Breathing New Life into Diagnostics

The healthcare community has always taken respiratory infections seriously – but the COVID-19 pandemic underscored just how central these infections are to clinical care in general. Respiratory infections are a leading cause of sepsis and, in many healthcare systems, the largest driver of antibiotic use. They are also the infection type most likely to cause a pandemic.

With so many ramifications for patients, it is imperative that clinical laboratories deploy the best possible tools for testing samples associated with respiratory infections. Today, laboratorians use a broad range of technologies – from cultures and singleplex PCR tests to syndromic molecular panels, and from low-throughput, sample-to-answer systems to high-throughput, industrial-scale operations. Each new pathogen that must be detected can mean implementing a broader panel test, adding a pathogen-specific kit, or even bringing in an entirely new diagnostic platform.

In the future, there may be a more straightforward approach that would have the inherent ability to detect all pathogens without requiring any new kits or equipment. Metagenomic testing is a sequencing-based alternative that can detect all organisms in a patient sample. And because no prior hypothesis is needed about the causal pathogen, metagenomics offers an unbiased view that can catch co-infections, hidden causes of infection, and other clinically relevant information missed by conventional diagnostic approaches. In addition, it can detect genetic markers of antimicrobial resistance to inform treatment selection.

The age-old approach

Even with the growing popularity of PCR-based diagnostic tools, the go-to test for respiratory infections hasn’t changed in nearly a century. Culture-based testing is cheap and easy, with widely available consumables and well-established methods for preparing the test and for reading results. Indeed, cultures are still regarded as the diagnostic gold standard; unfortunately, they take days to generate results, and there are many slow-growing or even unculturable organisms that can cause respiratory infections. 

Beyond pathogen identification, it takes even longer for culture-based tests to produce information about antimicrobial resistance. In healthcare systems where respiratory infection testing is based solely on cultures, there is no opportunity to adjust a patient’s treatment – such as to de-escalate antibiotic treatment when the pathogen is found to be a virus or to shift a patient from a broad-spectrum antibiotic to a targeted antibiotic in a clinically relevant time frame. Not only does it lead to poorer patient outcomes, but it also contributes to the growing epidemic of antibiotic resistance.

Many clinical laboratories have adopted molecular diagnostic platforms to reliably identify pathogens faster and provide patients with the right treatment as soon as possible. But these tests also have their disadvantages. Labs have a set menu of tests – even when send-out tests are included, which means that only the usual pathogens can be detected; in other words, rare, novel, or emerging pathogens are missed by these tests. To manage costs, most labs run these tests serially. If the causal pathogen is the very last on a long list of suspects, it could take just as long to get the answer as it would with a culture-based test. In addition, most molecular assays focus on pathogen identification and do not profile antimicrobial resistance markers.

The metagenomic era
 

The idea of turning to sequencing-based metagenomics to identify all viruses, bacteria, and fungi in a community is not new — scientists have been using this approach for years to characterize microbial communities in soil, deep ocean vents, biofilms, and more. But metagenomics can just as effectively be deployed to look within ourselves.

In recent years, clinical metagenomics has been evaluated for various types of patient samples and has proven an efficient and useful technique for detecting microbes found in and on humans. Until now, though, there have been good reasons not to roll this method out more broadly. For example, most studies relied on short-read sequencing technologies. The short snippets of DNA produced by these tools can be challenging to align for an accurate pathogen identification, and they are not amenable to connecting genetic markers of resistance back to their microbial hosts. In some countries, reimbursement issues have arisen when metagenomics workflows report all organisms found in a sample rather than just the one or ones most likely to be responsible for infection.

A different technology and an altered approach may be what’s needed to make metagenomics a valuable component of the respiratory testing toolbox. Long-read sequencing platforms can help address the alignment challenges associated with short-read data; they can also fully resolve more complex genomes, such as those characteristic of fungal pathogens. With long-read data, it is also possible to link resistance-carrying plasmids to their host genomes.

Nanopore-based sequencing, which can be used to generate long or short reads as needed, can also produce data very quickly. In a recent pilot project at the Guy’s and St. Thomas’ Hospital NHS Foundation Trust in London, a clinical laboratory team evaluated nanopore sequencing to support a rapid respiratory metagenomics workflow (1). They tested nearly 130 samples from more than 85 individuals with lower respiratory infections, setting detection thresholds equivalent to culture-based testing to avoid reporting microbes that were unlikely to be clinically relevant. For most samples, results were reported to the clinical care team on the same day the sample was collected. Interestingly, nearly half of the results led to shifts in antimicrobial selection (in some cases escalating and in others de-escalating the initial treatment choice). Several unexpected organisms and cases of co-infections were reported; these would not have been found with conventional tests.

On trial
 

Clearly, further investigation is warranted to determine whether rapid metagenomics could be a useful alternative to standard respiratory testing approaches in clinical laboratories. Clinical trials will be helpful in understanding whether the treatment-selection benefits seen in this pilot project will translate to other laboratories and broader patient populations. As these evaluations occur, there will need to be assessments of the ideal reporting thresholds to ensure that causal pathogens are included in results and that most other microbes are not. If rapid metagenomics realizes its promise, technology developers will have to do their part to facilitate widespread adoption by automating sample preparation, analysis, and reporting.

If these initial successes with rapid metagenomics for respiratory testing hold up in larger trials, it could be a much-needed solution to growing challenges in healthcare. With more respiratory pathogens circulating in the general population, the old approach of testing for one microbe at a time becomes less feasible. With the rise of antimicrobial resistance, it is more important than ever to adopt rapid tests that can inform responsible treatment selection in a matter of hours. Rapid metagenomics could also identify a hidden burden of infections that are clinically significant but cannot be detected with current tools. 

Overall, rapid metagenomics has the potential to serve as a one-and-done test for respiratory infections to ease the testing burden on clinical laboratories and help physicians deliver better outcomes for their patients.