NEW YORK (GenomeWeb) – At a US Food and Drug Administration public meeting held earlier this month to discuss clinical mass spectrometry, participants raised the question of what level of evidence the agency would require in a 510(k) submission for a mass spec test intended to replace an existing immunoassay.
If, for instance, the mass spec assay and immunoassay showed little correlation, could it be assumed this was due to poor performance on the part of the immunoassay, or would applicants need to gather additional clinical data to demonstrate that the mass spec assay did, in fact, represent an improvement over its immunoassay predicate?
Because it was a meeting of mass spectrometrists, there was some support among participants for the former option. Offering the opposing view was Johns Hopkins researcher Daniel Chan, who suggested that, whatever their limitations, FDA-approved immunoassays presumably had clinical value that should not be set aside so lightly.
"I had to get up and say, 'If you compare the two methods, and the existing method is an immunoassay that is FDA approved and we are using it in clinical laboratories, we know that it is not perfect but we know that it is giving useful clinical information,'" Chan, who heads Johns Hopkins' clinical chemistry division and pathology core, told GenomeWeb this week. "And so, when you come in with a mass spec assay and you have no correlation [to the existing immunoassay], people cannot assume that you are right and the existing assay is wrong. You have to prove, not just to the FDA but to the clinical lab director and maybe the clinicians and patients, that indeed your assay is better."
The discussion was a relatively narrow one focusing on the FDA's thinking around the evidence required for 510(k) submissions in cases without quality predicate devices, but it raised a larger question fundamental to the business of clinical mass spec—to what extent thus far has mass spec actually proved in practice a broadly superior technology to immunoassays?
Mass spec's advantages — among them, better specificity and accuracy and improved multiplexing — are well known and frequently discussed within the community. But are these advantages enough to actually dislodge existing immunoassays or to make mass spec the dominant technology for proteomic tests to come?
Despite significant technical advances in recent years, mass spec's complexity and long (at least compared to common immunoassays) turnaround time continue to make it a difficult sell as a clinical technology. It has managed to gain a foothold for certain assays like thyroglobulin testing, where existing immunoassays had obvious flaws, but, such examples aside, it has made little headway as a replacement for conventional tests.
And while companies including Integrated Diagnostics and Sera Prognostics have released new proteomic tests on mass spec platforms, the majority of such tests still come to market in immunoassay format. And, in fact, comments from Indi CEO Albert Luderer indicate that even firms like his, that have pioneered clinical LC-MS/MS, would under certain conditions consider a shift from mass spec to immunoassay.
Where within clinical proteomics might mass spec make significant inroads? What path, or paths, might it take that could lead to its broader adoption?
One possible route, particularly with regard to replacing existing tests, is to leverage the technology's multiplexing capabilities, suggested SISCAPA Assay Technologies CEO Leigh Anderson.
"The idea that we just march in with mass specs and replace all the Roche and Siemens analyzers is obviously naïve, because those are cost-effective ways of doing these tests today — as long as you are doing them in the standard way," Anderson said. "If you are competing on a single-analyte test, the mass spec assay as it is implemented in clinical laboratories today, which is, generally speaking, not highly automated, is not cheaper than an immunoassay."
Once you begin multiplexing assays, though, the costs savings offered by mass spec could become attractive, he said.
"In an immunoassay model, each assay adds a fixed increment of cost, but in mass spec the workflow is what costs the money, and the incremental cost per added analyte is much lower," Anderson said. "So, as we go further with panels, we are going to get to the point where it is substantially less expensive, and that is a totally different argument."
Chan agreed that mass spec's multiplexing ability would likely be key to any broad clinical adoption. He suggested that mass spec assays combining 10 or 20 different analytes could prove very attractive to clinical customers, though he added that the technology wasn't at the point at which such multiplexing could be implemented routinely in the clinic.
"For research we routinely multiplex 100 different molecules, but I would not encourage that for clinical use," he said.
Another factor Anderson said he believed would drive adoption of mass spec is the move—at this point still largely speculative — toward less invasive patient sampling techniques like dried blood spots.
Dried blood spots have been used for decades as a sample collection method for newborn screening and have more recently drawn interest for drug development and clinical trial work. They also appear potentially useful as sample format for proteomic testing, with a number of researchers and firms including Anderson and SAT looking into the feasibility of using them in clinical mass spec assays.
Because dried blood spots involve a simple finger prick as opposed to a blood draw and can be sent unrefrigerated through standard mail, they offer significant advantages in terms of convenience and cost. And, were the clinical testing industry to move toward dried blood spots as a sample source, mass spec could prove an effective technology for running such assays.
"As long as we are drawing big tubes of blood, then there is plenty of sample to do as many immunoassays as you need," Anderson said. "But in a world of microsamples like dried blood spots, there isn't enough material to divide between all the immunoassays you might want to run. So with that kind of sample you need to be able to multiplex your tests on a very small volume."
"Clearly [dried blood spot sampling] is a small application currently, but it has the opportunity to be huge because of the enormous improvement in patient convenience and the really profound cost improvement compared to people having to go to a clinic or a doctor's office and get blood drawn by a phlebotomist and then ship those samples cold by FedEx," he added. "And I think the applicability of mass spec specifically to this new sampling paradigm is one of the key things that is going to drive it."
Anderson questioned, though, the extent to which large reference labs like Quest and Laboratory Corporation of America would be at the forefront of any drive toward mass spec. "People who are in the diagnostics business on a large scale right now are, generally speaking, pretty comfortable with its dependence on the existing immunoassay technology," he said.
At the same time, mass spec is beyond the capabilities of many smaller hospital labs, Chan said, citing the example of the recent development of mass spec-based thyroglobulin assays.
Those assays "are mainly performed by a few large reference labs — LabCorp, Quest, ARUP, Mayo Clinic," he said. "And part of the reason is that mass spectrometry is much slower than immunoassays, it requires a lot of technical expertise to perform, and the instrument costs a lot. So, for a small hospital lab, they cannot afford it."
Chan also noted that mass spec vendors, by and large, don't yet provide the level of instrument support required by a clinical lab.
"In my clinical lab at Johns Hopkins, mass spec is the least reliable of all the instruments I have," he said. And yet, "the support we receive from mass spec companies today is not up to [the standard] of most clinical diagnostic companies."
"A good example is that many mass spec companies only provide five-days-a-week service," he said. "But what about on the weekends? If we are going to use the mass spec for 24 hours a day, seven days a week, then you have to provide that kind of service."
Development of novel test content provides another possible route for mass spec's expansion into the clinic. Here, Indi's Luderer told GenomeWeb, the technology's great utility as a discovery tool plays significantly into its appeal as a clinical platform, letting test developers move more quickly from discovery through to validation and commercialization than they could if they had to convert their tests over to an immunoassay format.
"We measured hundreds of [proteins] simultaneously and ended up looking at probably several thousand different proteins, and you can't do that with antibody technology," he said. And transferring those mass spec discoveries to an immunoassay format would have been no trivial matter.
"What you see in a mass spec is sometimes hard to replicate on an antibody basis," Luderer said. "I think if we would have done it, it probably would have added two years to our commercialization date." Indi launched its mass spec-based Xpresys Lung test, an 11-protein panel for assessing the malignancy of lung nodules, in 2013.
"I think what we have shown is that there is a lot of information that you can glean from taking a mass spec approach, and you can create some very powerful diagnostics," Luderer said. "It is very hard to actually execute a technical strategy [in proteomics] without using mass spec, and we think that proteins are where the action is going to be. So, I think if there is going to be a future for new analyte discovery in the proteome, mass spec is going to be a central player."
And, if firms follow Indi's route to market, that would mean increasing adoption of mass spec as a clinical platform. Though, Luderer noted, not all companies will prioritize speed to market as Indi did.
"If you are in a large company that has lots of finances and would rather populate a global immunoassay platform, you might hold off and wait for FDA approval of a kit," he said.
And, in fact, even Indi might not stick with a mass spec format in the long run, Luderer said, noting that as the company gathers more data on its test, it might become possible to winnow down the number of analytes in the panel, which would make transitioning to an immunoassay format more attractive.
Asked if the company would switch the Xpresys test to an immunoassay format were the panel to become small enough, Luderer said that he believed it would.
"The time to get an answer is going to be a lot shorter because the analytic preparative steps for immunoassay are just much simpler than they are for a mass spec," he said. "Time is money, and turnaround time also [affects] the expectations of physicians ordering the test."
Such a path, from mass spec to immunoassay, is not uncommon, Chan said, citing the example of small molecule assays like drug testing, for which mass spec has been used for many years.
In many cases, he said, tests that were once done on mass spec have moved to immunoassay as good antibodies to the analytes being measured became available.
"Most of the high-volume drug [tests] are on immunoassay systems now, with the exception of, for example, immunosuppressants or new drugs that come out and don't yet have good antibodies," Chan said. "In my clinical laboratory I offer several mass specs just for those molecules, but we have now narrowed it down to very few markers. So, it may be the same thing for proteins, that the majority will still go to immunoassays, because they are very fast and they are reliable.
"When everybody tells me that immunoassays are no good and that we are going to go to mass spec, I often have a problem with that," he said. "I'm trying to say, 'Hey, wait a minute, that is not necessarily the case.' We are moving toward mass spec more and more, but we are still not there yet."