Changes in Clinical Diagnostics and Tracking Infectious Diseases

Thank you for joining us. I'm Dr. Phoebe Thorpe and it's my pleasure to welcome you to CDC, Public Health Grand Rounds for October 2016, "Changes in clinical diagnostics and tracking infectious diseases". We have a very exciting session. So let's get started. But first, a few housekeeping slides. Public Health Grand Rounds has continuing education credits available for many disciplines, physicians, nurses, pharmacists, veterinarians, health educators, and others. Please see the Public Health Grand Rounds website for more details. In addition to our website, we're also available on social media and we are live tweeting today. So please use #CDCGrandRounds for all your tweeting needs. We also have a featured video segment on YouTube called "Beyond the Data" which is posted after the session. And we are partnered with the CDC Public Health Library and Information Center to feature scientific articles relevant to this session. The full listing is available at CDC.gov/scienceclips. Here is a preview of our upcoming Public Health Grand Rounds session. Please plan to join us live or on the web. In addition to today's outstanding speakers, I'd also like to take a moment to thank the many contributions of the individuals listed here. Thank you. And now for a few words from CDCs Director Dr. Tom Frieden. 

Technology continues to advance, change and improve our lives. One example is a new generation of Culture, Independent Diagnostic Tests or CIDTs. These CIDTs make it possible for healthcare providers to determine which pathogen is causing a patient disease, often while the patient is still at the doctor's office. When a specific cause is known quickly, patients benefit from faster diagnosis and targeted treatment. CIDTs are available for many infections, including tuberculosis and chlamydia. They can confirm infections that weren't easy to test or culture before such as legionella and C. difficile. But as with other technological advances, there are unintended consequences. Although, CIDTs provide rapid diagnosis, there is no organism available for further testing, including by traditional culture. If clinical care ends with the CIDT results, we may not get answers to important questions. We can't determine whether one person's illness is part of a larger outbreak. Whether a bacteria is resistant to antibiotics and how likely it is to cause severe illness. Information that's key to understanding where infections come from and how they spread is especially important in the case of infections spread through our food. Additional laboratory testing is still needed for positive CIDTs, without it the food supply will be less safe. Every year, 48 million Americans get sick from food borne illness; 128,000 are hospitalized and 3000 die. Twenty years ago, along with our partners CDC established PulseNet and FoodNet to track the spread of food borne diseases and work with partners to trace illness back to contaminated food. Over that time many outbreaks have been stopped, but not before infecting too many people across the US and in other countries. In 2015, there were 35 multistate outbreaks of foodborne disease. And each year, the number of outbreaks increases. Only by connecting these individual infections can public health investigators trace illness back to the source and stop further spread. Without cultures and detailed genetic information, connecting affected individuals in multiple states is more difficult, takes longer and leaves contaminated food on the shelf. As you'll hear today, our scientists are using many approaches including collaborating to develop new technologies to help address these issues. But these advances take time to implement and until they're in place, public health will have to continue to rely on cultures to detect, traces and stop foodborne outbreaks, and keep our food supply safe. 

And now for our first speaker, Dr. Chris Braden.

Good afternoon. Today I will be giving you an introduction to the advances in Culture, Independent Diagnostic Tests. Culture-based diagnostic tests have been used for over 150 years, having been invented by Louie Pasteur. More recently, culture independent tests have been developed with antigen-based tests coming into wide use in the 1980s and 1990s. These tests detect antigens which are specific to a certain pathogen type. In the 2000's, molecular detection test based upon polymerase chain reaction methods were developed. These tests detected short genetic sequences specific to a single pathogen type. Since 2010, we have seen a major expansion in the production of multiplex PCR panels which detect one or multiple pathogens simultaneously. These panels are often designed around disease syndromes such as gastroenteritis, respiratory diseases, meningitis, or bloodstream infections. Though we will be focusing mainly on diagnostics for enteric bacterial pathogens today, culture independent tests based upon antigen or nucleic acid detection have been applied to many different types of pathogens and are often the only practical way to detect viral pathogens. For enteric pathogens the number and type of culture independent diagnostic tests have increased rapidly. In the last 5 years there has been an increase from 5 antigen-based tests for Campylobacter and Shiga-toxin-producing E. coli or STEC to 8 antigen-based tests, several laboratory developed PCR test used in large commercial laboratories, and to date 5 multiplex PCR panels have been FDA cleared. Still others see CIDTs including multiplex PCR panels have been FDA cleared for respiratory, meningitis and bloodstream pathogens. FoodNet, a network covering about 15% of the US population monitors CIDT use as part of its surveillance activities. For the bacterial infections tracked by FoodNet, on average 7% were diagnosed by CIDT in 2012 through 2014. As depicted by the blue bars in this graph. This more than doubled to 16% in 2015. Uptake of CIDTs in clinical laboratories varied by pathogen and the use of multiplex PCR panels is growing. The use of CIDTs may impact surveillance trends. Compared with the average incidents in 2012 to 2014. In 2015, the incidence of Cryptosporidium and non-0157 sugar toxin producing E. coli almost doubled. This increase might, in part be caused of increased use of CIDTs. As you can see from this graphic, the workflow and CIDTs including multiplex PCR panels is relatively simple and fast. The clinical specimen is prepared for testing in a sample tube containing reagents, which is inserted into the automated diagnostic instrument. Detection of specific PCR products yields a positive result for one or more pathogens included on the panel. Automated analysis can produce a result in as few as 1 to 2 hours. Reflex cultures should be done to provide an isolate of the pathogen for additional testing such as conventional identification, antimicrobial susceptibility, and pathogen characterization. More on why this is important in a moment. However, conducting reflex culture of CIDT positive specimens may not always be clinically necessary and entails additional costs. If reflex culture is done for public health reasons, it is not clear whether the clinical or public health laboratories should conduct the culture testing and bear these costs. Also, in some instances, it may not be possible to conduct reflex cultures. For example, if the original specimens obtained with a swab, the entire specimen may be used and inactivated when transferred to the sample tube with reagents. Test procedures should include the option for reflex culture from the original specimen. The benefits of culture independent diagnostic tests are impressive, especially for clinical care. They can provide much faster results for targeted early treatment of infections, including whether or not antibiotics are needed. The panel test can detect or rule out multiple pathogens in a single test and are likely more sensitive than culture. These tests may provide faster information for local public health action. However, CIDTs alone do not provide cultured isolates. Isolates are needed to characterize the pathogen. In addition to the clinical utility of antimicrobial susceptibility results to tailor treatment, hospitals and public health programs need this information to track resistance trends in their facilities and jurisdictions. Determination of virulence factors can inform clinicians and public health officials of the likely severity of illness. Serotype and genotype information is used to identify specific strains, whether a strain is causing an outbreak and possible routes of transmission. Genotype information has been used extensively in programs such as PulseNet to connect cases, identify outbreaks and prevent over 270,000 illnesses each year. PulseNet is in the midst of transformation to more detailed DNA fingerprint analysis to facilitate outbreak detection and investigation using next generation whole genome sequencing technology. But this technology also requires cultured isolates. Each year, 48 million people get sick, 128,000 are hospitalized and 3000 die from foodborne disease. Public health programs need to enhance efforts to make our foods safer and we need to conduct DNA fingerprint surveillance of foodborne pathogens to do so. Other possible drawbacks of culture independent diagnostic tests are related to the difficulty interpreting test results in some situations. Because DNA from dead microbes can produce a positive test result, clinicians may not know a patient is still contagious and public health officials won't know if it is safe for a patient to return to work or to daycare. Also, a single test may detect multiple pathogens, some of which may not be causing illness. A study of one brand of a multiplex PCR GI panel found that over 30% of positive tests detected more than 1 enteric pathogen. This may lead to diagnostic confusion and unnecessary treatment, and to confusion for public health surveillance systems. CDC has strategies to meet the surveillance challenges posed by culture independent diagnostic testing. First, we need to support the use of reflex cultures for positive CIDT samples by encouraging states to update referral and reporting guidelines and loss. We also need to decide who performs the reflex culture and who pays for it. We can develop expedited isolate recovery methods to make culture simpler and cheaper, and we need to work with the diagnostic device industry to make sure specimens can be cultured using our procedures. Second, in the short term, we need to implement a whole genome sequencing technologies for surveillance. Though this will also require continued use of isolates from reflex cultures. Developing large whole genome sequence databases will allow the type of research and development required to achieve the third tier of our strategy for the long term, the development of metagenomic diagnostic tests. Meta-genomic technology has the potential to provide both the essential clinical and public health information direct from specimens without the use of cultures. Dr. Besser, will be providing some insight on different methods for metagenomics, including amplicon sequencing, and shotgun metagenomics. As technology advances, we must pursue a path that benefits both patient care and public health surveillance. The solution includes building a broad understanding of the issue among many partners, including experts from public health, clinical medicine, diagnostic device makers, Congress, regulators, and insurance officials. In 2012, CDC convened a multi-stakeholder forum to raise awareness of the impact of CIDTs on public health programs. In 2013, the infectious disease Society of America published specific recommendations to improve diagnostics for infectious diseases in a way that also benefits public health surveillance. FoodNet partners are tracking the growing use of CIDTs in evaluating changes over time. And The Council for State and Territorial Epidemiologists is discussing how best to count infections identified by CIDT and public health surveillance in updating surveillance case definitions. And now I'd like to turn the program over to Alicia Cronquist. [ Applause ]

Good afternoon. CIDT is having a major impact on surveillance in Colorado. Our most recent data show that 15 labs are using multiplex PCR testing. To put this in perspective 40% of enteric bacterial pathogen cases reported in Colorado so far during 2016 have been tested using PCR, 89% of the salmonella Shigella and Shiga-toxin-producing E. coli are xTAG tested using PCR have also had a reflex culture tested performed either at the clinical lab, or the state public health lab. In Colorado laboratories of various sizes and from urban and rural settings are adopting multiplex PCR since 2013. And we've been busy responding to this switch. Our activity falls loosely into 3 categories, ensuring accurate case reporting, facilitating isolate recovery and adapting our day-to-day public health practice in light of the new types of cases being reported. Until recently surveillance for enteric bacterial pathogens captured only culture confirmed cases. The classification scheme we use in all 50 states, or the case definitions have been modified to incorporate CIDT positive results as probable cases and we need to adjust accordingly. We monitor the local uptake of CIDT by regularly serving clinical laboratories. It's important to know the types of tests each lab is using so we can understand the surveillance information we receive from them and ensure we're receiving that information correctly. We started this in late 2009 when a few Colorado laboratories began using antigen-based tests to detect Campylobacter. Currently, we survey laboratories within our FoodNet catchment area which is the 7 county Denver metropolitan area, twice a year. We do a similar survey once a year with laboratories in the rest of the state where resources are not as robust. While this activity is very important, it is also labor-intensive. Knowing which labs are using see CIDTs is only the first step. Next, we needed to modify our disease surveillance database to capture the new tests, and ensure data are properly exported when queried. For us, this required working with our IT department. We have discovered there are many ways in which errors in reporting culture independent results occur. And we continue to work with disease reporters to improve the quality of the CIDT case reports we receive. For example, some labs using electronic lab reporting needed to change their electronic settings so that CIDT positives flowed to us correctly. Other laboratories had trouble modifying printed reports they sent to us. They would report culture positive for salmonella, for example, when we knew they were no longer performing culture for salmonella. And finally the thorniest issue is one of incorrect test interpretation and reporting. Some multiplex panels include pathogens that are not reportable, but sound like conditions that are reportable. We continue to have cases incorrectly reported to us. Shigella, Shiga-toxin-producing E. Coli or STEC, and Plesiomonas shigelloides sound similar, but only two are reportable in Colorado, and each is a different pathogen. We've created guidance documents and held several meetings and webinars with infection preventionists, laboratories and local public health partners. We reach out individually when we identify a particular issue and we continue with frequent stakeholder outreach. We feel we cannot over communicate with our partners about these issues. Of course case reporting is only one part of surveillance, we need isolates for subtyping for cluster detection and monitoring trends. The big question for us is where the culture happens. We've taken a hybrid approach to isolate recovery in Colorado. Historically, we have had excellent relationships with our clinical labs. We received approximately 95% of the isolates needed without having any regulations for isolate submission in place. We're building on these established relationships to find a path forward. We prefer when isolation is performed at the clinical laboratory. Culture results are available faster, it decreases concerns about transit of raw specimens, and the chance that pathogens will die in transit. It also means that susceptibility testing can still be performed at the clinical level where it could impact patient care. We encourage each lab that adopts CIDT to perform reflex cultures at a minimum for salmonella, Shigella and Vibrio. We review our isolate submission protocols with them to ensure that any isolates that are generated are sent to the public health lab. When reflex cultures are not done, clinical laboratories are asked to forward clinical material which is generally stool, that tests positive to the state public health laboratory, where we can then attempt to isolate the organism. A lot has gone into creating a workable plan. First we had to prioritize pathogens. We chose STEC, salmonella, Shigella, and Vibrio. These organisms have the highest likelihood of public health action based on confirmation or subtyping results. Next, we sought and continue to seek additional funding to perform culture which adds considerably to our costs. To ensure we would receive specimens we ask the Board of Health to modify our disease reporting regulations to require submission of isolates or clinical material for select pathogens. This is the first time specimen submission has been required rather than voluntary in Colorado. We've also worked to facilitate rapid delivery of viable specimens to the state public health lab. We reviewed the courier service we provide to ensure couriers are regularly going to all labs that need it, and we provide transport media and written guidance about how to ship clinical material to us. This is an area that's evolving as APHO is performing studies to determine the best methods for isolate recovery. However, none of this is perfect. Despite all this effort are specimen submission rate has dropped to 90% and we continue to see a range of issues. For example, stereotyping during an outbreak investigation was substantially delayed when one rural lab mailed clinical material instead of sending it to us via the courier. The final area is the one where we still need to do the most work. With CIDT, the data we're acting on has changed and we need to figure out how to use the information we receive. We see an increase in case reports with less certainty about each one. Reporting is faster with respect to the time between specimen collection and public health notification. However, we now have a longer period between the time public health first learns of a case and we have subtyping information to pair it with. First off, we need to implement the new CSC case definitions to classify and transmit surveillance data correctly. To do this, we routinely collect more detailed test data. In Colorado, we attempt to collect negative results when they're pertinent. For example, if a PCR is positive and culture was negative, we try to capture that in our surveillance data. We've worked to train internal and local public health staff to appropriately assign case status based on the new criteria. This includes creating new cheat sheets and guidance documents for information that's ever more complex. We're working to establish and evaluate guidance for which cases should be investigated based on their case status. While it would be great to interview all suspected and probable cases, we need to consider local resources and prioritize across the spectrum of diseases they are asked to investigate. For example, a suspected, highly infectious respiratory case would be a higher priority than a probable PCR positive Campylobacter case. The timing of case investigation is also critical. For the case categories we decide to investigate we launch an investigation based on PCR results and don't wait for culture. Other jurisdictions might make different choices about that. When should we initiate worker and childcare exclusion or restriction for PCR positive results? We have chosen to treat a PCR positive the same as culture in terms of excluding or restricting workers. Yet, is that's the right choice? For follow-up testing we accept a PCR or a culture result. Although culture is often the test used, since much of the follow-up testing is performed at the SPHL at no charge to the patient. And how do we handle patients who are CIDT positive for 2 or more reportable conditions. From a public health perspective, we need to decide what type of exclusion might be necessary and which questionnaire to use. Our guidance is currently to choose the disease control measures for the pathogen with the greatest risk of transmission, and to interview the patient using the pathogen specific questionnaire that is most comprehensive, or merge the 2, when applicable. For example, when a patient is supported with a PCR positive for Shigella and Campylobacter, we instruct local public health to follow Shigella exclusion policies, and interview with a questionnaire that's a blend of the Campylobacter and Shigella questionnaires. But clearly much more research is needed. We have focused our response to CIDT on accurate case reporting, isolate recovery and adapting public-health actions to the new types of cases being reported. These areas have several factors in common which are a critical assessment of resources, prioritizing based on public health goals of detecting and mitigating disease risk and frequent ongoing communication with partners using a variety of modalities including guidance documents, phone calls, and training sessions. Thanks. [ Applause ] And now I'd like to introduce Brad Spring. 

Thank you Alicia and good afternoon everybody. I'm going to be representing industry, although I work for BD. I do represent an organization called AdvaMedDx which I'll provide a little bit more detail later. But I'm going to give you the industry perspective on how we develop tests, but also how we can partner with public health and other stakeholders to solve some of the challenges we're seeing. So, as described earlier, we do have advantages and challenges of CIDTs. Multiplex PCR test can be more sensitive than culture and provide results in a few hours. They may also save costs and significantly improve healthcare. But because PCR tests require the organism to be lysed, or the cells be broken apart to extract the DNA or RNA, we don't have viable organisms available for surveillance purposes. Companies do recognize the need to partner with public health agencies and FDA to collaborate on solutions. And AdvaMedDx which is an association represented by over 50 diagnostic companies, can leverage our strong relationship with lab communities, government agencies and public health laboratories to work towards a common goal of providing rapid diagnostic solutions while meeting the public health needs. So I just want to give you an overview of how we develop tests in the industry by covering our product development process. And we typically use what's called a phase gate approach. And as you can see here we start in the concept phase, go through definition, development, qualification of the validation stages, and then we'll launch the product. But the first step in concept and then through definition is to conduct what we call voice the customer activities. And this is where we build upon what are the requirements for the new test. And this is really where we come up with the ideas and some of the challenges our customers, or future customers may have and how to address them. So as we do these voice the customer activities, we gather requirements. And the requirements may not be specific. A customer may say I want a test that's easy to use. May provide fast results. Or it's small in size. Now these aren't measurable. So we translate those requirements into specifications. So for example, a fast result may be translated into a result within 2 hours including specimen processing. And we go back to the customer to verify that essentially meets their requirements. But certain needs may only be met by existing technologies that could create what we call conflicting requirements. So for example, to achieve a fast result, PCR may be the technology of choice. And in the cases we're discussing here today, you're not able to get a viable organism when you use PCR. But just because we have a conflicting requirement, meaning using PCR doesn't allow us to culture the organism, at least from the PCR specimen, we don't abandon these requirements and we try to find workarounds. So while requirements may conflict, that doesn't mean we just give up as an industry. We try and find other ways to address them. So I think, you know going forward, and even looking back on previous product development activities, we need to encourage engaging public health laboratories, CDC, state labs, and others in these voice of customer activities. So as another example, we may go to a clinical lab, gather the data, but rarely are we getting data that says we need to be able to culture the organism for submission of public health. And then we go to public health, we'll hear, well we've got to make sure that the organism's cultured. So, I'm going to cover too, a couple of potential opportunities to solve this problem and I think the other speakers will be addressing these as well. So one example is our instructions for use need to remind laboratories that they have to follow state and federal requirements for collection and preservation of organisms needed for epidemiological or the public health purposes. So in those instructions for use, and a customer needs to follow those instructions for use, we will hopefully achieve a little bit better compliance to state requirements. There are opportunities in product development for new technologies. As you'll hear more about later. There are some on the horizon that will help reduce compatibility issues between CIDTs and public health. Metagenomics holds the most promise, today, as it may be possible to both detect the organism in a metagenomic specimen and differentiate strains for epidemiological purposes. However, there is much work still required to develop this technology and you'll hear more about this later. One thing to consider, though, is when an industry actually may decide not to create a test for manufactured test for clinical lab, a clinical lab may create that test on their own. So there are a number of lab developed tests out there today that are basically put in place to meet an unmet need. So, if there is an enteric pathogen that isn't currently detected using one of these multiplex PCR tests, a lab may create that test on their own. So as we look at culturing organisms, it's not just the industry instructions for example, they might have to change, we have to do more to educate the labs on what their requirements are, especially for LDTs or lab developed tests. So, just to kind of cover in more detail the instructions for use opportunity here. Some manufacturers, not all, do put specific precautions in their labeling. And under the clinical laboratory improvement amendments, which are regulations that govern laboratories, CLIA does require that labs follow manufacturer's instructions. So if our instructions state, for example, the APHL recommended language around laboratories must follow state and/or local rules pertaining to reportable pathogens, the lab must do that. Otherwise they during inspection may get observation regarding that. So going forward, you know we as an industry do agree we need to put such labeling and standardize it, these statements in our labelling going forward and that will help. It doesn't do everything, but it will help with the public health labs and ensuring that these pathogens are also cultured. So we believe this type of information is one step in the right direction. We do believe there are continuing opportunities on these collaborations. So we will, as an industry, work with CDC, other public health laboratory agencies, as well as FDA to discuss efforts to explore additional measures to aid in the surveillance. So, you know, we've talked about the labeling. There may be other opportunities that we can, as we come together collectively, figure out how best to meet this challenge. We have conducted and will continue to conduct educational outreach meeting with key stakeholders. So like I said, the public health labs, other microbiology groups such as maybe the IDSA, ASM and through other industry meetings. And as a manufacturer, you know we have a salesforce, a pretty large salesforce out there. And we currently use those salesforces obviously to, not just sell our products but to also educate customers in a variety of issues. One example being the get smart about antibiotics week coming up here in November. We'll distribute material to our customers to help educate them on antimicrobial resistance. So this is another opportunity to at least increase awareness and educate customers. And then finally, one thing we are looking at is just to provide better informational resources and we're talking, collectively talking to FDA about posting a list of approved or cleared molecular diagnostics on the FDA website. It will serve essentially as a one-stop shop to finding out what CIDTs are out there and what are the procedures used to conduct those CIDTs. So, and thank you very much. And now I'd like to introduce Dr. John Besser. [ Applause ] 

Thank you Brad and good afternoon. I'm going to be speaking today on efforts to develop tests for characterization of our public health pathogens that can be done directly from specimens. I'm going to be focusing on the bacterial enteric pathogens, our PulseNet pathogens. And I will be attempting to squeeze a graduate level in molecular biology into 8 minutes. So let's see how that goes. But first why are we developing tests. Well, most importantly we want to ensure continue compatibility between our public health tests, and the commercial systems even if the specimens are biologically inactivated. And the second compelling reason is that these tests have the potential of being quick, which is of course one of the same reasons that makes CIDTs desirable for clinical diagnosis. The process of detecting outbreaks through pathogen based surveillance, can be a lengthy process, as illustrated here. Cases must be identified and linked to other cases, food vehicles identified, and depending on how a direct specimen tests are implemented they can have the potential for significantly reducing the amount of time to an actionable result, which in turn means that more outbreaks would be solved more quickly. More illnesses prevented. Developing direct tests for public health is not a trivial task. Stool is an immensely complex environment, containing a wide variety of different DNA and RNA species. There could be significant human DNA from white blood cells, epithelial cells. All the food that you eat, the plants, the animals they all have DNA. There's bacteria, parasites, viruses, fungi. Every one of us has an average of 1000 species of microbes in our gut at any one time. And every gram of stool can have up to 100,000 organisms per gram. And of particular concern to us is that many of our PulseNet pathogens are virtually identical to some of the commensal flora. Every line in this figure represents the genome of a bacterium that might be in our gut. The commercial developers need to find a marker which is conserved within the species of interest. And specific to it. Meaning, that that marker doesn't appear in any of the commensal flora. For public health tests, we must also specifically detect the pathogen, but also detect other markers for strain type, resistance virulence, and also detect variability within those markers so that we can detect differences between strains, illustrated here in the colored bars. And these should not be confused with similar sort of regions in our pathogens. Of course, when we're actually doing these tests, we don't have any of that information. In fact, what we're actually dealing with is many millions of pieces, little pieces of DNA or short reads as in some of our assays. And there's two major problems that we confront. One is signal-to-noise, finding our pathogen which may be very rare when compared to the background flora. And the second is called phasing. Getting all the pieces together that we find of the pathogen, specifically connected to the marker, and not including any pieces that may be from some of the commensal flora. There's three general strategies that we're pursuing to accomplish this task. The first is called amplicon sequencing, a method that's been around, especially for the viral pathogens for quite a while. We believe that this can be accomplished using technology that is readily available today. And we're currently pursuing two approaches. The first, we're looking at for STEC, we've identified a heterogeneous region in Shiga-toxin-converting phage, which is normally integrated into the circular chromosome. And here is a map of the phage genome. Within the phage is a region containing the regulatory genes, and also the Shiga-toxin gene, which defines the pathotype, which helps us solve the phasing problem. Our scientists have developed a multiplex PCR assay using conserved primer binding sites which yield overlapping amplicons covering an 11 to 17 KB region. These amplicons can then be assembled into a single continuous strand of DNA or consensus sequence, also known as the contig. The contig, then can be compared to sequences from other patients to find genetic differences, represented here in red, which in turn can be used to assign strain types. And the strain type information is what we use to detect clusters and are used as part of epidemiological case definitions. For salmonella, we're taking a slightly different approach, since it's less related to commensal flora of an STEC. We've developed a pipeline, or a series of processes that are listed in the bullets here, to identify specific and informative targets to approximate a whole genome multilocus sequence typing or wgMLST. The second major strategy that we're pursuing is shotgun metagenomics, as was mentioned earlier. What is metagenomics, it's the sequencing of all the nucleic acids in a sample. This is widely used already to characterize the whole microbial community, or microbiome. In our case, the stool environment. Researchers here at CDC, and elsewhere, have shown that it is possible to both detect pathogens and differentiate strains for epidemiological studies directly in stool using shotgun metagenomics. However, the method is still fairly insensitive, expensive, slow. It produces a lot of data that has to be handled. But the potential power of metagenomics really can't be overestimated. It likely will make it possible to find solutions to a whole array of public health problems that are currently difficult to address. Our group, and other groups, here at CDC are exploring multiple approaches at various points in the specimen to sequence process, to solving these signal-to-noise and phasing problems, some of which are listed here. These approaches, if successful should also reduce the amount of sequencing and data for public health surveillance. We've also begun to explore with external collaborators, droplet based single cell sorting or barcoding, which should allow us to obtain the same type of information we're currently obtaining through whole genome sequencing of isolates. Individual cells are encapsulated in the lipid droplets which are subject to PCR and run through a high-speed cell sorter. In this animation, there's a laser below the red arrow, which elicits a signal from PCR positive cells. Positive cells, which are shown here as solid circles, are then routed down a different path from the other bacteria, shown as empty circles. At least in theory, this technology could result in a tube of cells from a single pathogen, which could then be subject to whole genome sequencing. In summary, direct from specimen tests could increase compatibility between CIDTs and our public health activities. And we believe that the current technological limitations will likely be overcome with research that's occurring with multiple partners. And in the end, we believe that all of these advancements will make PulseNet and our food supply safe, more efficient and more effective. In summary for our entire session today, CIDT technology has been used for improving patient care and as Dr. Braden has showed us, and public health is in the process of adapting to these changes in a variety of ways. Advances in technology in the public and private realms will make our lives safer in general, but we must pursue a path that benefits both patient care and the public's health. Now I'd like to turn the podium over to Dr. Braden again, thank you. [ Applause ]

Well I'll stay here at the panelist desk and as we go into the question and answer session, so first of all I would like to know if we have any questions from the external audience.

We do, thank you. Reports are being charted as final results at some facilities, and I suspect they should be considered screening tests. On average do these panels have acceptable sensitivities and specificities, precision and accuracy, etcetera?

I'm going to turn to Dr. Besser and Brad for comments.

There are some studies in the literature that address the sensitivity and specificity of these tests. Frankly, we don't know all the answers. Some kits are more specific and sensitive than others. And this actually has created some problems in creating the probable and confirmed case definitions with CIDTs. But I think as time goes on we will learn more and it's incumbent upon us to encourage studies of the performance characteristics of these tests so that this question can be answered.

Yeah, just a comment on industry. So our tests do go through the FDA review process and the FDA looks at what's our intended use. Is it for screening, diagnosis, and the sensitivity and specificities related to that intended use. So FDA would not approve or clear a test that didn't meet a certain threshold. That said, I think the FDA and industry, as well as labs will look at is it better than what we have today. So I think when you look at that, it's hard to say is it good enough, but at least it may be better than what we have now, so. Can't quite answer on that on the screening tests.

However, to add from a disease reporting perspective, it it's a reportable condition that the person asking the question is asking about, then yes, we would still want that reported to us, along with what type of test was performed. We need to have all of that reported to us so we can understand what's going on with disease surveillance and with these tests.

I would also add that as the tests, culture and diagnostic tests have advanced over time, I think the sensitivity and specificity has improved. And these new multi-analyte PCR panels have been shown to be quite sensitive. Another question? How about from the audience? Yes, Steve.

I'm Stephen Rose Associate Director for Laboratory Science and Safety. My question is really directed toward Alicia and falls on the previous question about sensitivity and specificity. And I think if I heard you correctly you said that if the CIDT only, then it's a probable case. Do you code the cases that you alluded to that are CIDT positive, culture negative any differently from one that is CIDT only?

We try to collect that level of information and we strongly encourage laboratories to report pertinent negatives to us. So if it's antigen positive and culture negative, we surely would like to have that culture negative piece in our dataset. From a coding standpoint to meet case classification, those would both be coded as a probable case.

Peter Gunersmidt from Enteric Diseases Laboratory branch here at CDC. This is more a question for Dr. Spring I guess. The metagenomics that we're working on trying to develop some methods. Those methods, they really belong in the clinical laboratories, don't they? So my question to you is where are the, what are the diagnostic industry doing with this technology and how far are you before we can implement metagenomics in the clinical laboratories.

Yeah, I'm probably not going to give you the answer you're expecting, which is I don't really know. But I will say this, that there are companies out there, and most of the time they're startups, right. They'll get venture capital funding to investigate this. And I know there is one company at least that I'm aware of out there that does provide testing services using metagenomics. I don't believe they have a commercial kit for laboratories. I don't know how far away we are. Again I haven't seen any come through FDA. I think FDA has to set up standards for these. You know how do you validate them analytically and clinically. So I think they're probably going to be shown more in a laboratory maybe as an LDT before they show up as a commercially available assay.

I believe that metagenomics will be used regularly, sooner rather than later for specimens from normally sterile sites like CSF and blood. And I think that in the short term that probably will occur in laboratories. But I would imagine in the very long-term it's probably most appropriate for this to happen in the clinical setting. And then the state and local public health agencies retrieve information that they need, federal agencies retrieve what they need. But it seems closer to the patient the more efficient the system gets. But I think the technology is a long way from operating at that level.

Do we have another question external?

We do from Twitter. Are there any national summaries on CIDT surveillance and impact on patient care?

So there are some summaries of CIDT use that are provided by the FoodNet publications. Every April FoodNet, as a surveillance system provides interim reports of the previous year's surveillance data and within that data there is reports of CIDT use. However, we don't gather the information about how it's affecting clinical care, which I think is the other part of that question. Putting on my clinician's hat, you know if I can get the information about the pathogen that may be causing an illness faster, same-day or even within hours, that makes a lot of difference. And so I just, you know in that antidotal way it really does improve what we can do with patient care. Any other comments?

There's a very good publication by IDSA [inaudible] on CIDTs and clinical care that one can find for free on the web, that's worth reading. Another question from the external audience.

We get laboratory reports indicating more than one positive on the same specimen. Do you know if the labs are then obligated to perform cultures, serology tests, etcetera, depending on the microorganism?

Brad, do you know that? Or Alicia?

I'm not sure I understand the question. It seems like the laboratories, if the condition is reportable, they need to report, I would guess all of the conditions is that correct Alicia?

Yeah. Yep. So from a public health standpoint, they would need to report anything that was positive. And depending on the state where they're located, they might be required to perform reflexive culture. Or, they might be required to send some of the clinical material to the state public health lab, depending on the state and the pathogen. But I think the question is more from a clinical perspective and asking about determining which pathogen is making the patient ill. And I think that's a very difficult question that I believe clinicians are trying to sort out as well.

As we are hearing from the community that the incidents of multiple positive signals from CIDTs is fairly hard. And I think there are many aspects of CIDTs that we're going to need to resolve over the years in interpretation, that being one of them, and it could be the polymicrobial infections are much more common than we knew, or perhaps some of these findings are spurious and we just don't know, and time will tell.

I think, there's not some kind of regulatory responsibility for the laboratories to do more testing to see which pathogen may be causing illness, but certainly the physician may follow up with additional tests requested in order to clarify.

Any other questions from the audience?

David Bell, Division of Viral Diseases. Could you say anything more about the role of quantifying some of these results and assisting with the diagnosis? You know in PCR we have CT values and I don't know as much about the other, but in some of the issues you raised, it might be relevant if the pathogen was present in a higher quantity as opposed to a lower quantity. You know some of these dual findings and so on.

So just one comment, that I mean as you look at the microbium and normal flora, you know sometimes you have to distinguish between what is the causative agent and what is normal flora. So quantification, or quantitation is required for some of these assays, because you may be actually detecting a healthy level of some bacteria. So it depends on the test, and depends on what the treatment options may be. But I think there is value in quantitation.

If you're referring to clinical diagnosis, is that? Or the? Is that what you're referring to?

Well, well both. Also surveillance because it's not clinically relevant. It may or may not be public health, but it's both.

Sure, I believe most of the, or all of the CIDTs operate with thresholds built into the system already. So the positive result you get is a quantitative result. In the area of metagenomics, and sequencing, and shotgun metagenomics, quantification is, quantitation is an approach that we're exploring extensively to help with phasing, have been re-binning clinical quantitation of pathogens to align our various bits together. So for characterization of our pathogens, what we call in situ pathogen characterizations, quantitation is an important part of our overall strategy.

So to follow up on David Bell's question on quantitation and the issue of exclusion either from workplace or daycare, and of course, my experience is with norovirus where it's quite clear that the virus is detectable much, much longer than somebody is probably a real risk for transmission. And so is there any thought of using quantitation in making those determinations about when somebody is or is not excluded from the workplace or from daycare?

It would be great. I don't know if it's at all in the horizon, but even before that I think that some follow up studies are desperately needed to track how long people shed. I think that we exclude people fairly routinely from occupations if they have, say E. coli 0157 or Shigella, we will exclude them from their jobs, which has a tremendous impact on them and their lives. And so form a public health standpoint, we are very concerned about doing that based on a PCR positive and are very eager to participate in studies, perhaps where stool is collected, and both cultured and then put through a multiplex PCR so we understand that better. I don't believe those studies have been done yet.

Do we have another question?

FoodNet depends on public health lab surveillance to describe changes in illness incident rates overtime and the incident rates have targets that come from healthy people 2020. How has CDC incorporated the impact of CIDT in tracking illness incidents overtime and marking our progress in meeting foodborne illness national health objectives.

I will start and Alicia may comment also. So there were a couple of times, and especially Alicia alluded to the fact that CSTE is revising case definitions. When we revise case definitions it is going to affect the surveillance data trends. What is going to have to occur is to the extent possible to take the data that we have about the increase in specificity and sensitivity of some of these tests, and apply that to the models that we use to then generate our surveillance trends to be able to translate what we've seen before with what we're seeing now. That's not an easy task, but it is something that a number of groups are trying to do. So that we can then determine if we are actually still on track with meeting our goals, or we're not. Alicia?

As Dr. Braden alluded, we're working on a variety of data collection projects and thinking about different ways to model the data to assess the impact of these changes on our trends so that we can accurately say whether disease is going up or down to meet these national objectives. One key piece of information that we didn't have time to discuss during the formal presentations, is another way in which we hope to partner with clinical microbiology labs, which is the collection of information from them about how much testing they're performing and who they're performing testing on. And this is absolutely critical information, so that we can understand the data that's flowing to us. One area with CIDT it's possible that more testing is being performed, or testing is being performed on different groups of people than it used to be performed in. And we need to understand that to understand the surveillance data that's coming to us. So in addition to the partnerships that we've talked about with clinical microbiology labs and submitting isolates or clinical material to state public health laboratories, another key area for participation in partnership between public health and clinical micro labs in the near future is going to be working with us to figure out how to obtain this, what we're calling test volume data.

Another question from the audience.

Hi. Lee Katz Foodborne Diseases Laboratory branch. I was wondering from the panel if there are any products being used for source attribution. So in Dr. Besser's presentation, he noted that some food might also be in a sample, in a stool sample, if there are products that might be used to detect them that are being in research and development. And if those food sources are detected, I'm wondering if at Colorado and other state health labs, if it might be used to inform epidemiology.

That's a very good question. No there is no such test out there that I'm aware of, but it's an excellent suggestion. If you were able, I'm sure that if Alicia were able to tell her cases what they ate and ask them where they got it from, that would be very helpful.

I think there are some technologies that may adapt a clinic test, right, and maybe a sample preparation that has to be modified and then allow it to work in a CIDT. I know there are some next generation sequencing technology, well the technology itself of NGS is used in say in some food screening. I know they even use it in like looking for male and female eggs in selecting, you know which ones go forward and which ones don't. I think they're also using those in detecting potential pathogens.

One area where this technology has been used for a while has been in food fraud. So for instance, fish, the fist that you buy at the supermarket may or may not be what it says it is. And the FDA has been using technology of the sort to confirm the identification of seafood.

Thank you so much for joining us and I'd like to take a moment to thank our speakers. Thank you very much. [ Applause ] And please join us next month for Public Health Grand Rounds.