Literature Review of Pathogen Agnostic Molecular Testing of Clinical Specimens From Difficult-to-Diagnose Patients: Implications for Public Health

To better identify emerging or reemerging pathogens in patients with difficult-to-diagnose infections, it is important to improve access to advanced molecular testing methods. This is particularly relevant for cases where conventional microbiologic testing has been unable to detect the pathogen and the patient's specimens test negative. To assess the availability and utility of such testing for human clinical specimens, a literature review of published biomedical literature was conducted. From a corpus of more than 4,000 articles, a set of 34 reports was reviewed in detail for data on where the testing was being performed, types of clinical specimens tested, pathogen agnostic techniques and methods used, and results in terms of potential pathogens identified. This review assessed the frequency of advanced molecular testing, such as metagenomic next generation sequencing that has been applied to clinical specimens for supporting clinicians in caring for difficult-to-diagnose patients. Specimen types tested were from cerebrospinal fluid, respiratory secretions, and other body tissues and fluids. Publications included case reports and series, and there were several that involved clinical trials, surveillance studies, research programs, or outbreak situations. Testing identified both known human pathogens (sometimes in new sites) and previously unknown human pathogens. During this review, there were no apparent coordinated efforts identified to develop regional or national reports on emerging or reemerging pathogens. Therefore, development of a coordinated sentinel surveillance system that applies advanced molecular methods to clinical specimens which are negative by conventional microbiological diagnostic testing would provide a foundation for systematic characterization of emerging and underdiagnosed pathogens and contribute to national biodefense strategy goals.


Introduction
A dvanced molecular techniques have demon- strated utility in the identification of pathogens in difficult-to-diagnose clinical situations where infection is suspected, yet routinely available microbiological testing yields negative results. 1 Recent examples include patients with meningitis and encephalitis, 2 undiagnosed respiratory illnesses, 3 prosthetic joint infections, 4 and an international outbreak of acute severe hepatitis in children. 5Techniques, such as metagenomic next generation sequencing (mNGS) (also referred to as shotgun metagenomics), have been used for pathogen agnostic testing in research settings; however, these are not yet widely used as a routine diagnostic tool in clinical laboratories due to inequitable availability, variations in sensitivity, resource constraints, regulatory burdens, and concerns for contamination or detection of nonclinically relevant pathogens. 6,7These techniques are potentially underutilized as a strategy for the early detection of emerging or reemerging human pathogens.However, they could be used to strengthen clinical care and public health preparedness and response if used appropriately and more widely.
Even after extensive testing in clinical laboratories, up to 60% of patients with serious infections may not have a causative pathogen identified. 1,8Current microbiological methodologies and diagnostic workflows are often limited in providing accurate diagnoses because they largely rely on pathogen-specific tests and are not always timely when searching for rare and dysgonic pathogens that are difficult to culture.Regardless of the type of microbe, pathogens can be detected directly from clinical specimens using mNGS. 9,10mNGS can characterize all nucleic acid in any specimen type and can detect rare organisms that may otherwise be missed by conventional diagnostic assays. 11,12urthermore, mNGS techniques have been used to increase diagnostic yield for patients with acute 2 or subacute infections 4 or who are immunocompromised. 13lthough mNGS holds much promise in terms of the depth and breadth of its applicability, there are numerous challenges to expanding its usage in clinical scenarios.Barriers such as the high initial and maintenance costs; need for highly trained specialists to conduct, analyze, interpret, and collate testing; long turnaround time and elaborate work flows; variations in analytic sensitivity; and difficulties with result interpretation all represent significant challenges to widespread adoption. 14,15Additionally, there are multiple limitations in clinical pathogen detection, including the shortage of complete genomic sequences for certain organisms; the need for clear criteria, guidelines, and interpretive standards; and the decreased test performance for some specimens (eg, intraocular fluid, joint fluid). 14Furthermore, the difficulty of validating mNGS assays according to Clinical Laboratory Improvement Amendments (CLIA) regulations may limit the number of laboratories that can perform mNGS.
However, as usage of mNGS increases in the postpandemic era, the costs may decrease, 16 and more comprehensive database libraries should enhance interpretation, sensitivity, and clinical translation.Additionally, the development of guidelines may provide a framework to define clinical utility and criteria. 15Despite these recognized challenges, mNGS has the potential to be a timely, unbiased pathogen agnostic approach to detection that does not depend on predetermined targets or culture techniques and can be deployed in pathogen agnostic clinical scenarios to identify novel emerging and reemerging pathogens. 1,6,7,14,15iomedical literature offers a robust and deep source of critical information, which, when paired with other sources and selectively curated, can be used to identify academic collaborators, track the performance of emerging technologies, and develop best practices to support a national pathogen agnostic sentinel surveillance system to advance biosecurity.We conducted a targeted review of the published biomedical literature to improve the understanding of how mNGS is used in clinical, research, and pathogen surveillance activities.Findings from this literature review will increase understanding of the potential for coordinating data sharing and reporting from entities that perform these tests to support regional and national public health surveillance and provide insight regarding threats from emerging pathogens.

Methods
For the purposes of this review, the term ''metagenomics'' was defined as a pathogen agnostic, unbiased, sequencing technique that can produce sequence data of all microorganisms found in each specimen.Additionally, 16S/18S/ ITS amplicon-based sequencing was included due to its importance in surveillance for emerging pathogens.
A literature search and selection process were used to address the aims of this study.We searched for peerreviewed literature published before February 2023 and selected illustrative examples (''exemplars'') of publications that reported notable discoveries using a pathogen agnostic approach on human clinical specimens.

PATHOGEN AGNOSTIC MOLECULAR TESTING OF CLINICAL SPECIMENS
For the National Library of Medicine National Center for Biotechnology Information PubMed literature database, we developed a search strategy that included articles published from 1966 to February 2023, using multiple keywords and phrases.Terms were organized using ''OR'' and ''AND'' operators to represent entities (''who''), activities (''how''), and findings (''what'') (see Table 1).The search focused almost exclusively on pathogen agnostic approaches from 2006 to February 2023, specifically mNGS, and was limited to infectious disease among humans.The initial search of PubMed was conducted in October 2022, and a final scan for relevant publications was performed in February 2023.
The search resulted in an initial retrieval of more than 4,000 records.A formal, systematic screening of the initial literature retrieval was not feasible due to the volume of records retrieved.Given the overall exploratory objective of the review, an informal selection process was used to narrow the initial retrieval to a subset of publications that focused on pathogen agnostic approaches and were limited to infectious disease in humans.This resulted in a subset (n=104) that was representative of the main search terms.Additional publications were identified (n=56) through referrals by partners or through informal and exploratory searching including review of references, Google searches, and author searches.This subset was then further examined to identify a representative sample set-the ''exemplars.''For the selection of exemplars, the search was limited to case reports, case series, and studies involving clinical trials, research, and surveillance activities.Exemplars had to include 1 or more human subjects or specimens obtained from humans experiencing illness and describe the use of mNGS to detect or identify 1 or more pathogens.Opinion and theoretical pieces and publications on foodborne illness surveillance and antimicrobial resistance were excluded.The review was limited to empirical studies to provide insight into practical and actual applications and use of the diagnostic approach (inclusion of nonempirical publications generally provides insight into themes or perspectives and requires a wholly different, qualitative, or thematic approach to synthesis).US and non-US-based publications were considered.A total of 34 exemplars were selected for data abstraction and summary.
We developed a study characteristics and content data extraction sheet in Microsoft Excel to capture data elements such as author names, clinical presentation, molecular techniques, study type, and laboratory protocols, as well as other data elements of interest (eg, patient immune status and travel history) as available.To supplement the entity analysis, information about funding sources was extracted from the funding and acknowledgments sections of the peer-reviewed articles.

Results
Of the 34 exemplars, most publications were case reports or case series (n=21 publications, Table 2) while other studies involved prospective clinical trials, surveillance studies, research programs, or outbreak situations (n=13 publications, Table 3).Case reports were most often on single patients (n=17) or a series of patients (n=4) (Table 2).Metagenomic techniques were used to successfully identify pathogens directly from patient specimens, including cerebrospinal fluid, 13,[17][18][19][20][21][22][23][24][25] respiratory secretions, [26][27][28][29][30][31]  and soft tissue or bone and joint tissues/synovial fluid, 32,34 heart valve tissue, 35 and urine. 36A review of the methods sections revealed significant heterogeneity in sensitivity and specificity of testing due to variability in study design, reference standards, clinical factors, and technical variation among the laboratories performing the testing.For exemplar US publications, laboratories of the University of California at San Francisco were featured prominently as testing laboratories.A total of 13 publications were categorized as using metagenomic or traditional sequencing techniques (eg, targeted sequencing and fragment analysis) and performed for prospective clinical trials, surveillance studies, research programs, or outbreak situations (Table 2).These studies included patients with respiratory symptoms or infections (including tuberculosis), neurological syndromes, gastroenteritis, and fever syndromes.A study involving respiratory infections utilized knowledge from prior identification of a pathogen (eg, polyomavirus or human parainfluenza virus) that resulted in the screening of large numbers of specimens from patients with respiratory infections to look for those viruses. 39Another study involved identifying the etiology of pneumonia, 41 a condition that is often managed without specific microbiological diagnoses.Several studies were noted to be conducting surveillance on patients with central nervous system infection syndromes where the focus was on attempting to identify a microbiological etiology for the infection. 43,48Several studies implemented advanced molecular testing in outbreak situations to identify a potential pathogen, 42,45 including a study from China to identify the novel SARS-CoV-2 virus. 47esults from mNGS supported the identification of multiple new pathogens. 3,17,38,47Researchers used mNGS to detect single pathogens, 18,24,26,33,34 multiple pathogens, [29][30][31]39,48,42 and rare pathogens. 22,28 Reults from mNGS were also used to correlate with clinical symptoms, 42 test specimens from patients with atypical or no symptoms, 21,40,49 and facilitate more accurate 28 and improved clinical care.50 Additionally, mNGS analyses yielded several significant detection advantages, including detection from a very ill or immunocompromised patient, 13,19,20 when all other advanced tests and cultures performed were negative 27,32,33,35,50 or inconclusive. 26,50Metagenomic sequencing was used to determine associations between genotype and clinical phenotype, 46 elucidate etiologies, 23,41,43,44,46 investigate a pathogen's epidemiology by subsequent whole-genome capture, 50 and characterize clusters of divergent strains.51 Overall, mNGS exhibits good sensitivity and specificity for diagnosing central nervous system infection and diagnostic performance during clinical application by assisting in identifying the pathogen.However, the efficacy remains inconsistent, which warrants subsequ-ent studies to further improve performance during clinical application. 49,52 Insome cases, diagnostic sensitivity was increased by more than 18% to 20%, 49,51 without false positives, 51 independent of specimen type, 53 and without the need for viable or culturable pathogens.35 However, mNGS can be less sensitive than polymerase chain reaction for detecting very low-level infections, 48 and as with other genomic methods, mNGS output still requires careful interpretation and substantial bioinformatics support. 45

Discussion
Advanced molecular methods, such as mNGS, can provide unbiased, pathogen agnostic testing for human clinical specimens.mNGS methods are available and can be leveraged to support clinicians and patients in identifying a potential pathogen in difficult-to-diagnose patients.However, it is not clear how widely and equitably these modalities of testing are available to clinicians and patients who would benefit from such testing.These techniques have the potential to drive significant improvements in patient care and health outcomes by facilitating microbiological diagnoses in patients whose underlying etiology may be difficult to identify using more conventional and readily available laboratory testing. 7,54he literature reviewed was generally robust regarding descriptions of techniques and information about testing of clinical specimens using mNGS.The literature review of exemplar publications identified 3 general types of pathogen detection: (1) clinical infections caused by previously known pathogens that were difficult to diagnose using conventional testing of patient specimens, (2) presence of previously known pathogens in a site that may represent a new manifestation of infection, and (3) potential new human pathogens that were previously unknown that may be causing the infection.
Broadly, there appear to be at least 4 clinical situations and scenarios where mNGS testing has been implemented: (1) individual patients with acute, subacute, and chronic clinical syndromes where infection is suspected, yet conventional and available microbiological testing yields no known pathogens-this may include identifying coinfections with pathogens in patients that may allow for targeted therapies; (2) diagnosing 1 or more patients who may present with a clinical syndrome that has recently been described to be caused by a new or previously unknown pathogen; (3) prospective clinical trials, retrospective casecontrol studies, surveillance studies, and research programs; and (4) outbreak situations where metagenomic sequencing is systematically implemented on patient specimens to establish microbiological diagnoses and understanding of the clinical syndrome.
From a review of the biomedical literature, it is not clear how most individual patients benefited from the advanced testing, as journal publications could take several months or years to prepare and publish in the public domain after peer With wider availability of metagenomic testing and improvements in the techniques themselves to overcome challenges, such as specimen contamination and interpretation of results, we envision several clinical and public health benefits of periodic reviews and incorporation of results from metagenomic testing on human clinical specimens.These include developing targeted diagnostic assays using mNGS identification of a potential pathogen in the early days of an outbreak or pandemic to increase the effectiveness of the clinical and public health response to emerging pathogens. 55Knowledge gained after an outbreak has subsided may also support understanding and preparation for subsequent outbreaks, including developing, incorporating, and deploying new and emerging pathogenspecific diagnostic targets into US Food and Drug Administration-approved multiplex testing panels, based on knowledge gleaned from metagenomic testing. 5Finally, systematic surveillance for emerging and reemerging pathogens in specific populations (eg, patients with immunocompromising conditions, nursing home patients, people who are incarcerated, persons experiencing homelessness, returning international travelers) may be sentinels for such pathogens.
Currently, there is no nationwide pathogen agnostic surveillance system for emerging or reemerging human pathogens in the United States.Developing a coordinated sentinel surveillance system that deploys advanced molecular methods to clinical specimens from patients whose specimens are negative by conventional diagnostic would provide a foundation for systematic detection of emerging, reemerging, and underdiagnosed pathogens.In addition, mNGS has the potential to identify bioengineered or novel naturally occurring strains of pathogens that may evade traditional diagnostics.If appropriately resourced and scaled after demonstrating utility and cost-effectiveness, such a system could significantly strengthen proactive preparedness and response capabilities and be responsive to national biodefense strategy goals. 56oordination and integration between systems supporting frontline clinicians and local, regional, and federal government are important components of a nationwide surveillance system that successfully monitors for, identifies, and responds to emerging and reemerging pathogens.Such a surveillance system would need to be equitably distributed across the country with availability at local and regional levels, and it should be agile enough to provide support for specific and vulnerable populations.Therefore, ongoing detection and discovery for surveillance of emerging and reemerging human pathogens are central to public health preparedness and outbreak detection and response at regional, national, and international levels.
The literature review has inherent limitations, such as publication bias because negative results are rarely published, delays in publicly reporting potential new biothreats, and reliance on publicly available information.The literature review alone leaves gaps in understanding of processes, pipelines, and key details about the mechanics of how specimens are identified, transported to specialty laboratories, and tested.Published information was not always able to clarify whether the mNGS tests were validated by the laboratory, whether the tests were regulated and approved through CLIA, thus allowing for reporting to clinicians and patients, or whether results were used for reflex testing with CLIA-approved assays.The search was not a systematic review, which likely resulted in an incomplete list of recent publications that may potentially impact findings.Additionally, the time needed to prepare and submit research articles for publication in a peer-reviewed journal creates a temporal gap that limits the usefulness of the information for literature response in a real-world setting.This review did not specifically target preprint servers.
There may also be institutions and practitioners who use advanced molecular testing but do not publish their results or findings; these are most likely small-scale, personal professional networks that play a role in routing point-of-care patient specimens for metagenomic testing, or the growing field of private commercial entities offering advanced molecular testing.Furthermore, some agencies and institutions may be unable to publish in scholarly outlets (eg, national security programs, commercial entities, public health agencies).Finally, this review does not account for the vast literature and experience available in the One Health arena where pathogens are routinely discovered in animal reservoirs. 57,58nderstanding the extent of the availability of such testing modalities to clinicians across the country with respect to academic, clinical, public health, and commercial laboratories is necessary.It is also important to understand the flow of patient specimens from the primary point of microbiological testing to laboratories that perform these specialized tests, the status of validation and CLIA-approval for these tests, the exchange of patient information and clinical details along with the specimen, and mode of payment (eg, insurance coverage) for these tests.Reflex testing with CLIA-approved clinical diagnostic assays is also required to inform patient care.It is essential to understand the equitable availability of these tests for patients who would most benefit across geographical, racial and ethnic, and social demographic barriers.As these technologies improve and are more widely used, consistent and standardized policies and guidelines are critical for clinicians, researchers, and public health professionals to recognize the limitations of these techniques, protect patient privacy, ensure accurate interpretation of results, and support the appropriate use of the results to benefit individual patients and public health. 54,55

Conclusion
Advanced pathogen agnostic testing using metagenomic sequencing is being performed in a limited number of settings to directly support patient care, and for research and surveillance purposes.Improved understanding of the current usage and availability of advanced molecular testing techniques in a variety of healthcare settings that directly provide care to patients with pathogen agnostic testing could help inform public health surveillance and intervention strategies.Overcoming limitations and barriers to wider adoption of advanced molecular methods for testing clinical specimens would provide a foundation for systematic characterization of emerging pathogens.Additionally, deploying advanced molecular methods to test clinical specimens from patients whose specimens are negative by conventional microbiological diagnostic testing can contribute to a more coordinated sentinel surveillance system and is responsive to national biodefense strategy goals.

Table 1 .
skin PubMed Search Strategy

Table 3 .
Exemplar Publications of Pathogen Agnostic Testing Using Advanced Molecular Techniques (n= 13)