Developing IVD Instruments: Complex But Rewarding

Gen-Probe's Panther system

Developing an amplified molecular instrument is challenging. It is a time-consuming and costly endeavor for manufacturers, and the end result must be a highly reliable, robust instrument. Add in the regulatory hurdles and issues related to marketing the product worldwide, and the task seems almost insurmountable.
To learn more about how such challenges are overcome and why the benefits of developing such instruments are worth the costs, IVD Technology editor Richard Park spoke with Brad Blake, vice president, research & development for Gen-Probe.

IVD Technology: What have been the biggest technological advantages in IVD instrumentation development during the past few years?
Brad Blake: Amplified molecular methods are at the top of the list. I have been in instrument diagnostics for over 25 years, and to me, the biggest advancement is the amplified methods where you amplify the genetic target and enable sample detection with very low copy levels. This is a profound improvement over standard signal amplified technologies like immunoassay and chemistry. A good analogy is finding the proverbial needle in a haystack. In non-amplified methods, you have to process every single piece of “hay,” hoping that the lone needle will pass by your sensor and be detected. And then, even if that needle does pass by your sensor, it most likely will be at such a low signal level that your sensors will not detect it above the “noise” of the machine. This can result in a situation where the instrument reports no needle is present, when in fact they were present but went undetected.
In comparison, for amplified molecular assays, a needle-amplifying reagent is added to the hay, which then duplicates the single needle into two needles. This process continues by duplicating the two needles into four needles, then 8, then 16, then 32, and on until you have millions of needles. So if zero needles were present in the original sample, you have zero needles. If there was one needle present, you end up with millions of needles.
This technology enables you to detect very low copy levels of your target cell, which can mean much earlier diagnosis and treatment for the patient and improved clinical outcomes. This kind of true amplification is important not only in diagnostics, but in blood supply testing as well, where we and our partner, Novartis, maintain more than 80% of the US market.
With the new molecular methods, the blood supply has been made much safer for people around the world. For example, in South Africa, where 20 to 30 percent of the adult population has the virus that causes AIDS, there has not been a single documented case where an individual has contracted AIDS from transfused blood in the several years since molecular blood testing began. So the biggest advancement I see is molecular testing, and it keeps on delivering more value as we move forward.

What sort of development efforts were required to make molecular diagnostics more automated and into more of an instrumentation platform?
That is a really great question. I have led the development of many diagnostics instruments-hematology, immunoassay, chemistry, liquid-base Pap, high-resolution imaging device, et cetera. Developing a molecular automated system is significantly more difficult than these other systems, and that’s why you see very few fully automated molecular instruments in the marketplace.
It took Gen-Probe more than 10 years and millions of dollars to develop the know-how to create Tigris, because molecular processing can take some 40 to 80 steps. Each step has to be done very precisely, with very tight temperature controls and with very small aspiration and dispense volumes of sample and reagents, all while having to address the difficult problem of contamination. To ensure each one of these steps is done correctly, your process controls have to be extremely robust.
 Taking all these factors into account, making an amplified molecular instrument is definitely the most difficult challenge in the diagnostics space. It is very, very difficult. It is very expensive, and it takes a great deal of know-how and system knowledge to create these instruments. Tigris has been in the market since 2004 and it has no strong competitors, in part because of the difficulties in bringing this type of instrumentation to market.

What would you say are the latest trends emerging in IVD instrumentation development?
The next step is to get the automation of molecular instruments to the same level as that of chemistry and immunoassay systems. The goal is to eliminate the need for specially trained operators to perform these unique processing steps in order to get a good result.

So what would it take to bring molecular up to the same level of automation as the assay systems?
We just released the Panther System in Europe, which we strongly believe achieves that end. Panther has random sample loading. We have host query, so the customer really just loads the sample onto Panther and it does all the work for them. Panther goes up to the host via the LIS and queries the host on what test needs to be run. It then runs that test and sends the result back to the host, so the operator has no need to interact with it. There is no need to batch or write what test needs to be run-you can simply load samples on the fly, while the system is running. You can also change reagents on the fly.
We think Panther has achieved that same degree of automation as the chemistry and immunoassay instruments. Panther was designed and built to be a major improvement in molecular lab productivity.

What are the primary obstacles that IVD manufacturers encounter when developing their IVD instruments and in trying to do business in this area?
Dealing with the tremendous complexity of these instruments. They have to process thousands of samples without a problem, work with over six different sample types, and achieve very high levels of sensitivity and specificity. The entire development process takes a large group of very talented people with diverse backgrounds and talents who must work cooperatively together in the same direction-not an easy task.
Additionally, there are always the regulatory approval challenges to work out-FDA, EU, Health Canada, et cetera. You must have a best-in-class development team to be successful in creating these instruments.

How do IVD companies go about overcoming such challenges?
First, by following a systematic development process. We use a phase-gate development process, in which development is broken up into separate stages. Then you have to hire very knowledgeable people in these various areas, looking at it from a system point of view. Gen-Probe has some of the best system engineers in the world.
Although it took us more than 10 years to develop Tigris, Panther took only three-and-a-half years. That demonstrates that companies need to invest in the “systematic knowledge” that is needed to create these instruments.
And once you start, it does get easier but it is never truly easy. Once you complete one instrument, you can do a second and third, but it’s critical to have that solid infrastructure of software engineers, system engineers, hardware engineers, assay scientists- all of whom must have the expertise needed.
You also need a systematic and controlled development process. For instance, our software follows a spiral model. This was developed by Carnegie Mellon and is viewed as superior process for developing software for complex instruments. We also use automatic requirements processing software, so when we write our requirements, they flow through all our development stages, right down to our test cases. Gen-Probe has also developed simulators that replicate all the hardware functionality before any hardware modules are designed and built, so we can fine-tune designs quickly and easily.
 Using all these automated tools makes the development processes easier, because again, the processes are so complex you must move beyond the proverbial “notes on the back of the napkinæ and “some smart people in a room.” You still need those smart people, but they have to be well-managed and well-coordinated with these automated tools. With our software verification and validation, we use automated tools to test all out external interfaces and do extensive simulation that stress our software to ensure we have a robust product.
The key is using this whole suite of the latest development tools to help create a winning product.

When developing IVD instruments, what model do IVD companies tend to follow?
First, we spend a great deal of effort understanding what our customers will want as a next product, what our customers’ key requirements are that will improve their business, so they will be delighted with our product. We then look at how we can best address those needs. This is a very complex process. We look at multiple pathways and options until we are satisfied we have found the best one that addresses our customer needs. This is a place where we do not rush. It is critical to determine up front what product you need to robustly meet your customer needs. If you design the wrong product, you lose. If midway through development, you determine the need to change your design, it is very costly, time consuming, and problematic-again, you lose. Simply, it very important to spend the time and effort up front to ensure you are designing the right product for your customers.
Determining if you should design a product from scratch or modify an existing design (a derivative product) is based on the product requirements, time to-market, cost, et ceteral. If you can meet the product requirement with a derivative product, that is great. It speeds up the time-to-market and by utilizing design that has been proven, you increase product reliability. This is good for both the company and our customers. Customers get a great product that meets their needs sooner.
Sometimes, customer requirements dictate a completely new design that will robustly meet those needs. In these cases, a completely new, from-scratch design is created, which typically increases the time to market.

When you do build a completely new instrument, how much time, money, and energy goes into developing it? When would you see a need to build a completely new IVD instrument that has never been on the market before?
You can divide it into at least two areas. The first is if the technology is old and getting near the end of its useful life, there is a need to move to a new technology to stay competitive. The second case is when you have good technology, but you really want to make a quantum leap and take it to a new market where it may not have been the most appropriate. Then you need to make a major jump in technology.
But normally, you keep evolving your technology-keeping it fresh and up to date, keep improving it in steps so you stay ahead of your competition. This approach is often better for customers, because developing a whole new technology takes a longer period of time.
To summarize, the goal is to develop a product that best meets the customer’s needs. Typically, if you keep evolving your technology so it stays ahead of competition, you will have a winning product using that technology. However, there are cases when the best solution is to use a new technology, but that is more rare.

What role has connectivity played in IVD instrumentation development?
Connectivity plays a large role. As we have discussed, labs are looking for more and more automation. Labs are desperate for more automation for three important reasons: One, it is more and more difficult to find skilled laboratorians; two, labs are under constant competitive pressures and are looking to increase productivity of their workers; three, people make a great deal more mistakes than instruments. The more manual steps that humans have to perform, the more likely a mistake will be made. This is costly, decreases productivity, and in some cases, leads to incorrect results.
Labs are looking for a host query where you just load the racks on the instrument and press a button that says “go.” That request will go through the LIS connectivity to the host and ask what test needs to be run, then the instrument will run that test and report the results back to the LIS. The operator has no need to interact with the sample, except for placing it on the system.
Another place where I see connectivity is in remote service and diagnostics. This is becoming more important as it allows your service person or call center staff to remotely dial in to the instrument to gather information to help fix problems. Sometimes the call center can fix the problem right over the phone, saving a field service engineer (FSE) visit. If an FSE is needed, the technician will know in advance which part to bring along to resolve the problem, which again saves time. In all, instrument uptime is increased.
We have remote service and diagnostics on both our Panther and Tigris systems. Connectivity is playing a big role in helping to improve the quality, reliability, and customer experience of our instruments.

How has point-of-care testing affected IVD instrumentation development?
Another great question. For years, industry analysts have predicted that point of care is going to take a large percent of the market. Years back, Abbott’s handheld blood analyzer, iSTAT, was going to be the dominate player in the market, but iSTAT has never really achieved those market predictions.
Clearly, point-of-care systems are getting better, but the mid- to large-volume systems are here to stay. Look at LabCorp or Quest, and you see the huge number of tests that they run and the value they provide to customers. Point-of-care instruments just cannot offer this level of efficiency, for many reasons. First, if you just look at it from the cost perspective, typical point-of-care tests are much more expensive than mid- to large-volume systems. Unit-dose consumables can cost you five, six, or ten times what assays on a mid- to high-volume system cost. Considering this cost model, mid- to high-volume systems have the advantage. And typically, the sensitivity and specificity of large-volume systems are better than those of the point-of-care systems.
Another consideration is whether doctor office staff are trained to perform these tests. Do they have the room necessary for all the instruments? What happens when an instrument breaks down? What about increased liability of the doctor office? If you really look at the point-of-care model in totality, there are many variables to consider. The big labs like Quest and LabCorp offer a lot of value in doing these tests.
Point-of-care will lead parts of the market, such as in emergency rooms, where immediate testing is needed. Costs and specificity issues are not so critical there. I believe point-of-care testing has its place, but I do not believe it will displace mid- to high-volume systems or have a major impact on them. The mid- and large- instruments are still going to be performing most of the tests because of the economics and quality issues.

How do IVD companies like Gen-Probe engage or liaise with academia in developing IVD instruments?
These types of relationships are critical for us, both for the development of instruments as well as assays. We have relationships with the University of Michigan and many other institutions. We monitor very closely what’s happening out there. We look at the research, and where we identify a technology or method that looks promising, we will engage with those people, work with them, help fund them, be part of the process.
We have an entire research team whose sole function is to monitor and partner with academia. We keep our finger on this pulse, so that if a technology emerges that is going to be disruptive, we don’t miss it.

Is there any particular technology that you’ve seen recently that has come out of academia that has particularly impressed you and that you could possibly foresee making a significant impact in the area of IVD instrumentation?
We try to look at instrumentation and assays holistically, as an integrated solution. For example, we have a relationship with the University of Michigan related to our PCA3 and T2 assays, which are new markers that help diagnose prostate oncology problems in men. That is one really clear area.

What role does software development play in the IVD instrumentation development? And how do those developing the software reconcile the complexity of the instrumentation with creating a clean, simple user interface?
Software is probably our largest single expenditure for an instrument. Hardware and verification costs can be tremendous, but software and software verification are our largest costs, since we recognize the importance of ensuring the software works exactly as defined in our product requirements. For example, our Panther software was developed by about five different companies across six different time zones. We had experts in databases, experts in user interfaces, and experts in embedded controllers. We take advantage of the world’s experts in those areas. Software is a big part of the finished product and its development process is very closely managed. We have a whole department that does nothing but test software to make sure it performs as we said it would.
The key to good software development is to focus on the requirements stage. You must have a good vision of what product you want to make and what type of experience you want the customer to have. Many of our software requirement writers have actually worked in the field talking to our customers and running instruments.
We also have our marketing people who focus very closely on customer needs. One of our key marketing people used to work in customer service, directly answering customer calls and solving their problems over the phone. So our marketing personnel are critical in helping us frame and structure our software.
Our process involves analyzing all the various software available in the world. We look at everything from iPad tablets to other diagnostic equipment. We look at user interfaces across the world and try to get the right combination. You touched on user friendliness. This is important for us because some of our devices are used in developing nations, while others are used by very sophisticated operators in the United States.
We look at the number of clicks required by the users and we try to keep that to a minimum. We want to keep the user experience simple but flexible. A key example is on Panther. When you open the lower drawer of the instrument, the software immediately goes to the appropriate screen. The user does not need to select it. When you close that drawer and open another to put new tips in, the hardware tells the software that a new drawer is open, and the software immediately calls up that screen-it’s very intuitive.
When we develop software, we bring in customers at the very beginning and show them what we’re developing-this results in very helpful feedback that allows us to incorporate further improvements.

What new trends can we expect to see this year and in the future of the area of IVD instrumentation development? What future challenges will emerge in IVD instrumentation development?
For starters, we have a changing regulatory environment, which is always a challenge. We also see one of our biggest challenges crop up as the world continually becomes smaller. We are placing instruments in Indian, Chinese and South African markets, so we must ensure that our systems meet the unique needs of those markets. They have to be more reliable, more easily serviced.
Additionally, in this global environment, our customer interface must be translated into various languages while still maintaining the same look and feel. As for instrumentation trends, I believe our customers want more productivity, they want more automation, they want more connectivity, they really want to see us continue to advance the technology.
Most importantly, our customers want more assays on a single box. They want consolidation-they don’t want five different boxes. In the current amplified market, if you want to run various tests, you need to run this one on this system, that one on that system, et cetera. Clearly a consolidation of platforms is warranted, so you can run various technologies all on one instrument. If you look at the throughput of instruments such as Panther, they have the ability to do that. I believe this will be an important future development pathway for molecular companies such as Gen-Probe.

Brad Blake is vice president, research & development, Gen-Probe. He has more than twenty years’ experience in instrument development. He can be reached via Alyssa Eggum at alyssa.eggum@gen-probe.com.