William Check, PhD
“Fear no FISH,“ Dan M. Hyder, MD, advised pathologists at CAP ’10. No, Dr. Hyder was not referring to the infamous shark Jaws. Speaking at the symposium “You Can Do It Yourself! Molecular Testing for Community Pathologists,“ Dr. Hyder, medical director of laboratory services at Northwest Cancer Specialists, Portland, Ore., and Vancouver, Wash., was encouraging attendees not to be intimidated by the molecular technique fluorescence in situ hybridization. His positive message was typical of the answers the three symposium speakers provided to questions that many pathologists have about embarking on molecular testing.
Samuel K. Caughron, MD, addressed two basic questions that pathologists ask themselves: Should I get into this area? Is there a market? “Every community pathologist needs to be looking at molecular testing,“ Dr. Caughron, director of the molecular genetic pathology lab at MAWD Pathology Group in Kansas City, Mo., told CAP TODAY a few weeks after the symposium. “At some point every community pathologist will be getting into molecular testing; it is just a matter of their current practice dynamics and the market opportunities in their area.“ Eventually, he added, molecular tests are going to be part of routine practice, no matter the area of pathology. As for markets, Dr. Caughron recommends looking at tests in the area of women’s health as initial offerings.
General laboratory testing is growing about five percent per year, while the pace of growth for molecular tests is closer to 20 percent, Dr. Hyder says. “This rapid growth is forcing pathologists to consider getting into molecular pathology and driving people to perform make-buy decisions,“ he told CAP TODAY. “They are asking themselves, Do I want to keep sending it out and be responsible for charges? Or do I want to do it myself?“
In his opening talk at the symposium, Jeffrey A. Kant, MD, PhD, of the University of Pittsburgh Medical Center, conveyed a positive message, too, in responding to one of these pervasive questions: Do I need special training to start or expand molecular testing? “In general, no,“ Dr. Kant, professor of pathology and human genetics and director of the Division of Molecular Diagnostics, replied. Much contemporary molecular testing is amenable to being done by trained lab technologists without special expertise.
Automated equipment and growth in the number of IVD assays have been instrumental in making molecular pathology more accessible to community practitioners. “There are now a number of user-friendly platforms with FDA-cleared assays,“ Dr. Kant said. “Most are in microbiology, but there are cleared assays in genetics and oncology as well.“ Dr. Kant called the growth of platforms “a very dynamic area,“ with new ones emerging all the time.
In recognition of the ever-increasing importance of molecular testing for pathologists and of the contribution of advances in platforms and assays to this transformation, this issue of CAP TODAY marks a milestone—its first product guide in the molecular arena. With almost a dozen companies represented, the guide to automated molecular platforms signals that molecular techniques are assuming a standard place in the clinical laboratory.
“This product guide is long overdue,“ says Daniel H. Farkas, PhD, HCLD, VP of clinical diagnostics at the Sequenom Center for Molecular Medicine, Grand Rapids, Mich. More and more labs are getting into molecular diagnostics, he says, and a product guide like the one in this issue offers a “route to comparison shopping“ for molecular platforms. “There is a lot of routine molecular diagnostics now,“ Dr. Farkas adds. “Molecular need not be in a freestanding lab. It can be part of a core lab.“
Frederick L. Kiechle, MD, PhD, a member of the CAP Publications Committee, calls the product guide “an important step“ and says, “The editorial staff decided that CAP members are crying out for help.“ Dr. Kiechle, medical director of clinical pathology at Memorial Healthcare System, Hollywood, Fla., says CAP?members are asking, What’s out there? How can I distinguish various products? “The product guide has tons of information and signals that people in clinical practice are becoming very interested in this technology and want help in figuring out what devices to include in their armamentarium.“
Worth noting, Dr. Kiechle adds, is that the guide shows which assays are FDA cleared and what’s approved only in Europe. “So it gives you a good idea what each company is up to.“ Important in molecular testing is how many specimen types you have to validate. In microbiology, particularly, this can be a challenge. Many microorganisms can infect various body sites—tissue, blood, urine, cerebrospinal fluid, even synovial fluid. Dr. Kiechle’s maxim: “If you haven’t validated dirt, you’re not done.“
Of course, there are differences between molecular testing and more traditional lab specialties. Collection and processing are one example. “Molecular is not like coag or chemistry or hematology, where everyone knows about preanalytical issues,“ Dr. Kiechle says. For a hematology product guide, for instance, collection is not discussed because everyone knows about various types of collection tubes. “These are things that everyone takes for granted,“ he says. In the area of molecular testing, however, preanalytical steps need to be discussed explicitly, because they can affect cost. “For the most part, molecular platforms detect the presence or absence of mutations, so the price of the instrument underestimates the total cost of testing,“ he points out. Molecular assays have three steps: extraction, amplification of the target, and detection of a sequence. Information about how many of these steps each instrument performs is missing from this first version of the product guide, Dr. Kiechle notes. He cites Cepheid’s GeneXpert, which performs all three steps in a single cartridge, as a perfect example of the importance of this information. “With a GeneXpert you don’t need to own nucleic acid extraction and amplification devices,“ he says.
Dr. Kiechle also points out that some companies that make molecular testing devices are not in the guide, such as Illumina, which provides microarrays for mutation detection as well as a third-generation sequencer, and Asuragen, which uses Luminex technology. These omissions highlight the difficulty of knowing where to draw the line on what to include in a molecular product guide. CAP TODAY will publish a guide to array-based and sequencing instruments next year.
Readers will also notice that Roche Diagnostics is not represented in the product guide. Despite having had input into the original template used to gather the data presented in the product guide, Roche decided not to participate for reasons not understood at CAP TODAY. A Roche spokesperson wrote, “We have been having extensive dialogue about it on our end and our molecular marketing team has decided not to participate in the automated molecular systems survey after all.“
Important to realize is that this is a first try at a molecular product guide and thus imperfections are sure to be found that will be fixed in subsequent years. Readers are welcome to bring to CAP TODAY’s attention anything they think can be improved. (Write to firstname.lastname@example.org.)
In his introductory talk at CAP ’10, Dr. Kant addressed several issues related to setting up molecular testing. Laboratorians who want specialized training can get it at courses offered at meetings of the CAP, the ASCP, and the Association for Molecular Pathology. Virginia Commonwealth University offers a two-week practicum in molecular pathology. The three symposium speakers represent three different routes to obtaining molecular expertise and work in three different practice settings. Dr. Kant is a formally trained molecular pathologist who works at an academic institution.
Dr. Caughron recently completed an accredited molecular genetic pathology fellowship at Vanderbilt University Hospital and passed the MGP board examination. Molecular genetic pathology is one of a limited number of subspecialties—like cytopathology, hematopathology, and dermatopathology—in which a pathologist can become board certified. Greater than 90 percent of pathologists completing general training today go on to a fellowship, Dr. Caughron notes. Molecular genetic pathology is a small but growing subspecialty. As training programs have become aware of the significance and importance of molecular genetic pathology, the number of training positions available has increased. Dr. Caughron now works in a community practice.
Dr. Hyder was an academic pathologist with a focus on hematopathology. “Molecular diagnostics probably began in hematopath, with T and B cell gene rearrangement studies,“ Dr. Hyder says. He had no formal training in molecular techniques; rather, he pursued a “seat-of-the-pants“ course. He first learned from younger colleagues in academic practice, then went to a private hospital where he set up a molecular lab, learning from the literature and colleagues. “As hematopathologists we were forced to deal with molecular testing, and I found it much easier to do it myself,“ he says. Dr. Hyder now works in an oncology-focused group practice. To maintain his skills, he is taking the AMP molecular pathology board review course.
Finding out what your clinician clients want can help the laboratory determine what tests to offer, Dr. Kant said, in addressing another basic decision. However, he advised distinguishing real needs from corporate marketing initiatives. As to methods, while labs depend on lab-developed tests for many purposes, LDTs are not the best route when first setting up molecular testing. Many FDA-cleared assays are available for infectious diseases, with some in genetics and oncology as well. If you do plan to use LDTs, consider hiring a savvy PhD, who can also be helpful in other lab areas, Dr. Kant suggested.
Whether you will need two rooms for molecular will depend on whether you adopt conventional open PCR instruments or the newer real-time platforms, which are closed systems. You will also need to decide whether to distribute molecular tests among specialty labs or to have a central molecular facility. In addition, Dr. Kant said, “Integrating reporting into the EMR can prove to be difficult.“
While nucleic acid chemistry does differ from the chemistry of other laboratory testing, Dr. Kant said, “This is really simple stuff.“ Molecular testing is unique in that it is largely tissue-based. “A fresh piece of tissue is always preferable to paraffin,“ he said. Also, it is better that heparin not be present. All molecular analytes are unregulated—they did not exist in the clinical laboratory when CLIA ’88 was formulated. Schedules for proficiency testing don’t exist, nor is the number of challenges clearly specified. “Sophisticated proficiency testing is delivered primarily through CAP,“ Dr. Kant said. There is also a CAP program, the Sample Exchange Registry for Alternative Assessment, for labs doing tests not offered in formal Surveys. “It’s really a great deal—it’s free and satisfies the PT requirements in the molecular pathology checklist.“
In the commercial area, Dr. Kant asked: “Can you trust what vendors tell you?“ As in any other lab field, colleagues are key, he said. You also need to know whether the vendor provides a call center, as well as on-site service. Get the vendor’s ROI spreadsheets and study the assumptions.
Another financial issue is CPT coding, which can be complex for non-microbiology applications. “‘Stacking codes’ for oncology and genetic tests are Byzantine,“ Dr. Kant warned. Make sure you are informed before you start testing. As for genetic tests themselves, he said: “Like it or not, you’ll be dealing with them. You need to be able to consult on genetic counseling and informed consent issues for heritable diseases.“
Dr. Hyder discussed applying molecular testing to oncology. He is responsible for hematopathology, cytogenetics, hemostasis and thrombosis, and pharmacogenomics. “There are molecular components to all of these areas,“ Dr. Hyder said. He uses both FISH and amplification techniques—PCR and real-time PCR.
Clinical utility and the economic market must be considered when deciding whether to offer a test. “I like it when an official group says the evidence justifies doing a test,“ Dr. Hyder said. Choice of a method is influenced by whether it is FDA cleared and whether the platform can perform many tests. In addition, “Intellectual property considerations are cropping up more and more,“ he said.
Verification of FDA-cleared assays is spelled out in CLIA. A more extensive list is in the CAP molecular pathology checklist, June 17, 2010 revision, which requires verifying accuracy, precision, analytic sensitivity, interferences, reference range, and reportable range. “You need to prove in your lab that the test does what the manufacturer says it does,“ Dr. Hyder said. “We do this all the time with other tests.“ A list of FDA-cleared molecular tests can be found on the AMP Web site: www.amp.org/FDATable/FDATable.doc. For non-FDA-cleared assays, there are additional steps, of which verifying diagnostic sensitivity and specificity are the most difficult, Dr. Hyder said. “It may be impossible for you to do this in your lab,“ he added. “I highly recommend using an FDA-cleared test if possible.“
Speaking of specific methods, Dr. Hyder said he enjoys doing FISH. “It is not that difficult if you know your limitations and stick to probes you’re capable of interpreting,“ he said. FISH has greater diagnostic sensitivity than cytogenetics to detect certain genetic abnormalities in leukemias and lymphomas. PCR has greater analytic sensitivity than FISH but isn’t always best for diagnosis, owing to the size of its primers. In situ hybridization can also be viewed with chromogen (CISH) or silver (SISH) nanoparticles; it requires only a standard light microscope. Dr. Hyder advises community hospital labs to stay away from metaphase FISH testing, which requires a cytogenetics lab, and to stick to interphase testing, which is used to determine the number of one or more chromosomes and to detect some specific characteristic chromosome deletions or amplifications. Interphase FISH can be done in 24 hours.
Counting is done with chromosome enumeration probes (CEPs), while deletion or gain at a specific gene region or fusion of gene regions is detected with locus-specific indicator probes (LSIs). “Avoid using fusion probes,“ Dr. Hyder advised. “They are very complicated to do and require a cytogenetics facility. Stick to enumeration of chromosomes or gene regions with CEPs or LSIs.“ As an example of an FDA-cleared enumeration FISH assay, Dr. Hyder points to Abbott’s UroVysion test, which has three CEPs and one LSI. “UroVysion is not specific for urothelial carcinoma,“ he cautioned. “Results must be correlated with the clinical history. And you must follow FDA procedure.“ Commercial FISH assays for the HER2 locus are also available to qualify breast cancer patients for Herceptin. Volume of HER2 testing is high. “Pretty much every lab can bring it in-house,“ Dr. Hyder said. Two FISH assays and one CISH assay are cleared for this purpose; other CISH and SISH assays are before the FDA.
In hematology a lot of testing requires only counting, not detecting fusions. This is true of chronic lymphocytic leukemia panels. A Vysis FDA-cleared five-probe CLL panel is done in two tubes. Three vendors (BioView, Genetix, and Applied Spectral Imaging) offer automated FISH instruments for $100,000 to $200,000. Such instruments capture a single image for the technologist to review by stacking several focal planes to maximize information.
Among amplification techniques, Dr. Hyder noted that real-time PCR is robust, performing amplification and detection simultaneously in a closed system. One high-volume application of amplification techniques is factor V/factor II/MTHFR mutation analysis. Four companies (AutoGenomics, Nanosphere, Roche, and GenMark Dx) make FDA-cleared kits for this purpose. Dr. Hyder uses only mutation analysis for factors V and II because, in his view, MTHFR is not clinically justified.
Potential high-volume tests done by amplification are JAK2 V617F, BCR/ABL1, and MPL W515K and L, used to evaluate clonal myeloid disorders. “You can do these even if you’re not a hematopathologist,“ Dr. Hyder said. “All are relatively easy to perform, but none are FDA cleared, so you must validate them yourself.“
Pharmacogenomics, Dr. Hyder said, “will hit us in a big way.“ PGx assays are used to predict treatment response—the risk of adverse drug reactions or the probable efficacy. For instance, 20 percent to 50 percent of patients don’t respond to antidepressants, and 30 percent to 70 percent don’t respond to statins.
One class of PGx assays measures the levels of allelic variants of cytochrome enzymes that metabolize drugs and thus affect efficacy or toxicity. Examples are tamoxifen for breast cancer, irinotecan for colorectal cancer, ifosfamide for lung cancer, and gefitinib for prostate cancer. Dr. Hyder has brought in CYP2D6 testing to help decide on tamoxifen versus the more expensive aromatase inhibitors.
A potential PGx application is to help determine starting doses for warfarin to optimize efficacy and reduce the risk of bleeding. Several FDA-cleared assays are available; they measure alleles of CYP2C9 and VKORC1.
In addition to measuring patients’ alleles, pharmacogenomic assays look for mutations in tumor genes that affect efficacy of targeted therapies, such as the EGFR inhibitor erlotinib for metastatic colorectal cancer. Mutations in exons 12 and 13 of the KRAS gene predict resistance to EGFR inhibitors. Testing for these mutations is officially recommended.
Dr. Caughron says molecular assays in the area of women’s health is a good place to start. Pathology groups doing cytopathology have a fair number of clients providing women’s health services. It’s a mature area of lab testing with a fair number of FDA-cleared assays. He reviewed three applications: cystic fibrosis screening, human papillomavirus testing, and detection of Chlamydia trachomatis and Neisseria gonorrhoeae (CT/NG).
Cystic fibrosis screening was the first universal genetic testing program, Dr. Caughron noted. It measures the risk of transmitting a CF mutation to a couple’s offspring. More than 1,500 mutations have been reported in the CFTR gene responsible for the disease. The American College of Obstetricians and Gynecologists and the American College of Medical Genetics recommend screening, and reimbursement policies are good. Screens include any mutation that occurs with at least 0.1 percent frequency in the general population. The current basic panel has 23 mutations. Some vendors offer extended panels, Dr. Caughron said, but he is not a fan. “They offer limited additional value for added cost. Also, in some cases we don’t know their clinical significance,“ he explains. Since it’s not possible to predict the extent of clinical disease from mutation results, Dr. Caughron advised, “Be cautious in clinical forecasting.“
Laboratories that perform CF screening should always calculate residual risk after a negative result. Calculating residual risk, which Dr. Caughron calls standard of care, requires obtaining ethnic origin and family history at the time of collecting specimens. In Northern European Caucasians, with a mutation detection rate of 90 percent, a negative result lowers the probability of being a carrier by tenfold, to one in 241. For Hispanic Americans, on the other hand, with a detection rate of 57 percent, a negative result lowers the residual risk by only twofold, to one in 105.
Three FDA-approved CF assays are on the market. With multiple targets, all targets require positive controls. Including positive controls for all targets in every run is not feasible; it is accepted practice to rotate positive controls among all targets. A negative control should be run with every reaction to monitor for contamination.
Testing for high-risk HPV has two indications: first, triage of equivocal Pap results (essentially AS-CUS), and, second, in conjunction with a Pap test for primary screening in women age 30 and over. There are two FDA-cleared tests for this purpose: Qiagen Hybrid Capture 2 (formerly Digene) and Hologic Cervista (formerly Third Wave). Both are approved with ThinPrep but not SurePath, so use with the latter reagent requires limited validation. “Both are good assays,“ Dr. Caughron said. Hybrid Capture “is known to have a problem of cross-reaction with low-risk genotypes if there is high viral load,“ he said. However, it has more extensive clinical validation, since it was approved in 2003, six years before Cervista.
Five companies have approved assays for CT/NG: Roche, Gen-Probe, Abbott (real-time PCR), BD, and Qiagen. Dr. Caughron chose to make an LDT for this purpose, relying on the literature to design the assay. To validate analytical sensitivity and specificity, he compared the LDT to an FDA-approved test. He established clinical validity from the literature.
Dr. Kant closed the symposium by noting additional considerations, such as the importance of clinical utility, which goes beyond clinical validity and is an increasing focus for payers. For instance, the CMS decided to pay for pharmacogenomic testing for warfarin in the context of approved clinical trials. Dr. Kant also discussed economic considerations in decisions about sendout testing, saying the potential savings for bringing a test in-house can be large.
Looking to the future of molecular pathology, Dr. Kant brought up the “$1,000 genome“—sequencing of the whole genome of an individual for $1,000. “This will happen,“ he is certain. “Technically, it is a given.“ But what will we do with all the data? Who will interpret it? “You guys,“ Dr. Kant said, looking at the audience. “This will be a good thing for pathologists. It will be a good opportunity for pathologists going forward to establish a position of consultative expertise.“ And it all starts with that first step—setting up molecular testing now.
William Check is a medical writer in Wilmette, Ill.