NEW YORK (GenomeWeb News) – This month, proteomics researchers and industry players met with US Food and Drug Administration regulators to discuss the current status of and future outlook for the field's clinical ambitions.
In large part, the challenges facing protein test developers are the same as those facing molecular diagnostics. However, several questions more specific to proteomics efforts emerged from the discussions.
Of particular concern was the question of how high a bar the agency will ultimately hold mass spec-based protein tests to compared to their immunoassay counterparts.
As Leigh Anderson, founder of SISCAPA Assay Technologies, noted in a presentation during the meeting, mass spectrometry-based clinical proteomics could offer substantial improvements over conventional immunoassay technology which often struggles with issues like specificity and interferences.
However, he added, "there is a potential downside" to mass spectrometry – at least from the perspective of getting tests through FDA – because researchers can "learn so much from mass spectrometry about the performance of an assay… that there are more questions to ask about a mass spectrometry assay" than an immunoassay.
The proteomics community "would love to be on the positive side where the additional capability [of mass spec compared to immunoassays] proves to be an advantage in the [FDA] registration process and not a disadvantage, raising more questions than they answer," Anderson said.
Addressing a similar question during a Q&A session at the end of the meeting, Julia Lathrop, a scientific reviewer with the FDA's Office of In Vitro Diagnostics and Radiological Health said that this was "a point the agency is aware of."
"The more you look, the more the universe of possible things you can look at, and the bar becomes impossible to clear," she said, regarding the possible challenges raised by the large amount of information generated by mass spec assays. "That's not what we want to have happen."
Lathrop suggested that one venue for addressing these concerns would be during the agency's pre-submission process where instrument and test developers could, in a sense, prepare the field, addressing what data they believed was and was not relevant to their submissions.
"One of the reasons we want to have these discussions is to better understand, [for instance,] you have all these [mass spec] peaks, but these [specific peaks] aren't the things that we really need to be worried about," she said. "We need that sort of feedback."
Last year saw FDA grant the first 510(k) clearances to mass spectrometers as the agency cleared Shimadzu's Axima Assurance MALDI-TOF and Bruker's Microflex MALDI-TOF as part of, respectively, BioMérieux's Vitek MS and Bruker's IVD MALDI Biotyper clinical microbiology systems. At the meeting, Yvonne Shea, one of the FDA reviewers involved in those clearances, provided a look at the process.
Like Lathrop, Shea stressed the importance of the pre-submission process, suggesting that this was one of the keys to the success of the Vitek MS and Biotyper 510(k) submissions.
"We are very, very open to talking about the studies. We may suggest something, and if you don't think it's reasonable or you don't agree, then we want that discussion," she said.
Shea also detailed critical aspects of the agency's review of the instruments, noting that one of the biggest areas of focus was the instrumentation software.
FDA is interested in "how you are acquiring your spectra, how you are analyzing it," she said. "So we asked for all the software documentation, the verification and validation of all the software."
Additionally, Shea said, the agency wants "to know if there is a quality control tool to monitor your mass spec" and what algorithm is used to ensure that the instrument is working properly.
One of the "most important things," she added, "is that we want to see the sample report – we want to see what you are reporting to the user."
While clearance of the Vitek MS and Biotyper was no doubt an encouraging development for proponents of clinical mass spec, the systems' intended use – using MALDI mass spec for identification of microorganism colonies cultured from humans – is somewhat different from the protein biomarker assays being widely pursued in the field.
For these assays, which typically involve measurement of surrogate peptides using a triple quadrupole instrument, a number of technical and regulatory hurdles remain, Anderson said during his presentation, noting that while several vendors have registered triple quadrupole instruments with FDA as Class I devices, none of these instruments are quality system regulation-compliant (which FDA requires for IVD products) or FDA-cleared.
Anderson added that an exercise that could help move clinical proteomics along would be for a group to actually attempt a 510(k) submission for a triple quad-based protein assay – an effort that, he suggested, would help both the field and FDA better assess the state of such assays and work through the issues that must be resolved to move them into the clinic.
A significant portion of the day's discussion revolved around these issues – in particular the challenges of developing accurate, reproducible mass spec-based protein assays. As at previous meetings focused on clinical proteomics, the reproducibility of trypsin digestion was a major concern, with considerable attention paid specifically to the question of how differential rates of peptide degradation might affect assay quality.
During his presentation, University of Washington researcher Andy Hoofnagle suggested that good assay calibration was essential to addressing these questions, calling calibration "the cornerstone of repeatability."
Citing data from the IGF-1 LC-MS/MS Working Group, an effort by an international group of researchers – including Hoofnagle – to develop and standardize mass spec assays for IGF-1, he noted that using calibration curves, the team was able across five labs in three different countries to achieve an average CV for their assay of 16 percent. He added that this could be further improved by using single-point calibration, in which a single human sample is used to calibrate a batch of assays.
"So it is possible with LC-MRM assays to do fantastic standardized assays around the world," he said.
In addition to the specific technical and regulatory challenges of mass spec-based proteomics, the meeting also tackled more general questions around matters like structuring clinical studies, demonstrating clinical utility, and determining intended use for a test.
While these questions aren't specific to clinical proteomics, examples given by some presenters suggested that the field remains somewhat naive regarding such issues.
Lathrop, for instance, cited several examples that she said were "hypothetical" but drawn from existing tests or submissions, including a test that claims to identify all cancers using a single analyte, and a submission with an intended use so broad that, she noted, "it would take years and millions of dollars, when in reality they really want a much narrower claim."
She cited, as well, cases of devices developed using entirely incorrect patient populations, something FDA Medical Officer Abraham Tzou likewise noted as a common flaw during his presentation on study design.