Medical Device Link.

IN PERSON: EXPANDED ONLINE VERSION

New developments in detection technologies, such as detection of pathogens that cause infectious disease, show promise but present many challenges. Among these challenges are working with very-small-scale microfluidics, perfecting sample preparation, and achieving high sensitivity and specificity in a small footprint.

To learn more about these and other detection-technology issues, IVD Technology editor Richard Park spoke with Christopher Cooney, PhD, director of engineering at Akonni Biosystems (Frederick, MD). In this interview, Cooney talks about advances in lateral-flow technology, making point-of-care devices smaller and more portable, the so-called “microplumbing problems” surrounding the point-of-care genetic-based biochip, and more.

IVD Technology: What have been the most significant advances in the area of detection technologies during the past few years?

Christopher Cooney: In general, we’ve seen more market penetration from genetic-based tests such as polymerase chain reaction (PCR) and microarrays, with FDA approving the first microarray test about two years ago. Sequencing has also made impressive strides. And although sequencing is not directly applicable to diagnostics yet, it does allow diagnostic tests to be developed more rapidly. So, rapid sequencing translates into rapid identification of genetic primers and probes for many diagnostic tests, such as microarrays. For example, the sequencing of multidrug-resistant tuberculosis allows companies like ours to develop a diagnostic for the alleles, for the drug-resistant tuberculosis strains.

Additionally, we’ve seen broader use and lower cost of lateral-flow strips, which have translated to more point-of-care testing. And then there’s been more automated sample prep for genetic-based assays, which has been another significant advance. And we’ve also seen increased multiplexing.

Molecular diagnostics is an interesting, growing part of the IVD market. How has the development of molecular diagnostics in the IVD market influenced detection technologies and the development of such technologies, and how will molecular diagnostics continue to influence the direction of detection technologies moving forward?

There has been an improvement in sensitivity and specificity compared with traditional immunoassays. However, molecular diagnostics bring the extra burden of sample prep, which is typically the more onerous part of genetic-based tests. Often, many companies are based or founded on the detection technology with not as much focus on the sample-prep technology.

Many of the more established companies, however, recognize that sample preparation is critical to improving sensitivity and specificity. Sample prep and fluidics are also the pieces of the puzzle that make the difference as far as having an integrated system for a point-of-care use goes. And in some ways that’s really why lateral-flow tests are still so ubiquitous and continue to grow in popularity.

Trends and Challenges

What are the latest trends in developing detection technologies for IVDs?

One of the latest trends is genetic detection of the responsible pathogens of infectious disease. Many companies for some time have been trying to develop a low-cost, stand-alone, point-of-care device to genetically identify infectious disease. But a pitfall has been the plumbing. As the fluidics scale down, the complexity seems to increase nonlinearly, which opens up the opportunity for many companies to explore novel microfluidic schemes. For example, we’re starting to see lateral flow applied to genomic tests, and we’re seeing more droplet microfluidics, such as emulsion schemes used for sequencing. Our group is pursuing a hybrid microfluidic lateral-flow approach.

In general, the challenge has been: How do you take a 1-milliliter sample, which, for many applications, is the typical sample size that has clinical relevance, and reduce that down to a biochip the size of your thumbnail? For example, if a physician takes a spinal tap to test for encephalitis, the sensitivity requirement may be as low as 1 to 100 PFUs per milliliter. Even if the test has a high sensitivity, you still are going to need a significant volume to be able to resolve that 1 to 100 PFUs per mil.

In addition to sample volumes there are other problems, such as biomolecular absorption of enzymes because of the high surface-to-volume ratios of the chips. There are bubbles clogging small channels. Then there’s manufacturing tight tolerances down at the small scale.

All of these challenges have in some way plagued the point-of-care genetic-based biochip. Many groups and companies are being founded now to address, or are at least focusing a lot of effort on, what I’m calling these “microplumbing problems,” rather than taking the more traditional route of focusing solely on the detection part of the puzzle. We’re also seeing in general more-sophisticated automation instruments and higher throughput at standardized laboratories. Again, other trends are multiplexing and improved ease of use.

What then are the primary challenges that IVD manufacturers encounter when designing and developing their detection technologies for their products?

Microfluidics is a challenge. Sample preparation is another challenge. Sensitivity and the specificity for the tests, going from sample to answer is a challenge. A small footprint that’s easy to use, that’s also a challenge. Manufacturing in general—with respect to achieving the required tolerances at small scales—is also a real significant challenge.

You’ve mentioned multiplexing a couple of times now. It’s been around for a while, but it is still a relatively new technology that has generated a lot of interest in the IVD industry and among IVD manufacturers. What sort of specific challenges does multiplexing present to detection technologies?

Aside from the technical challenges, such as multiplexing PCR reactions, there have also been issues on the reimbursement and regulatory sides. Microarrays, for example, have been around for a while and have been used by a number of groups. Although we’ve seen FDA approval of a microarray test for breast cancer, microarrays have not really penetrated the IVD market as one would have thought a number of years ago, for two critical reasons, in my opinion: One is the health-insurance reimbursement scheme, and the other is the regulatory side.

Here is an example of the former: A physician calls for a test, which can identify multiple pathogens on the same platform, but the physician only calls for one test. Well, all of those tests are going to be run, even though the physician is only interested in one of them. And the result that is returned will only be the one because that’s only what the health insurance company is going to pay for.

So then that puts the burden back on the IVD manufacturer: Is the manufacturer willing to invest in the potential for a lot of information on a single test that may not be used, or will they choose the route of multiple single tests? Additionally, each test has to be validated on its own and in conjunction with the other probes and primers present. So that is a big and costly effort.

Now, the upside is that you could potentially have a broad spectrum of tests, without necessarily changing your manufacturing. So you could have broader deployment if you’re willing to invest in the validation studies. And maybe the health insurance wouldn’t reimburse for the full test. But in the long run, as far as manufacturing in volume, you might reduce your cost because of the sheer number of tests that you’re developing.

But I think more and more we’re starting to see that multiple tests are necessary, especially when you consider something like tuberculosis. If it’s known that it’s tuberculosis, well, is it a drug-resistant strain? And, if so, what drugs is it resistant to? So there are a number of drug-sensitivity tests that have to be run. The normal recourse would be to prescribe multiple antibiotics. The treatment is then adjusted, based on the results of these tests, which can take possibly weeks for a result. But then there’s a lot of cost associated with prolonging the treatment because of a lack of information at the beginning of the treatment. I think physicians and healthcare providers are starting to realize that knowing more information up front and treating correctly and more effectively is going to reduce cost.

With smaller sample sizes being demanded to enable certain tests to be run, what sort of specific challenges do those small samples present to developing detection technologies?

There are primarily two sets of challenges: One is the control of liquids. And the other is biomolecular absorption of enzymes and sample. Shrinking down geometries translates to increasing surface-to-volume ratio. That’s where the biomolecular-absorption issues arise

With respect to the challenge of liquid control, it’s difficult to control microliter droplets. If you’re only able to control, for example, ±0.2 microliter—which is impressive—and if you’re manipulating a nanoliter-size droplet, that’s 20% of your sample. This variability would likely have a significant impact on reproducibility. With respect to biomolecular absorption, there is a reduced efficiency as the sample is moved from location to location. If the sample is precious (i.e., has a low copy number), a significant fraction could be lost to fluidic connectors, channel walls, pipette tips, and so forth. It becomes critical to know the sensitivity that you require, and then to back-calculate from that what volumes are necessary, especially with respect to how efficient your sample-preparation process is.

Typically in sample preparation processes you’ll see efficiencies are 60 or 70%, so there is sample loss just by using traditional sample-preparation strategies. When you try to incorporate that on a small scale, efficiencies decrease even further.

Considering all of these various challenges that we’ve been discussing, how do IVD manufacturers overcome them?

We work on multiple projects in parallel. So we have near-term and long-term goals. And we allow the technology to develop along with the science. This strategy has helped us make advances we didn’t think were possible. For example, we’ve been working on our microarray flow cell and discovered almost by accident that we can design a flow cell that isolates a single chamber for performing PCR and microarray hybridization from a waste chamber by the use of a specific geometric configuration.

When thermocycling is complete, a pipette introduces wash buffer into the PCR/microarray chamber and displaces all the liquid into the waste chamber. That was one path that came to surface that we weren’t necessarily pursuing, but we believe it had a big impact. In general, I think it helps to have parallel paths to allow risk mitigation. So if one path doesn’t work out, you still have an alternative plan.

Developments in Detection

What detection technologies have been developed by your company, Akonni Biosystems?

We’ve actually been working on quite a number of activities. Our company is based on gel-drop microarrays. They are porous, hemispherical gel drops. They have a greater surface area than conventional 2-D arrays, and this allows for improved signal-to-noise ratios with fluorescent imaging. This improvement can translate into lower-cost fluorescent detection systems. We’ve also been working on a low-cost fluorescent reader—considerably lower cost than what we see on the market for microarray imaging. The technology of the gel-drop microarrays can be applied both to DNA microarrays and protein microarrays, and it’s more or less the same process. They’re printed onto a surface and then polymerized on that surface. The surface can be conventional glass, or it can be plastic without needing to functionalize the plastic. Our gel drops give high signal-to-noise ratios and are low cost because of the ability to print on plastics. It’s a printing method rather than a photolithography process.

We’ve also been working on sample prep. We have a sample-prep device that’s in the form of a pipette tip. It can process complex samples in about four minutes. So far we’ve used it for blood, sputum, nasal wash, urine, and soil samples. We’ve done that for both DNA and RNA microbes such as TB, flu, Strep, and Staph. As I mentioned earlier, we’ve developed a “flow cell” biochip that allows us to do PCR in the microarray chamber, and then wash the microarray. But the wash solution stays in the biochip, so there’s no risk of contamination.

We’ve demonstrated this more than 250 times for our TB, MRSA, and Strep assays and arrays. We have also worked on an automated version of this sample-preparation device as well. And we have been working on a thermocycler device that allows us to efficiently perform PCR inside of our flow cells, which are typically flat substrates. Coupling to those substrates with conventional thermocyclers is difficult. This device uses a flexible membrane that couples to the flow cell, so it provides efficient heat transfer.

We’ve integrated this all into one complete package, and we’ve demonstrated sample-to-answer for our microarrays. And one of the challenges with sample-to-answer from microarrays is typically the wash step. So we’ve integrated the wash step as well. That’s in a breadboard format, and that’s more long-term than some of these other things. What we’re working on near-term are the assays, the flow cells, and a microarray reader. What we’re working on longer term is the automated sample-prep device, and the point-of-care sample-to-answer integrated microarray system. And we’re opening up these projects to interested companies as a way of focusing on our near-term goals.

You’re talking not only about detection technologies but also various other kinds of related technologies. Specifically regarding detection technologies that Akonni has developed, how have they been applied, or how do you envision them being applied or implemented in the diagnostic test?

The first way is that the diagnostic test would be a multistep process. So we have that sample prep pipette tip that a trained laboratory technician would use as a means of purifying the sample. And then that purified sample will be combined with our assay for TB, for example. And then that combination of the sample and the PCR master mix will be introduced into our flow cell. The flow cell contains Akonni’s gel-drop microarrays, which is our detection platform. The flow cell is then inserted into a conventional slide-based thermocycler. The flow cell then gets washed by the laboratory technician and read on a microarray reader.

The longer-term vision is to have a point-of-care device that allows us to go sample to answer where the answer is a readout from our gel-drop microarrays. This is our integrated system that I described earlier.

Has your company ever established any sort of partnerships or strategic alliances with other IVD companies and manufacturers to develop detection technologies?

We’re currently surveying a number of different manufacturers for this purpose. We’re at the point where this is the next major development step for us.

Creativity and Simplification

As you know, there are efforts to continually automate processes and make them simpler and easier to use. Do those efforts put more pressure on developers and manufacturers to create and produce more-sophisticated and advanced technologies to ensure that the tests are as specific and sensitive as possible?

I think to some extent that’s true. As you completely automate these tests, it does certainly shift the burden from the user to the instrument to be able to handle that. On the other hand, I think that there are a lot of creative solutions coming out now with respect to microfluidics that make that not as onerous as one would have thought perhaps a number of years ago. So I believe what we’re starting to see is a simplification of the fluidic schemes and more-creative solutions with respect to the fluidics. For instance, not needing as many pumps and valves because of the application of clever microfluidic approaches.

Of course there’s always the risk of the user not introducing the sample properly into the disposable or into the system. And that is usually one area that is difficult to control and to automate. And I think that can only be solved by working closely with the users and the user community to understand and to clearly define protocols on how the sample should be introduced and to spend effort in designing that sample introduction method and process.

Is working so closely with the users in that way very expensive? And how is it done?

Yes, it is expensive. And that’s a challenge, especially for smaller companies. We have an early-adopter program and we’ve identified certain clinical facilities that have been eager and willing to work with us on our development. We get feedback from them, we iterate on it, we give them new improvements. Engaging these clinical sites early on is helpful and necessary to make that happen and to gain that input.

Keeping Costs Low

What challenges will emerge that IVD manufacturers will have to overcome and address in the future in developing detection technologies?

The challenge I see on the horizon that is hard to miss is designing low-cost tests. It seems that everything is pointing to this right now. Hospitals are trying to cut costs anywhere they can. Health insurance plans now are trending to high deductibles, making patients more vigilant with respect to the cost of tests. And of course labor is a significant part of the cost. Designing low-cost systems that do not require a lot of hand-holding will be necessary, which is particularly challenging for companies targeting point-of-care diagnostics.

How does an IVD company balance its efforts to keep tests low in cost with the healthcare community’s demands to make tests simpler and easier to use?

It’s an ongoing struggle and challenge for every company. And usually what happens is that someone sketches up what seems like a great idea on a white board. But then as you go to implement it, issues start to come up, and difficulties arise, and problems need to be solved. And the solutions often make the system more complex, which of course adds costs and makes the system more complicated and maybe not as robust as what was initially intended.

I think it’s a real skill to be able to take market research and distill it into good design requirements by having a clear understanding of what’s actually a requirement and what’s a “nice to have.” Understanding that as the system or the product or the prototype is moving through the development process, there are some things that might not be able to make it into the product, some extras or some “nice to haves” that might not make it in simply because of cost. It’s an iterative process that has to be completely looked at. I believe that—as I mentioned earlier—having parallel paths and risk mitigation for a specific project helps in that regard.

Having constant communication between the various groups in the organization to understand where the cost creep comes from and how to control that is critical. Of course, involving manufacturing early on is a big part of that, and understanding the implications of the design early on. Most manufacturing engineers want to be involved early because when a concept starts on the white board, they can see the manufacturing challenges, even though it seems like a great idea, and maybe you even get proof-of-principle results. But when you go to manufacture it, it requires a difficult and expensive manufacturing process.

Innovation Versus Tradition

There have been certain detection technologies that have been around for a long time. Considering the emergence of such innovations as molecular diagnostics and multiplexing, do you see traditional tried-and-true detection technology sticking around and continuing to be used in the future?

Yes. It certainly seems that way. Lateral-flow strips, for example, are getting more and more prevalent and are being applied to more and more different tests. So yes, I see that they’re certainly going to be around for quite a while. But they will be improved upon. And they’re now being used more quantitatively than they’ve been in the past, and there are interesting approaches to doing that using fluorescent-based beads and using readers to read the lateral-flow strips, whereas before it was just done with a visual read. We’re seeing an increased use and implementation of the tried-and-true methodologies for a broader range of tests. Of course there are also some cutting-edge new approaches that are being applied specifically in the area of flow control that are being advanced as well.

What new trends can we expect to see this year and in the future in the area of detection technologies?

We’ll see more-portable point-of-care devices. I believe this year we’ll start to see more-streamlined lateral-flow devices—systems, for example, that use a single absorbent pad that can accomplish all steps for lateral flow: separation, wicking, conjugate release, reaction membrane, and absorbent. In the future I believe we’ll see more systems that use capillary action, novel lateral flow methods for fluidic control, and hybrid devices consisting of microfluidics combined with lateral flow. I believe we’ll also start to see increased multiplexing on a single test. I realize that health insurance has limited this innovation to some extent, since the patient should not have to pay for a test that the physician did not prescribe. However, if the detection strategy were a microarray, for example, it’s conceivable that a universal array could be used for a battery of tests, perhaps a different set of primers and/or reagents are added to the same universal chip. This of course has regulatory implications since each individual test will require validation.

We’ll also see a lot of growth in pharmacogenetics testing. That’s where I think the trend is going.