NEW YORK (GenomeWeb) – Clinical mass spectrometry has made significant strides in recent years, with several new tests coming to market and instrument vendors registering a number of their devices for clinical use.
Nonetheless, as a US Food and Drug Administration public meeting held last week to discuss clinical mass spec made clear, the field is still in its early days and has yet to fulfill what proponents have long billed as its great potential.
At the gathering, FDA officials and leading clinical mass spec researchers tackled a variety of challenges currently facing the field, from questions of assay development and implementation to regulatory philosophies and commercial demand.
The meeting spent relatively little time on mass spectrometry itself, a trend apparent at other meetings on clinical mass spec, focusing instead on the sample preparation and data interpretation steps that have proven the actual sticking points in the process.
For instance, Christopher Shuford, technical director at Laboratory Corporation of America, provided a detailed look at that company's process for selecting the peptides used in its mass spec-based thyroglobulin test, as well as the variety of approaches he and his colleagues used to optimize the assay's tryptic digestion step.
Another key issue was the appropriate selection of internal standards and calibrators for mass spec-based proteomic assays, which NIST researcher Mark Lowenthal tackled in a presentation detailing best practices for developing and employing these tools.
Broadly speaking, the field has managed to demonstrate that mass spec-based protein quantitation is, in certain instances, clinically feasible. Now the challenge becomes to develop, implement, and popularize the standards and processes that will enable it to become a more widespread tool.
This latter process is moving slowly, however. In theory, mass spec offers a variety of advantages over immunoassays, including better specificity and accuracy and improved multiplexing capabilities, but realizing those advantages remains a challenge. For instance, Andy Hoofnagle, head of clinical chemistry at the University of Washington, presented data from a recent College of American Pathologists proficiency study comparing various platforms for measuring 25-OH vitamin D. The study found that in practice, mass spec offered little if any performance improvements over commonly used immunoassays.
Calling the study "a real call to the field that says we need to be thinking very carefully about how we are doing LC-MS," Hoofnagle noted that a number of publications provide guidelines for proper implementation of such assays. He highlighted in particular the Clinical & Laboratory Standards Institute's C62-A document that includes guidelines for a large portion of the LC-MS workflow.
Hoofnagle added, however, that the document does not cover several topics central to proteomic assays, including peptide selection, digestion optimization, and the use of peptide and protein internal standards and calibrators. Given these gaps, he suggested that a protein-specific CLSI document was needed.
Beyond technical issues, questions of commercial viability were also discussed, with Arizona State University Biodesign Institute researcher Dobrin Nedelkov noting that to date, only 12 CLIA tests using mass spec for protein measurement have been developed. Of those 12, Nedelkov said, three — assays to measure cystatin C, beta-2-microglobulin, and retinol binding protein developed by his former company Intrinsic Bioprobes — were discontinued due to a lack of commercial demand.
"After 30 years of the electrospray and MALDI revolution, we have [fewer] than 10 mass spec assays," he said, comparing this to the initial development of ELISAs in the 1970s, which moved at a much more rapid pace. The recent launch of Sera Prognostics' mass spec-based PreTRM test brings the number of tests to an even 10, though the basic point still stands.
On a similar note, Hoofnagle observed that of the roughly 7,700 CAP-accredited laboratories, only around 2 percent are currently running LC-MS/MS assays, and the number using LC-MS/MS for protein assays is significantly lower.
Nedelkov suggested that a major issue for the field was a lack of demand for mass spec-based assays, given the popularity of immunoassays. Even tests developed using mass spec for discovery often move to ELISA or other immunoassay platforms for commercialization, he noted, citing the examples of Vermillion's OVA1 and Applied Proteomics' SimplyPro Colon tests.
"I would argue that there is very little commercial demand currently for mass spec-based protein assays," he said, adding that a related issue is the complicated nature of LC-MS, and of trypsin digestion in particular.
Essentially, Nedelkov argued, mass spec-based tests are unlikely to gain significant traction as upgrades to existing immunoassays. Rather, he said, researchers need to identify biomarker content that can only be measured using mass spec.
"We can measure post-translational modifications, we can measure changes in the protein sequence," he said. "We need to discover something and come up with a killer app that will detect protein changes that are related to certain diseases that ELISAs cannot detect."
Daniel Chan, director of clinical chemistry at Johns Hopkins University School of Medicine suggested that, in the short term, automation could help simplify mass spec-based proteomics assays and potentially make them more competitive with immunoassays commercially. Longer term, he said, perhaps the field could move toward the measurement of intact proteins, which would eliminate the need for trypsin digestion.
Digestion has long been one of the thorniest steps of the proteomic workflow, with researchers struggling to speed and standardize the process while accounting for factors like the differential decay of peptides during digestion and differences in behavior between endogenous proteins and recombinant protein or peptide standards.
These matters are still being worked out, but despite the challenges involved, protein digestion assays are fundamentally similar to drug metabolite assays, which likewise involve converting a substrate to a product which is then measured by mass spec, suggested Russell Grant, vice president of research and development at LabCorp, "so this isn't new." In fact, he noted, given the larger amount of information generated by a protein mass spec experiment, compared to a small molecule experiment, mass spec could prove even better suited for protein analysis in some respects.