Mass Spectrometry in the Clinical Laboratory: How Have We Done, and Where Do We Need to Be?

Mass spectrometry has long been used in clinical laboratories. The first important applications were the use of gas chromatography–mass spectrometry (GC-MS) for drug screening and the diagnosis of organic acid metabolic disorders. More recently liquid chromatography–mass spectrometry [LC-MS; liquid chromatography–tandem mass spectrometry (LC-MS/MS1 )] has been introduced successfully into clinical laboratories, where it has become the gold standard for newborn screening. The prices for basic LC-MS/MS systems have decreased, and these systems are adequate for many clinical applications. At the higher end, LC-MS/MS systems with enhanced capabilities now permit the analysis of clinically important compounds that are present in very low concentrations in biological fluids. Although GC-MS and LC-MS instruments provide advantages over other techniques, financial, clinical, and analytical implications must be considered when deciding which technique to choose or interpreting results that may differ from other conventional assays. To gather the observations, opinions, and predictions of individuals with varied experiences in the mass spectrometry arena, we present here a virtual roundtable with a panel of 5 experts: a clinician, a researcher, 2 clinical chemists, and a clinical business manager from an instrument manufacturer. As a pediatric endocrinologist, have you seen an influence of mass spectrometry on your diagnosis and treatment of patients? Joseph Majzoub2 : Absolutely yes, in both positive as well as challenging ways. Mass spectrometry provides better results with much improved specificity over immunoassays for the analytes that I monitor. As a clinician, however, I often deal with multiple laboratories with different levels of testing or different methods. Some laboratories do not do prepurification of samples before immunoassay, which results in different answers because of different effects of structural analogs. The methodologies used by laboratories have also changed over time, so it is challenging sometimes to decide if a change is in the patient or the assay. As an investigator in cancer epidemiology, are there areas where the incorporation of mass spectrometry holds promise in cancer research? Ann Hsing3 : Indeed there are. In my prostate cancer research, we measure sex steroid hormones and other biomarkers. We have compared RIA with GC-MS and LC-MS; the mass spectrometry techniques give us better analytical sensitivity and specificity as well as reproducibility. GC-MS and LC-MS also allow us to see multiple metabolites or obtain a profile of compounds with the same amount of serum, which is an advantage over the conventional RIA. Are there new diagnostic areas where mass spectrometric assays will be important? Joseph Majzoub: There are some analytes for which assay improvements are needed. In my experience, estradiol and testosterone have been difficult to quantify with optimal specificity, and clinicians would like to see improvement here. Another example is urinary free cortisol. I believe that once we migrate to standard platforms, or know that assays have been validated against MS as the reference method, we will get better results. For large-scale studies that involve precious sample types, what characteristics do scientists and funding agencies want in assays? Ann Hsing: Prospective and nested case-control studies usually include analysis of samples that were collected many years before the diagnosis of disease (the endpoint) in case subjects. These samples cannot be replaced easily. Because of this, sample volume is an important issue to consider in prospective studies. Biomarker studies that require large sample volumes (>0.5 mL) should be planned with caution. As a clinical chemist, could you comment on your experiences using LC-MS/MS for drug screening? Do you see advantages or challenges? Alan Wu4 : LC-MS/MS has a number of advantages compared to GC-MS. First, sample extraction and pretreatment are minimized with LC-MS/MS. There is less dependence on the need for a drug or metabolite to be volatile, and no need for derivatization to create a volatile compound. Polar analytical columns allow analysis of glucuronides. Nonetheless, GC-MS is a good technique and we use it as well. With GC-MS the chromatography is often sharper and the peak resolution better. And the price tag for a good GC-MS system can be a fraction of the cost of an LC-MS/MS system. In the toxicology arena, the software libraries, algorithms, and information such as degree of fit are more advanced with GC-MS systems. What new scientific approaches do you see unfolding in MS that could increase analytical sensitivity and specificity? Alan Wu: At the detection end, there has been an increase in the toxicology world in examining time-of-flight (TOF) analyzers for drug screening. TOF can provide part-per-million mass accuracy and could reduce the number of candidate compounds causing the signal, thus giving improved specificity. One would rely on retention time and exact mass without the need for fragmentation. Of course this would be qualitative, but unknown screening would be a great candidate for TOF analysis. Alan Rockwood5 : Some laboratories have had success using microspray and nanospray interfaces, as well as microcolumn separations to enhance sensitivity. Microfluidic device applications work well in some applications, particularly in R & D environments, but often are difficult to use and do not hold up well in a high-volume production environment. Newer developments and products such as the Advion NanoMate and the Agilent HPLC-Chip Cube may hold promise for more robust and easy-to-use microfluidic applications in clinical mass spectrometry. Our laboratory has had success in improving specificity by using laboratory-developed 2-dimensional separations coupled with mass spectrometry. In our hands, multidimensional separation approaches provide improved sample cleanup and peak purity, as well as faster run times, compared with single-dimensional separation methods of comparable selectivity. They also keep instruments cleaner. Are there new instrumental designs or adaptations that will extend our capabilities in quantitative mass spectrometry? Donald Mason6 : Most recent advances in mass spectrometry systems used in clinical chemistry laboratories have been evolutionary, not revolutionary in nature. Instrument manufacturers are making incremental modifications to improve sensitivity, specificity, and stability by enhancing ionization efficiency, increasing the efficiency of ion transfer, and improving the sensitivity of detectors. Whereas each change provides an incremental improvement in performance, together the improvements are significant, allowing clinical laboratories to achieve lower detection limits with greater precision. One area for improvement, however, is in systems integration. Future mass spectrometry–based clinical analyzers must one day incorporate sample pretreatment with separation and detection. There are early examples of this today. The Spark Holland systems integrate solid-phase extraction and chromatography and can be coupled to a mass spectrometer. Turbulent-flow chromatography has also aided sample preparation, as have robotic sample pretreatment platforms, e.g., Tecan. However, there are still significant developments required before mass spectrometry–based clinical analyzers become a reality. Alan Rockwood: Ion mobility and related techniques such as traveling-wave ion mobility spectrometry and field-asymmetric waveform ion mobility spectrometry (FAIMS), when combined with mass spectrometry, are finding their way into clinical laboratories. These techniques add another degree of selectivity. For example, in a validation of a cortisol assay in our laboratory, we have seen large improvement in selectivity using FAIMS and a comparable improvement during research studies on a testosterone assay. Another design improvement is the linear ion trap, which allows enhanced selectivity through MS/MS/MS analysis. Linear ion traps may be stand-alone instruments or an option on triple-quadrupole instruments. I suspect that, as we use this type of detector, we may see big improvements in selectivity and improved quantification. Given the choice of throughput, limit of quantification, or specificity, which do you see as most desired in the clinical laboratory? Joseph Majzoub: Assay specificity is important for the tests that I routinely request. With the exception of cases of ambiguous genitalia, turnaround time is not an issue. As far as detection limits, there is really not anything that we cannot detect adequately. So I would say that specificity is at the top. Alan Wu: Limits of detection and quantification are going to be important as we find new markers and need to get down to attomolar and femtomolar concentrations. In the toxicology area, there are drugs such as fentanyl that have been difficult to detect due to their low concentrations. Donald Mason: Actually, all are important and have been expressed equally as needs by clinical laboratories. Clearly there will be trade-offs. For example, you may achieve very high throughput at the expense of sensitivity or specificity. Some examples may help clarify these different needs. Assays for steroids such as estradiol or testosterone, particularly in females and pediatric patients, require greater precision at the lower end of the reference ranges. This precision cannot be sacrificed for the sake of higher throughput. For high-throughput newborn screening, modern mass spectrometry–based systems provide sufficient sensitivity. Thus, some sensitivity may be sacrificed for the sake of the high throughput required to screen the 4 million babies born yearly in the US. By contrast, the metabolic testing that follows a presumptive positive result requires lower limits of quantification and greater specificity, but does not require the throughput of mass screening. These factors must be taken into account when designing any assay for routine clinical use. Alan Rockwood: The simple answer is that all three are important. But if I had to choose I would say limit of quantification, followed closely by specificity. These two get us excited because they improve patient care and expand the range of applications of mass spectrometry in clinical chemistry. Ann Hsing: Both throughput and accuracy are important elements in large-scale epidemiologic studies; however, one should never sacrifice accuracy for throughput. How much sensitivity and specificity are needed depends on the biological and intra- and interperson variation of a particular marker in the study population. Of course, our goal is always to minimize assay variation. Do you have any additional comments? Joseph Majzoub: Diseases are very often not physically evident in endocrinology. Thus clinicians rely on laboratory results to diagnose and monitor patients. We want to be assured that LC-MS assays have been properly validated according to guidelines or regulations. There is an opportunity for clinical chemists to come and talk to us and show us the analytical benefits of LC-MS assays. Similarly, it is important that clinicians communicate the type of results that help us most. Ann Hsing: We have not used GC-MS widely in epidemiologic studies due to its high cost and labor-intensiveness. However, this is the direction we are heading. New LC-MS technologies will help improve throughput, making it easier to incorporate this state-of-the-art technique into large-scale epidemiologic investigation. Alan Rockwood: We may have too often regarded mass spectrometry as a low-consumable-cost alternative to immunoassays. This is not always a clear-cut advantage for mass spectrometry, however, and it is better to regard mass spectrometry as a complementary technique having technical advantages over immunoassays for specific applications. Primarily, this revolves around the very high specificity of hyphenated analysis, such as LC-MS/MS. This often translates to a quality advantage, leading to better patient care. It is the quality advantage that is leading the current wave of clinical applications, such as the widespread adoption of mass spectrometry for steroid analysis. Also regarding specificity, I believe that ion ratios or secondary ions should be more widely used for identifying the occurrence of interferences in patient samples. This powerful feature is nearly unique to mass spectrometry. For example, in our laboratory, approximately 1% of specimens received for testosterone testing are found to contain interferences. These interferences would remain “hidden” if we did not use ion ratios, and this would lower the quality of patient care. Last, instrument robustness and reliability are big issues for clinical laboratories. Over the years, I have seen large strides made in instrument reliability, which has helped institutions successfully integrate MS technologies and assays into the laboratories. As a customer, I appreciate this. Donald Mason: One still unaddressed area in the clinical chemistry community is standardization or harmonization of mass spectrometry–based assays. For example, it has been well documented that results obtained for 25(OH) vitamin D3 by LC-MS/MS are extremely variable between laboratories. This is due, in part, to a lack of available standard reference materials. The availability of commercial calibration materials may also offer a route to improving interlaboratory comparability. Until suitable standardization is achieved, it will be necessary to continue to have site-specific reference ranges, greatly compromising research into the ever-expanding list of diseases associated with vitamin D deficiency. Alan Wu: We still need to be mindful of the matrix effects with electrospray ionization. For those thinking about acquiring LC-MS instrumentation and setting up assays in their laboratories, it is important to remember that the instrumentation is more complex than the typical turnkey analyzer, and that some level of expertise in the laboratory is a must. Some companies are better than others at recognizing and responding to the specific needs of the clinical laboratory.