDS were identified. Secondly, the list of the included articles was reviewed to identify any further relevant literature. The search strategy was as follows: (spinal infection OR spine infection OR spondylodiscitis OR spinal epidural abscess OR vertebral osteomyelitis OR pyogenic spondylodiscitis) AND (metagenomic next‐generation sequencing OR next‐generation sequencing OR high‐throughput nucleotide sequencing OR mNGS OR NGS). Chinese characters were used for retrieval in CNKI, Wanfang, and VIP databases, and the primary retrieval method was consistent with that used for other sources.

2.2. Inclusion and Exclusion Criteria

Studies were included if they met the following criteria: (1) Studies such as cohort, case–control, prospective, or retrospective studies that assessed the diagnostic efficacy of mNGS were included. (2) The study targeted a Chinese population. The IDS patients in the study population were clinically diagnosed and/or confirmed with clear diagnostic criteria, and the control group included IDS‐related diseases that required clinical differentiation. (3) The number of IDS cases was over 10 in the article. (4) The age of study subjects is unrestricted. (5) The publication language was limited to English and Chinese. (6) Studies containing necessary data for extraction, such as counts of False Positives (FP), True Positives (TP), False Negatives (FN), and True Negatives (TN).

The exclusion criteria were as follows: (1) Duplicate publications in different databases were excluded. (2) Literature types such as master's and doctoral theses, case reports, conference papers and reviews were excluded.

2.3. Data Extraction

Two researchers (XC and HY) independently screened the literature, extracted data, and cross‐checked the results. Any disagreements were resolved through discussion or consultation with a third party (JZ). Data including author name, year of publication, country, type of study, sample size, diagnostic criteria, specimen type, pathogen type, and type of mNGS sequencing platform, along with other parameters, were included. Individual studies were assessed, and data regarding TP, FN, FP, and TN values for the assay were extracted to evaluate the diagnostic accuracy.

2.4. Literature Quality Assessment

Two researchers (XC and HY) independently evaluated the quality of the included literature using the Quality Assessment of Diagnostic Accuracy Studies‐2 (QUADAS‐2) tool, which is designed to assess the quality of diagnostic accuracy studies. Each item was assessed based on three criteria: “Yes”, “No”, and “Unclear”. Studies that met the assessment criteria were marked as “Yes”, those that did not meet the criteria were marked as “No”, and those that partially met the criteria or had unclear descriptions were marked as “Unclear”. A “Low risk” designation was assigned when all questions in a section were answered “Yes”; a “High risk” designation was assigned if any question was answered “No”; and “Unclear” was assigned in all other cases. Any disagreements were resolved through discussion or consultation with a third researcher (JZ).

2.5. Statistical Analysis

The primary outcome was a summary of the combined sensitivity and specificity of all studies. Secondary outcomes included subgroup analysis based on sequencing platform, diagnostic criteria, sample types, and pathogen types based on pre‐collected information. The grouping information was defined as different subgroups and added to the 2 × 2 table. Spearman's correlation analysis was calculated using Meta‐Disc 1.4 software to analyze the logarithm of sensitivity and the logarithm of (1‐specificity), in order to determine the presence of a threshold effect. The data from each independent study was extracted from the 2 × 2 table, and forest plots of sensitivity and specificity were generated using Bivariate mixed effects models (the “midas” package) by Stata 18 software. The heterogeneity of the studies was then assessed using the p and I ² values. A fixed‐effects model was used for analysis when the value of I ² was less than 50%, indicating low heterogeneity, and the random‐effects model was used when the value of I ² was over 50%, indicating high heterogeneity. The sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), and their 95% confidence intervals (CI) were calculated and analyzed for the included studies. The summary receiver operating characteristic (SROC) curve was plotted, and the area under the curve (AUC) was calculated for the included studies. Subgroup analysis was then performed using a random‐effects model to explore the source of heterogeneity for significant heterogeneity. Sensitivity analysis (leave‐one‐out method) was used to investigate the stability of the pooled results, and Deeks' funnel asymmetry test was used to assess publication bias. The p‐value less than 0.05 was considered statistically significant.

3. Results

3.1. Literature Search Results

A total of 314 papers were searched by search terms and 290 initial papers were obtained by eliminating duplicates. After searching the titles and abstracts, 51 articles were obtained by excluding irrelevant articles, reviews and case reports. After reading the full text, literature without the control group or four‐cell table, 15 papers met the criteria for inclusion in the meta‐analysis. The flow chart of literature search is shown in Figure 1 and Table 1.

FIGURE 1.

Flowchart of the literature search.

TABLE 1.

Summary of study characteristics.

First author	Year	Study design	Country	Cases/controls	Diagnostic criteria	Type(s) of specimens	Type of pathogens	Sequencing platforms	TP	FP	FN	TN
Liu [21]	2021	Retrospective	China	20/18	His	Tissue/pus/blood/brain fluid	Containing Mycobacterium tuberculosis	BGISEQ‐50 platform (BGI‐Tianjin, China)	16	8	4	10
Shi [22]	2022	Retrospective	China	136/35	CDC	Tissue/pus/lavage fluid	Non‐tuberculous mycobacteria	Illumina Nextseq	76	19	60	16
Yin [23]	2022	Retrospective	China	35/11	CDC	Tissue/pus	Containing Mycobacterium tuberculosis	BGISEQ‐500 (MGI, Wuhan, China)	32	1	3	10
Zhang [24]	2023	Retrospective	China	28/12	CDC	Tissue	Containing Mycobacterium tuberculosis	—	22	8	6	4
Lv [25]	2024	Retrospective	China	62/14	CDC	Tissue/pus	Containing Mycobacterium tuberculosis	Dif seq platform (Dinfectome Medical Technology Inc., Nanjing, China)	51	8	11	6
Wang [5]	2023	Retrospective	China	21/4	His	Tissue	Containing Mycobacterium tuberculosis	BGISEQ‐50/MGISEQ‐2000 platform	16	1	5	3
Wang [10]	2023	Retrospective	China	92/8	CDC	Tissue/blood	Containing Mycobacterium tuberculosis	MGISEQ‐2000 platform (MGI, Shenzhen, China)	90	0	2	8
Jin [26]	2023	Retrospective	China	111/92	CDC	Tissue	Mycobacterium tuberculosis	BGISEQ 50 platform (BGI‐Tianjin, China)	79	0	32	92
Cheng [27]	2023	Retrospective	China	50/28	His	Tissue	Containing Mycobacterium tuberculosis	Illumina sequencers	42	1	8	27
Li [28]	2022	Prospective	China	91/9	CDC	Tissue/pus	Containing Mycobacterium tuberculosis	MGISEQ‐2000 platform (MGI Tech Co. Shenzhen, China)	81	1	10	8
Lin [29]	2023	Retrospective	China	31/8	His	Tissue	Containing Mycobacterium tuberculosis	Illumina HiSeq 2500 (Illumina, San Diego, CA)	27	1	4	7
Li [30]	2023	Retrospective	China	41/85	CDC	Tissue/pus	Mycobacterium tuberculosis	Illumina MiSeq instrument	16	1	25	84
Zhang [31]	2024	Retrospective	China	56/58	His	Tissue	Containing Mycobacterium tuberculosis	BGISEQ‐50 platform	42	9	14	49
Chen [32]	2024	Retrospective	China	34/15	CDC	Tissue/pus	Containing Mycobacterium tuberculosis	Illumina NextSeq CN500 (Illumina Inc., USA)	27	3	7	12
Ma [33]	2022	Retrospective	China	27/4	His	Tissue	Containing Mycobacterium tuberculosis	BGISEQ‐50/MGISEQ‐2000 platform	19	1	8	3

Abbreviations: CDC, composite diagnostic criteria; FN, false negative; FP, false positive; His, histopathology; TN, true negative; TP, true positive.

3.2. Characteristics of Studies and Quantitative Data Synthesis

A total of 15 papers were included in this study, comprising 4 Chinese‐language and 11 English‐language publications, which encompassed 835 patients as the case group and 401 individuals as the control group. Cases included in this study were given more detailed diagnostic criteria, including a confirmed diagnosis of IDS after clinical findings and/or the identification of aetiological/pathological/clinical anti‐infective treatment that was effective (meeting IDS diagnostic criteria). The results of the QUADAS‐2 assessment are presented in Table 2, indicating a generally high quality of the included literature.

TABLE 2.

QUADAS‐2 scale risk assessment results for included studies.

First author	Risk of bias				Applicability concerns
	Patient selection	Index test	Reference standard	Flow and timing	Patient selection	Index test	Reference standard
Liu [21]	High	Low	Low	Low	High	Low	Low
Shi [22]	High	Low	High	Low	Low	Low	Low
Yin [23]	Unclear	Low	High	Low	Low	Low	Low
Zhang [24]	Unclear	Low	Low	Low	Low	Low	Low
Lv [25]	Low	Low	Low	Low	Low	Low	Low
Wang [5]	Unclear	Low	Low	Low	Low	Low	Low
Wang [10]	High	Low	High	Low	Low	Low	Low
Jin [26]	Low	Low	Low	Low	Low	Low	Low
Cheng [27]	Low	Low	Low	Low	Low	Low	Low
Li [28]	Unclear	Low	Low	Low	Low	Low	Low
Lin [29]	Low	Low	Low	Low	Low	Low	Low
Li [30]	Unclear	Low	High	Low	Low	Low	Low
Zhang [31]	High	Low	High	Low	High	Low	High
Chen [32]	Unclear	Low	Unclear	Low	Low	Low	Unclear
Ma [33]	Low	Low	Low	Low	Low	Low	Low

3.3. Diagnostic Value of mNGS Technology Applied to IDS

3.3.1. Heterogeneity Test

The Spearman's correlation coefficient was −0.155 (p = 0.580), and the SROC curve did not exhibit a typical “shoulder‐arm” shape, indicating the absence of a threshold effect. Heterogeneity tests revealed significant heterogeneity in the sensitivity (Q = 105.03, p = 0.00, I ² = 86.67), specificity (Q = 102.77, p = 0.00, I ² = 86.38), PLR (Q = 170.55, p = 0.00, I ² = 88.76), NLR (Q = 105.77, p = 0.00, I ² = 86.76), and DOR (Q = 1.9e+26, p = 0.00, I ² = 100) across studies. Consequently, a random‐effects model was employed for the meta‐analysis of sensitivity, specificity, PLR, NLR, and DOR.

3.3.2. Results of Diagnostic Value Parameters

The meta‐analysis demonstrated that the pooled sensitivity and specificity of mNGS technology for IDS diagnosis were 0.95 (95% CI: 0.88–0.98) and 0.60 (95% CI: 0.48–0.71), respectively. The pooled PLR was 2.3 (95% CI: 1.7–3.2), while the pooled NLR was 0.09 (95% CI: 0.04–0.22). The pooled DOR was 26 (95% CI: 9–75). To further evaluate diagnostic accuracy, SROC curves were plotted, revealing an AUC of 0.85 (95% CI: 0.82–0.88) for mNGS (Figures 2, 3, 4).

FIGURE 2.

Forest plot of sensitivity and specificity of mNGS for the diagnosis of IDS.

FIGURE 3.

Forest plot of positive likelihood ratio (PLR) and negative likelihood ratio (NLR) of mNGS for the diagnosis of IDS.

FIGURE 4.

The SROC curve of mNGS diagnosis of IDS.

3.3.3. Subgroup and Sensitivity Analyses

To investigate potential sources of heterogeneity, subgroup analyses were conducted based on the diagnostic criteria used in the literature (histopathology/composite diagnostic criteria), sample types (tissue/tissue and others), pathogen types ( Mycobacterium tuberculosis and non‐tuberculous mycobacteria/ Mycobacterium tuberculosis and others), and sequencing platforms (Illumina/MGISEQ and others). The various subgroups did not significantly affect sensitivity and specificity. Diagnostic criteria, sample types, pathogen types, and sequencing platforms are all considered non‐heterogeneous sources. Sensitivity analyses were conducted to evaluate the stability of the study findings. The results of the sensitivity analysis indicate that excluding a single study from the analysis did not significantly alter the results of the meta‐analysis, suggesting stable and reliable outcomes. The results of the subgroup and sensitivity analyses are presented in Table 3 and Figure 5.

TABLE 3.

Subgroup analysis of mNGS for diagnosis of IDS.

Variables	Subgroup	Number of studies	Sensitivity (95% CI)	p	Specificity (95% CI)	p
Diagnostic criteria	Histopathology	6	0.92 [0.82–1.00]	0.15	0.64 [0.46–0.82]	0.98
	Composite diagnostic criteria	9	0.96 [0.92–1.00]		0.57 [0.43–0.72]
Specimen type	Tissue	7	0.95 [0.89–1.00]	0.44	0.61 [0.44–0.78]	0.56
	Tissue and others	8	0.94 [0.88–1.00]		0.58 [0.43–0.74]
Pathogen type	Mycobacterium tuberculosis and others	12	0.94 [0.89–0.99]	0.72	0.61 [0.47–0.74]	0.87
	Mycobacterium tuberculosis /non‐tuberculous mycobacteria	3	0.96 [0.90–1.00]		0.58 [0.35–0.82]
Sequencing platforms	Illumina	5	0.94 [0.87–1.00]	0.70	0.61 [0.42–0.80]	0.63
	MGISEQ and others	10	0.95 [0.89–1.00]		0.59 [0.44–0.74]

FIGURE 5.

Sensitivity analysis of mNGS for the diagnosis of IDS.

3.3.4. Publication Bias Evaluation

Publication bias was assessed using Deeks' funnel plot. The outcome of the asymmetry test yielded a p‐value of 0.73, indicating no evidence of publication bias (Figure 6).

FIGURE 6.

Deeks' funnel plot of mNGS for the diagnosis of IDS.

4. Discussion

Previously, mNGS was primarily used for the clinical assessment of sterile body fluids, including cerebrospinal fluid, blood, and joint effusions [9, 10]. However, the application of mNGS has been relatively limited to non‐sterile tissues, such as infected intervertebral discs or paraspinal abscesses. Recent studies have indicated that mNGS has the potential to enhance the diagnostic accuracy of bone and joint infections, serous cavity infections, skin and soft tissue infections, endophthalmitis, urinary tract infections, and other tissue infections [9, 11]. While there are several reports on the application of mNGS for the diagnosis of IDS, its diagnostic value remains unclear, and no meta‐analysis on the diagnosis of IDS using mNGS has been published to date.

This study included 15 research papers to assess the effectiveness of mNGS in diagnosing IDS using DOR and SROC curves. The meta‐analysis found a pooled sensitivity of 0.95 (95% CI 0.88–0.98) and a pooled specificity of 0.60 (95% CI 0.48–0.71) for mNGS in diagnosing IDS. The SROC curve, based on the sensitivity and specificity data from these studies, resulted in an AUC of 0.85 and a DOR of 26 (95% CI 9–75), confirming the efficacy and clinical value of mNGS for diagnosing IDS. However, with a PLR less than 10 and an NLR greater than 0.1, the diagnostic strength of mNGS appears more effective in ruling out the disease rather than confirming it. Therefore, a positive mNGS result does not conclusively indicate infection and should be interpreted alongside other clinical indicators for a definitive diagnosis.

The diagnosis and treatment of IDS remain a significant challenge for clinicians. The gold standard for the diagnosis of IDS is the detection of the causative pathogen [2], and the selection of appropriate antibiotics based on microbial drug susceptibility is fundamental to the effective treatment of IDS. Tissue culture technique (TCT) is not able to identify all infectious etiological agents. These methods are limited by the volume of sample and growth in culture media resulting in false negativity. Also, the sensitivity of culture decreases, if tissue culture is collected after antibiotics initiation. The mNGS is an emerging method for the detection of pathogenic microorganisms, characterized by rapidity, high accuracy and broad coverage of pathogenic bacteria, especially for rare pathogens such as mycoplasma, fungi and brucella, etc. [8, 12]. Additionally, mNGS offers an advantage over conventional methods, especially when rare pathogens are implicated or the patient is on antibiotics [11, 13, 14]. Li's meta‐analysis provided pooled estimates for TCT with a sensitivity of 0.34 (95% CI, 0.27–0.43), specificity of 0.93 (95% CI, 0.79–0.98), and an AUC of 0.59 (95% CI, 0.55–0.63) [15], which is consistent with our conclusion. Although mNGS shows higher sensitivity and overall diagnostic accuracy compared to TCT, its lower specificity indicates that it might be most effective when used in conjunction with traditional methods to improve diagnostic reliability in spinal infections.

Qu et al. conducted a meta‐analysis of the diagnostic performance of mNGS in the central nervous system (CNS), reporting that the pooled sensitivity of mNGS for central nervous system infections was 77%, and the specificity reached 96%. The area under the SROC was 0.91 [16]. In Tan et al.'s meta‐analysis, it was reported that the pooled diagnostic sensitivity and specificity of mNGS for prosthetic joint infection (PJI) were 93% and 95%, respectively, with a PLR of 18.3 and a NLR of 0.06, and the AUC reached 0.96 [17]. Compared to mNGS for diagnosing IDS, mNGS shows excellent diagnostic performance in two diseases. Specifically, it has the highest sensitivity and specificity for diagnosing PJI. For IDS diagnosis, mNGS displays higher sensitivity than for CNS diagnosis but lower specificity.

Spearman's correlation analysis in this study did not reveal any threshold effect, and Deeks' funnel plot did not indicate significant publication bias. However, substantial heterogeneity was observed in the sensitivity across studies. To explore potential heterogeneous factors, the study conducted subgroup analyses, which included diagnostic criteria, sample types, bacterial species, and sequencing platforms that varied among different studies. Although the analysis results indicated that these factors could not explain the observed heterogeneity, this study has several limitations: (1) In the subgroup of sample types, ‘Tissue and others’ includes various types such as pus, blood, brain fluid, and lavage fluid, but this subgroup is still dominated by tissue samples, which may lead to insignificant heterogeneity between subgroups. In addition, the research results may also be affected by the site and timing of sample collection. Scholars like Anderson found that the positive rate of puncture pus is higher than that of bone tissue or intervertebral disc tissue, and the positive rate of upper thoracic spine biopsy samples is even higher [18]. The Alessandro team's research showed that there is no significant difference in the volume of the collected tissue or the biopsy site (bone tissue vs. intervertebral disc), but sampling in the acute phase can significantly improve the microbial detection rate [19]. (2) The inherent characteristics of pathogens, such as the thick outer wall of some pathogenic microorganisms or the difficulty in breaking the cell wall during nucleic acid extraction (e.g., mycobacteria), lead to incomplete release of nucleic acids, thereby causing insufficient performance of mNGS detection. In our study, the ‘ Mycobacterium tuberculosis /Non‐tuberculous mycobacteria’ subgroup only included three literatures. The small number of included literatures may affect the effective evaluation of heterogeneity. (3) Variability in any of the steps of NGS sequencing (e.g., sample type, sample storage, DNA extraction, PCR amplification, sequencing technology, read length, and/or bioinformatics analysis) can lead to data variability [20]. In the studies we included, the sequencing platforms mainly used were Illumina series and BGI series. The BGISEQ‐500 sequencing platform is similar to the general NGS workflow and step‐by‐step sequencing procedures of the Illumina series; however, the templates of the two are significantly different. The subsequent DNA nanoball technology of the BGISEQ‐500 platform is specially used for library construction, which is different from the library construction protocols used by the Illumina series. BGISEQ‐500 uses both single‐end (SE) and paired‐end (PE) modes, similar to the latest Illumina's HiSeq 4000. The data show that among the short‐read sequencing platforms, Illumina's HiSeq 4000 and HiSeq X10 platforms provide the most consistent and highest genome coverage, while BGI's BGISEQ‐500 and MGISEQ‐2000 platforms provide the lowest sequencing error rate [16]. It can be seen that although no statistically significant difference was found in our subgroup analysis, different models are included in the same platform, which hinders the in‐depth analysis of related heterogeneity.

This study has several limitations. First, the databases searched were not comprehensive, as they did not include specialized databases such as Web of Science and Embase, and the search strategy was not rigorous, leading to a risk of incomplete inclusion of literature. Second, most of the included studies were retrospective analyses, with only one prospective study, indicating lower methodological quality. Third, all the study populations were Chinese, which may limit the generalizability of the results. Finally, the study only considered Chinese and English literature, potentially introducing language bias.

5. Conclusion

The mNGS offers high sensitivity and overall diagnostic accuracy, but its lower specificity poses a risk of false positives. Its primary diagnostic value lies in its ability to serve as a rapid screening tool for disease exclusion. However, for diagnosing IDS, it is essential to integrate other clinical indicators for a comprehensive assessment to confirm the diagnosis. The current studies are still limited, and future research should focus on more prospective, large‐scale, multicenter, high‐quality studies to scientifically and objectively evaluate the effectiveness of mNGS in diagnosing IDS.

Author Contributions

Xuejiu Cai and Honglei Yi: investigation, screening publications, visualization, Writing – original draft. Kun Chen and Jianqiang Dai: methodology, software, data curation. Jianhua Yi, Bing Tu and Yidan Wang: methodology and data curation. Jia Li and Jingshen Zhuang: conceptualization, supervision, writing – review and editing. All authors contributed to the article and approved the submitted version.

Conflicts of Interest

The authors declare no conflicts of interest.

Cai X., Yi H., Chen K., et al., “Meta‐Analysis of the Application Value of Metagenomic Next‐Generation Sequencing Technology in the Diagnosis of Infectious Diseases of the Spine,” JOR Spine 8, no. 4 (2025): e70134, 10.1002/jsp2.70134.