AMP report maps progress toward identification of filamentous fungi and mycobacteria
Interview by Steve Halasey
Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) has earned a position of importance among basic and clinical researchers with a need to identify and characterize organisms found in a wide variety of settings. So far, clinical laboratory applications of the technology have been mostly focused on the identification of bacteria and yeast, where its capabilities have received high marks. But some researchers believe the technology can also be used for the identification of fungi and mycobacteria, as well as for antimicrobial susceptibility testing, potentially making MALDI-TOF MS a new workhorse technology for use in clinical microbiology laboratories.
To investigate the key opportunities and obstacles in this area, the Association for Molecular Pathology (AMP) appointed a working group on MALDI-TOF MS whose initial findings appear in a special report in the November 2016 issue of the Journal of Molecular Diagnostics.1 To find out more about the working group’s assessment of the current status and future prospects for the application of MALDI-TOF MS in routine clinical practice, CLP recently spoke with Christopher Doern, PhD, associate director of microbiology at Virginia Commonwealth University Health System and lead of AMP’s working group on the use of MALDI-TOF in infectious diseases.
CLP: A significant body of published research has demonstrated the superior performance of MALDI-TOF MS for routine identification of bacteria and yeast. How broadly and in what settings has the technology been adopted for such routine clinical uses?
Christopher Doern, PhD: Clinical microbiology laboratories in the United States are at varying stages of MALDI-TOF MS acceptance and adoption. I do not think that at this point anyone would question the superior performance of the technology when compared to conventional identification methods for bacteria and yeast. However, cost is still a significant barrier to entry for some laboratories, and the significant initial capital investment has prevented some from adopting MALDI-TOF MS. In addition, until recently, no FDA-approved MALDI-TOF MS instruments have been available, and that fact has prevented many institutions from pursuing MALDI-TOF MS for clinical use.
So with that said, and without any peer-reviewed data to support this claim, I believe most large academic medical centers in the United States have adopted or are in the process of adopting MALDI-TOF MS for routine identification of bacteria and yeast. MALDI-TOF MS has also been widely adopted by most major reference laboratories, and it is now very common for state and public health laboratories to use MALDI-TOF MS as their primary method of identification.
CLP: Application of MALDI-TOF MS for organism identification requires an extensive reference database of organisms as well as some sophisticated preanalytical processing. Have these requirements slowed the adoption of the technology for routine clinical use?
Doern: I can only speculate as to why laboratory directors have chosen not to adopt this technology. I suspect that some have been slow to adopt MALDI-TOF MS because it is such a departure from traditional organism identification. That choice may result less from the sophistication of the preanalytical process as from the completely different techniques that are required, as well as their departure from traditional biochemical analyses. As clinical microbiologists, we have been trained to think about organism identification in terms of biochemical reactions, not protein profiles. For most of us, it is somewhat intimidating to make that shift. It is not much of an exaggeration to say that for routine organism identification, MALDI-TOF MS has rendered the knowledge of most traditional biochemical analyses obsolete, and that change is something that many of us resist.
CLP: The focus of the AMP report is on emerging and future applications of MALDI-TOF MS, beginning with identification of filamentous fungi and mycobacteria. What is the status of the reference databases available for these organisms?
Doern: These databases are at varying stages of development, and are rapidly evolving. Because they are changing so rapidly, it is difficult to provide a specific answer to this question. However, I can say that there is not currently an FDA-approved database for either organism group.
CLP: Labs may create their own reference databases or use manufacturer-developed databases. What are the advantages and disadvantages of each approach for clinical applications?
Doern: The advantage of a laboratory-created database is that it can be improved over time. Organism spectra can be added to the database, so that it reflects strains found at a given institution that might not be present in manufacturer-developed databases. Effectively, the database can “learn” to identify organisms that it could not recognize initially. The disadvantage of such a database is that the process of adding an organism to a laboratory-created database is not trivial and requires a sophisticated and somewhat labor-intensive approach. In addition, the laboratory needs to make absolutely sure that it has confirmed the identity of the organism being added to the database. Ideally this would be done through a sequencing-based approach. In practice, few laboratories actually create their own databases.
The advantage of manufacturer-developed databases is that they can include a large number of strains that individual laboratories might not be able to access, and therefore could not include in their own database creations. Manufacturers are also capable of updating and improving their databases over time, much in the same way that laboratory-created databases are improved. A disadvantage of such manufacturer-developed databases is that most have not been approved by FDA, and are therefore labeled for research use only (RUO). For most institutions, adopting an RUO-status database for clinical applications would pose some significant regulatory challenges.
CLP: Findings highlighted in the report suggest that identification of varied organisms works best when the reference database incorporates distinct species of the target organisms extracted using the same method as that to be used in testing, and organisms of the same age as those used for testing. How will labs implement these instructions? Will commercially available databases also satisfy these needs?
Doern: Organisms are identified with MALDI-TOF MS through an analysis of their protein profiles, which may differ depending on the conditions under which the organism is incubated (ie, temperature, duration, atmosphere) and the manner in which the protein is prepared for analysis. Optimal organism identification only occurs when the same method that was used for database creation is also used for organism identification. This will be the case for both laboratory-created and commercial databases. Due to the slower growth rates and unique growth requirements of such organisms, laboratories will have to think carefully about their workflow. For example, our laboratory recently evaluated the use of MALDI-TOF MS for the identification of filamentous fungi. But our current protocol requires a 36-hour incubation period, which has proven to be an awkward time point that does not fit well in our current workflow.
CLP: The report points out that both direct cell analysis and preanalytical aerosol-generating manipulation of organisms can create safety issues. What are the risks, and what procedures should labs use to verify organism inactivation before use?
Doern: Laboratories should continue to follow the same protocols and procedures they have always used to ensure the safe identification and handling of potentially dangerous organisms, regardless of identification system. With that said, MALDI-TOF MS processing of mycobacteria and fungi does introduce some new safety challenges, if for no other reason than that there will be additional steps involved in handling the organisms. Laboratories must ensure the inactivation of mycobacterial and fungal organisms prior to placing them on the MALDI-TOF MS instrument.
The unfortunate reality is that these are slow-growing organisms, so performing “real-time” viability assessments is not possible. As a result, robust validation studies should be performed to ensure that organisms are reliably inactivated. In addition, it is prudent to perform ongoing analyses that ensure the process continues to inactivate organisms once validation studies have been completed and clinical testing has begun.
CLP: The report also examines the possible application of MALDI-TOF MS for antimicrobial susceptibility testing, and finds that there may be some significant challenges. What are the key obstacles that might inhibit use of the technology for this purpose?
Doern: Microbiologists understand that MALDI-TOF MS cannot actually be used to perform susceptibility testing; rather, it is used as a resistance test. The MALDI-TOF MS methods that are closest to being implemented for clinical testing can only detect a limited number of resistance mechanisms. While detection of such mechanisms can tell physicians a lot about what antibiotics they cannot use, it doesn’t tell them much about what antibiotics they can use. In other words, “susceptibility results” generated via MALDI-TOF MS don’t actually tell the physician what antibiotics an organism may be susceptible to.
A second limiting factor, related to the first, is that MALDI-TOF MS “susceptibility testing” doesn’t eliminate the need to conduct any other susceptibility testing, and therefore adds more work to already understaffed microbiology laboratories.
And a final obstacle has simply to do with a laboratory’s workflow. The methods required to use MALDI-TOF MS for true susceptibility testing are improving and becoming more streamlined. At the moment, however, they are still somewhat labor-intensive, and many laboratories would struggle to fit this testing into their routine workflow in such a way that it would constitute a real advantage over current methods.
CLP: Overall, how does the workflow of MALDI-TOF MS compare to that of other testing methods, such as PCR-based molecular diagnostics or next-generation sequencing?
Doern: It is difficult to compare MALDI-TOF MS to the workflow of molecular testing methods because they are used for such different purposes. MALDI-TOF MS is purely an organism identification tool. PCR-based molecular diagnostics are used primarily for direct-from-specimen testing. Next-generation sequencing (NGS) really hasn’t found a place for routine use in the clinical microbiology laboratory yet, although many feel that will soon change.
Differences in clinical utility aside; MALDI-TOF MS workflow is very simple and, on a per-organism basis, requires very little hands-on-time. This differs markedly from traditional PCR assays and NGS, which require unidirectional workflow and significant amounts of hands-on time. By contrast, new-generation PCR assays afford an extremely simple workflow that requires only minutes of technologist hands-on time and combines all facets of nucleic acid amplification and detection into a single, contained assay.
CLP: Clinical labs that are considering the adoption of MALDI-TOF MS will be seeking areas in which the technology can improve both their diagnostic and financial performance. Where are the greatest opportunities for achieving such improvements?
Doern: This is a great question and a fun one to answer for those of us who have been through the process of adopting this technology. I’ll briefly address the financial question first. On a per-test basis, most laboratories calculate that MALDI-TOF MS costs between 10 and 50 cents. In most cases, this makes for a fairly attractive return on investment (ROI) calculation that most institutions will readily support. Of course, the ROI will depend heavily on laboratory volume, purchase price, service contracts, and conventional test method pricing.
The far more interesting discussion centers around the impact of MALDI-TOF MS on diagnostic testing and clinical performance. In some ways, it is difficult to quantify the benefits of MALDI-TOF MS, because they are indirect. For example, MALDI-TOF MS works so well that my laboratory has discontinued carrying many of the biochemical tests that had previously been needed to trouble-shoot conventional methods. While this benefit is hard to quantify on the primary plate-reading bench, it is very noticeable in the quality control (QC) area, where we no longer have to manage all of the additional reagents.
Another example of this is in dealing with difficult-to-identify organisms. A small evaluation of our ability to identify organisms from cystic fibrosis patients (typically very difficult to identify) shows that with MALDI-TOF MS we can significantly reduce both the number of methods and the time required to identify those organisms. In this analysis, we found that the primary conventional method failed to reliably identify organisms in about 33% of cultures. By contrast, MALDI-TOF MS provided reliable identification in nearly every culture.
In my experience, MALDI-TOF MS consistently improves turnaround time not only for organism identification—usually by at least 24 hours—but also for susceptibility results. Considering that MALDI-TOF MS doesn’t provide true susceptibility results, this seems strange. But what this result says to me is that, when using conventional methods, we had in some cases been waiting on the identification of an organism before releasing susceptibility results. With the use of MALDI-TOF MS, we accelerated the time to organism identification and therefore also accelerated susceptibility result reporting. This effect is most notable, and probably most impactful, in the diagnosis of bloodstream infections, where the implementation of MALDI-TOF MS alone reduced our time to identification by exactly 24 hours. There are many other examples of the benefits of MALDI-TOF MS, but these are some of the most important ones.
CLP: The Clinical and Laboratory Standards Institute is developing a guidance document on the use of MALDI-TOF MS in clinical applications. How do you expect that guidance will reinforce or differ from the observations of the AMP working group report?
Doern: I think they will differ quite a bit. The benefit of the CLSI report is that it will standardize some of the variable elements of MALDI-TOF MS performance, and provide definitive guidance in areas where that had not previously existed. However, the document is focused exclusively on FDA-approved applications. While this is useful and an important step, it does not address what most laboratories are actually doing at this time.
Many laboratories are using non-FDA-approved systems for organism identification and are pushing the technology far beyond where FDA-approved applications have gone. The AMP working group report describes some of the cutting-edge applications of MALDI-TOF MS that many laboratories are currently evaluating for clinical use. A large number of laboratories will likely be interested in these applications long before they are FDA-approved. The AMP working group report is not meant to be a formal recommendation; rather, it is our hope that it will serve to highlight what is possible with MALDI-TOF MS and inform those looking to apply it for the care of their patients.
Steve Halasey is chief editor of CLP.
- Special article: emerging and future applications of matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry in the clinical microbiology laboratory; a report of the Association for Molecular Pathology. J Mol Diagn. 2016;18(6):789–802; doi: 10.1016/j.jmoldx.2016.07.007.