are there any applications of NGS on infectious disease?

[dukefan]

I am doing a course project on NGS applications in infectious disease.

It seems that there are no NGS products available on the market and nothing under development. I wonder why?

Is NGS too expensive for infectious disease? Or do you rarely need to target many markers that exceed the multiplexity of a PCR test?

Thanks a lot!

[Brian Bushnell]

I think NGS is ideal for infectious diseases; at least, those that are blood-borne. Whether there are 'products available on the market' is irrelevant in terms of technology, isn't it? Costs are monotonically declining, and what you can do 'on the market' is up to your own creativity.

I suspect that the cost of NGS is trivial compared to the cost of misdiagnosing a treatable disease... though of course that depends on the local medical rates.

[GenoMax]

It is true that current utility of NGS is perhaps more in surveillance when a completely new pathogen is suspected to be the cause. Though there is nothing to prevent anyone from sequencing a previously known E.coli strain and confirming its identity. Sequencing the pathogen could prevent use of a wrong antibiotic if the resistance profile can be determined based on sequence of the plasmids present.

http://www.illumina.com/applications...veillance.ilmn

http://res.illumina.com/documents/pr...ase-review.pdf

http://investor.pacificbiosciences.com/releasedetail.cfm?ReleaseID=793199

[mbblack]

If you search the literature, you will actually see that the CDC uses NGS for identification of virus and bacteria in infection outbreaks. In the 2010 yellow fever outbreak in Uganda they used NGS to identify yellow fever virus and quickly generate a nearly complete genome of the specific strain in patients (I think that was published just a year later, in 2011 sometime).

But many of those sorts of publications are in clinical medical journals, not traditional genetics or genomics type publications since the authors target audience is the clinical and medical community, not biologists and informaticians (e.g. Journals like Clinical Pharmacology and Theurapeutics or Clinical Microbiology and Infection have had NGS publications at times).

[dukefan]

thank you all!

I did some search and it seems that the applications of NGS in infectious disease is limited to epidemiology study of outbreaks and last resort for strange disease. There is no routine infectious disease diagnostics using NGS.

it kind of make sense; when you can multiplex with PCR or microarray, there is really no reason to use NGS.

NGS is great when you don't know what you are looking for but how do you get that kind of open-ended test product approved? i can't possibly validate every pathogen under the sun.

it seems NGS for infectious disease will not happen any time soon.

[dukefan]

I think NGS is ideal for infectious diseases; at least, those that are blood-borne. Whether there are 'products available on the market' is irrelevant in terms of technology, isn't it? Costs are monotonically declining, and what you can do 'on the market' is up to your own creativity.

I suspect that the cost of NGS is trivial compared to the cost of misdiagnosing a treatable disease... though of course that depends on the local medical rates.

I respectfully disagree.

I kind of think NGS is not good for blood borne disease since you really need the results very very quickly. So it is technologically demanding.


Also, I think whether there are products on the market is relevant. I did some search and found some good press releases on Illumina and Roche developing NGS based product for infectious disease. But then nothing happened, I wonder why. There must be some good reasons that NGS is just not good for infectious disease, otherwise we would have seen approved products. these are capable companies.

[mbblack]

thank you all!

I did some search and it seems that the applications of NGS in infectious disease is limited to epidemiology study of outbreaks and last resort for strange disease. There is no routine infectious disease diagnostics using NGS.

it kind of make sense; when you can multiplex with PCR or microarray, there is really no reason to use NGS.

NGS is great when you don't know what you are looking for but how do you get that kind of open-ended test product approved? i can't possibly validate every pathogen under the sun.

it seems NGS for infectious disease will not happen any time soon.

Generally speaking, regulatory and oversight agencies do not approve diagnostic tests that do not have clearly defined basis for positive diagnosis, as well as estimates of false positives and even more importantly, false negatives. So no, its nearly impossible to get a formally approved diagnostic test past regulatory oversight when the method is in reality a fishing expedition as far as final results go.

Remember when you are talking about anything to be applied to actual patients in a clinical setting, there is no country in the world where you are not dealing with a highly politicized and beurocratic oversight environment. If you want to see something used in clinical settings, it needs formal regulatory approval.

Second, NGS is relatively slow, labor intensive, and relatively expensive. Clinicians want relatively simple, rapid and inexpensive tests, and usually ones that are clear cut yes/no outcomes (e.g is agent "x" the infectious agent we're dealing with in this particular patient, yes or no). They also ideally want their tests in uniform, canned package systems - one single system and protocol used by any and all clinicians no matter where they may be. Clinical medicine is not greatly interested in exploratory scientific tools where the methods often have to be custom tweaked for any given experimental system, nor where the data requires extensive expert analysis to interpret.

Now, if were to invent the NGS tricorder, you'd get clinicians attention! Although I bet it would still take decades to convince the regulatory agencies to actually approve its use.

[Brian Bushnell]

PacBio has a pretty quick turnaround time. And Oxford Nanopore may not be a tricorder, but it's not that far off; small, inexpensive, and fairly quick. Hospitals in the US charge ridiculous prices for procedures, and the costs and labor associated with these would be much lower than for example an MRI, which is fairly routine.

The regulatory issue is of course a big problem in reality, but not one that I consider relevant to the question of whether NGS would be useful.

[mbblack]

So what do you, Dukefan, actually mean by an NGS "product" specifically for infectious diseases? Illumina and others do indeed market their existing products for infectious disease monitoring, but its just standard whole genome sequencing not something specific to infectious disease agents. WGS is WGS - nothing special about using it in the context of infectious disease research.

If you actually mean patient diagnostic testing, then yes, I would say the actual hurdle is the regulatory process and requirements. Those are the very people one needs to convince, not so much of usefulness, but of direct clinical efficacy. So, for example if using WGS from patient samples to diagnose infection by a foreign organisms, the issues of sensitivity and specificity will take some considerable work to gain regulatory acceptance.

And in most clinical situations, physicians want results in hours or a couple of days, not many days or a week, and no NGS system is at the point were it can come close to providing results in that sort of time frame.

As to cost, sure an MRI is expensive. It also has an established huge return on investment in terms of patient outcomes, and there are no or few cheaper alternatives in many cases. Any novel NGS based diagnostic tool will need to independently convince adopters that its specific cost is acceptable in terms of return on patient health and outcome. Arguing that just because something else is also expensive does not gain you any traction when trying to break barriers to new technology. I've been around several startups working on genetic based diagnostic tools, and cost to benefit is a huge issue for them if they are to gain traction with their products. With any novel technology, actual real world benefit is notoriously difficult to demonstrate, and so cost tends to dominate the discussion of adoptance.

Remember the encephalitis patient from this past June's news? Yes, his disease causing organism was successfully identified by NGS in only about two days, but the UCSF's work in that case still represents a very rare exception, and not at all the norm for medical diagnosis. It is also not clear to most how widely applicable that specific approach would prove to be when thinking of other forms of infections. So very much in a clinical experimental era yet. The physicians in that case were also able to treat based on the NGS diagnosis mainly because they were out of options if they hoped to save the boys life. Doctors do have considerable discretion in such situations, but not necessarily so in others.

UCSF's clinical microbiology laboratory is likely though to be the first to actually get formal approval in the USA to use their system routinely in certain infectious disease testing - http://genome.cshlp.org/content/earl....full.pdf+html and http://www.nejm.org/doi/full/10.1056/NEJMoa1401268

But note there is no specifically tailored NGS "product" involved.

[dukefan]

Thanks, mbblack, for your insights. I just noticed that we are less than a mile away, hehe.

By product, I mean a commercial product that you would have at clinical microbiology lab and is routinely used to identify pathogen and offer susceptibility results.

I've seen some health economics studies that show significant saving by identifying the pathogen and susceptibility early and incorporating that information into clinical treatment. While NGS can take up to a few days, culture could take a couple weeks.

I wonder how do you envision a first successful NGS product to be? I was asked to look at specific disease area where NGS can offer value and meet umet needs. I think NGS is great for MDR-TB patients to figure out all the resistance profiles and identify the most effective drugs. However, since TB mostly happen in poorer countries they probably can't afford NGS. I am now wondering if HIV/co-infections can be a good application. Instead of doing different cultures, you would just do a shotgun sequencing and figure out genetypes of HIV and whether there are other infections.

It is really hard to find a great entry clinical application for NGS.

So what do you, Dukefan, actually mean by an NGS "product" specifically for infectious diseases? Illumina and others do indeed market their existing products for infectious disease monitoring, but its just standard whole genome sequencing not something specific to infectious disease agents. WGS is WGS - nothing special about using it in the context of infectious disease research.

If you actually mean patient diagnostic testing, then yes, I would say the actual hurdle is the regulatory process and requirements. Those are the very people one needs to convince, not so much of usefulness, but of direct clinical efficacy. So, for example if using WGS from patient samples to diagnose infection by a foreign organisms, the issues of sensitivity and specificity will take some considerable work to gain regulatory acceptance.

And in most clinical situations, physicians want results in hours or a couple of days, not many days or a week, and no NGS system is at the point were it can come close to providing results in that sort of time frame.

As to cost, sure an MRI is expensive. It also has an established huge return on investment in terms of patient outcomes, and there are no or few cheaper alternatives in many cases. Any novel NGS based diagnostic tool will need to independently convince adopters that its specific cost is acceptable in terms of return on patient health and outcome. Arguing that just because something else is also expensive does not gain you any traction when trying to break barriers to new technology. I've been around several startups working on genetic based diagnostic tools, and cost to benefit is a huge issue for them if they are to gain traction with their products. With any novel technology, actual real world benefit is notoriously difficult to demonstrate, and so cost tends to dominate the discussion of adoptance.

Remember the encephalitis patient from this past June's news? Yes, his disease causing organism was successfully identified by NGS in only about two days, but the UCSF's work in that case still represents a very rare exception, and not at all the norm for medical diagnosis. It is also not clear to most how widely applicable that specific approach would prove to be when thinking of other forms of infections. So very much in a clinical experimental era yet. The physicians in that case were also able to treat based on the NGS diagnosis mainly because they were out of options if they hoped to save the boys life. Doctors do have considerable discretion in such situations, but not necessarily so in others.

UCSF's clinical microbiology laboratory is likely though to be the first to actually get formal approval in the USA to use their system routinely in certain infectious disease testing - http://genome.cshlp.org/content/earl....full.pdf+html and http://www.nejm.org/doi/full/10.1056/NEJMoa1401268

But note there is no specifically tailored NGS "product" involved.

[Joann]

A few comments. Here it would seem that the "product" is speed. In some cases, as clearly illustrated by the recent encephalitis case discussed (above and elsewhere on SEQanswers threads), any delays would have resulted in the death of the patient. Also, at present diagnosis is geared toward culturing methods only as there is decades of experience behind this approach and apparently little alternative. However, successful culturing depends on so many variables and an array of laboratory materials while sequencing can potentially simplify procedures and lower costs. How about pathogens that are difficult or dangerous to culture routinely? And add those conditions to an already difficult medical case. The ability to discern the presence of a known pathogen (to the extent these sequences are available in public sequence databases) really clarifies otherwise complex (and expensive) medical conditions and this approach also mitigates against human error in dangerous laboratory procedures dependent on culturing. So safety would be another "product".