September 2008
Feature Story

William Check, PhD

Earlier this year, M. Rajan Mariap­pan, MD, PhD, who had recently become an instructor in pathology at Beth Israel Deaconess Medical Center in Boston, gave a lecture to the hematology / oncology division. “I chose a non-traditional topic,” Dr. Mariappan says. He spoke on the 2008 revision of the diagnostic criteria for myeloproliferative disorders, focusing on the JAK2-positive disorderspolycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF)—by a World Health Organization ad hoc international expert panel. “I intended to speak for 60 minutes,” he says, “but it turned into 90 minutes. The topic generated a lot of excitement. There was a very heated discussion about whether this was a valid classification.”

Dr. Mariappan notes that the conflict was not between pathologists and hematologists but among hematology/oncology physicians themselves. “Some of the data underlying the revision came from these Boston practitioners,” Dr. Mariappan says. “However, there are differences in how they practice. They tend to take into account not only rigorous scientific merits but also cost-effectiveness and ultimately the net benefit to patients.” Some of them felt that the new WHO guidelines for myeloproliferative disorders emphasize increased use of the more expensive molecular-based workups, and because targeted therapy isn’t available yet for most of these disorders, that the classification scheme may be ahead of its time (Samuelson SJ, et al. Blood. 2008;111:1741).

In his own view, the new classification scheme is strong and a step in the right direction. “It is what we hematopathologists/molecular pathologists have been waiting for for a long time. It takes a lot of subjectivity out of the diagnostic workup and provides more objectivity.” It is the first time, he adds, that clinicians, scientists, and pathologists have worked together on a classification of myeloproliferative disease.

Dr. Mariappan is of the view, too, that the discovery of mutations in crucial genes “has made the diagnosis of myeloproliferative disorders almost completely molecular and hence a distinctive group among all diseases.” Still, he acknowledges, the new classification “has some weaknesses.”

Ayalew Tefferi, MD, professor of medicine and hematology at the Mayo Clinic and a member of the WHO ad hoc panel, explains the genesis of the revision: “There have been major discoveries in the molecular pathogenesis of myeloproliferative neoplasms since the third edition of the ‘Blue Book,’ which was published in 2001. These research advances were capped by the discovery of the JAK2V617F mutation and its publication in early 2005, which created the opportunity to refine the WHO histological classification of the myeloproliferative neoplasms and come up with a practical clinical, histological, and molecular diagnostic approach that no longer required red cell mass measurement.”

(The two major genes whose mutations figure in the revision are JAK2—Janus kinase 2, and MPL—the thrombopoietin receptor, which is essential for myelopoiesis in general and megakaryopoiesis in particular.)

Dr. Tefferi enumerates four major changes in the new classification scheme:

• The inclusion of JAK2 and MPL mutations as clonal markers for the diagnosis of PV, ET, or myelofibrosis.



• Minimization of the role of red cell mass measurement for the diagnosis of PV. “According to the new diagnostic criteria,” he says, “red cell mass measurement is almost never required to make a diagnosis of PV.”



• Lowering of the platelet count cutoff level to make a diagnosis of ET from 600 to 450 x 10(9)/L.



• Effective use of histological information in differentiating ET from prefibrotic myelofibrosis.

The new criteria allow diagnosis of PV in an earlier stage, according to Dr. Tefferi.

Already critiques of the revision have been published (Teofili L, et al. Blood. 2007; 110:3384– 3386; Spivak JL, Silver RT. Blood. 2008; 112: 231– 239). In response, Dr. Tefferi says the WHO committee that proposed the revisions for myeloproliferative neoplasms included the most experienced and widely published international authorities on the subject in pathology, clinical medicine, and research, including some of the scientists who discovered the JAK2 and MPL mutations and the clinicians responsible for the majority of the clinical trials work in patients with these diseases. “There will always be ‘I know better’ comments,” he says, “but I believe it is time to move on and find effective treatment for our patients rather than being engaged in fruitless arguments over consensus documents.”

Daniel A. Arber, MD, professor of pathology at Stanford University Medical Center, associate chair of pathology (hematopathology), and interim medical director of clinical laboratory and anatomic pathology services, was a participant in the WHO advisory committee. He says the revisions “really have to do with how to incorporate JAK2 studies into the diagnostic scheme,” which recognizes that these mutations can be useful in the differential diagnosis of chronic myeloproliferative neoplasms. “If there is proliferation, then JAK2 helps to rule out the reactive form,” Dr. Arber says. He notes that MPL, an alternative to JAK2, is mutated much less commonly.

“In my experience, chronic myeloproliferative neoplasms aren’t diagnosed solely by pathologists,” Dr. Arber says. “Morphology alone is not enough to distinguish in many cases. This classification gives better criteria for separating early primary myelofibrosis from ET, but really it depends on a combination of clinical and laboratory findings and morphology. Making this distinction will still be difficult for pathologists.”

Dr. Arber points to European literature describing morphologic differences, which are in the new classification as well. These have to do with location and nuclear features of megakaryocytes. “But these are really fairly minor changes compared with the inclusion of JAK2 mutations,” he says.

Dan Jones, MD, PhD, associate professor at the University of Texas M.D. Anderson Cancer Center in Houston, agrees that the “discovery of JAK2 and MPL were what precipitated the major shift” in the classification. He says, “They added those mutations for diagnosis of PV, ET, and primary myelofibrosis, and while they were revising, they added several other categories, including specific recognition of myeloproliferative neoplasms that have genetic alterations of plateletderived growth factor receptor [PDGFR].” These rare neoplasms, which have chromosomal fusions or translocations involving one of the PDGFR genes, have distinct presentation (often with eosinophilia) and have a good response to kinase inhibitors, favoring their clear distinction from other myeloproliferative neoplasms.

To Dr. Jones, the take-home message of the revision is that WHO is beginning to make the classification both a diagnostic and a therapeutic one. “In the past it has been solely diagnostically oriented. Now, because some of the genetic alterations in these tumors have therapeutic implications, the WHO scheme is beginning to transform from a diagnostic classification into one that encodes other useful clinical information.” However, there is little in the classification as yet about prognosis within specific disease entities, Dr. Jones says. “These neoplasms are generally indolent chronic disorders. But some have a short course while others go on for decades.” The presence or absence of JAK2 or MPL mutations doesn’t fully explain these different courses, so other factors will need to be added in the future if useful prognostic information is to be included.

Dr. Jones acknowledges the many concerns and criticisms regarding inclusion of JAK2 mutation status. “One issue,” he says, “is how specific the finding of a JAK2 mutation is diagnostically, given that it is seen in a range of disease. You can see some minimal polycythemia that have low levels of JAK2 mutation that some clinicians wouldn’t want to classify as a myeloproliferative neoplasm, and also some apparently reactive conditions that have low levels of this mutation. For diagnostic purists, including a genetic event as an essential part of the classification is somewhat problematic.” However, he points out, “The current classification doesn’t force the issue. It doesn’t equate presence of the mutation to a specific disease state. It requires both morphologic criteria and a mutation. So it is probably the right compromise.”

Y. Lynn Wang, MD, PhD, makes the same point. She is associate professor of pathology and laboratory medicine, associate attending pathologist, and director of the molecular hematopathology laboratory at Weill Cornell Medical College/New York-Presbyterian Hospital. “One of the strengths of the new WHO criteria,” she says, “is that they combine histology with molecular changes.” She considers the revision to be “more defined” and “more objective” and believes that the combined criteria will improve diagnostic accuracy.

Before the discovery and introduction of JAK2 mutations, the criteria for polycythemia vera were “soft,” with gray areas, in Dr. Wang’s view. “PV tends to be overdiagnosed in the community,” she says, with diagnosis sometimes relying solely on hemoglobin levels or hematocrit. It is important to distinguish primary from secondary causes of high hematocrit. “If you go to high altitude, you get secondary changes as the body reacts and makes more red blood cells,” Dr. Wang says. People with apnea also have secondary changes with elevated RBCs and high HCT. So it can be hard to distinguish reactive changes from true PV, which she calls an “intrinsic” neoplasm.

To emphasize the value of JAK2 mutations, Dr. Wang says they have been found in “almost every PV case we’ve studied using a sensitive PCR technique” (Chen Q, et al. J Mol Diagn. 2007;9:272–276; Wang YL, et al. Leukemia. 2008; 22: 1289). Some previous studies show lower frequencies, due in part to the use of insufficiently sensitive techniques. At the same time, JAK2 mutations are not found in reactive conditions, according to Dr. Wang, making it a very specific test. So far she has not found any false-positive JAK2 results among the approximately 300 patients without myeloproliferative diseases her group has studied.

She now looks for JAK2V617F as one of her first two tests when PV is suspected, along with erythropoietin levels. Low EPO goes with a JAK2 mutation in myeloproliferative neoplasms because of a feedback loop; in secondary or reactive polycythemia, in contrast, EPO is very high.

“For diagnosis of myeloproliferative neoplasms, the presence of a JAK2 mutation is one of the major criteria in the revised classification,” says Paul G. Rothberg, PhD, professor of pathology and laboratory medicine and director of molecular diagnostics at the University of Rochester Medical Center in New York. Only one minor criterion is needed. “Under the revised WHO standard, probably the most likely one, if you already have evidence for polycythemia, will be reduced EPO.” An alternative is analysis of bone marrow.

Roger D. Klein, MD, JD, calls the revision an important advance “in terms of placing the diagnosis of a category of hematopoietic malignancies on a more sound biological basis, as well as the use of a molecular marker to add objectivity to the diagnosis.” Dr. Klein, who is medical director of molecular diagnostics in the Division of Hematopathology and Laboratory Medicine at the H. Lee Moffitt Cancer Center and Research Institute in Tampa, says also that the new criteria “are exciting in what they portend for the potential development and implementation of targeted therapy.”

Like others, Dr. Klein cautions that a JAK2 mutation makes it possible to place a disorder within a new category of myeloproliferative neoplasms, but not to differentiate between PV, ET, and PMF. “We must still rely on clinical and traditional pathological criteria to separate those entities,” he says. Even so, it moves the diagnosis of PV from one of exclusion to one that is more an affirmative diagnosis, he says, though exclusion criteria remain integral to the diagnosis of ET and PMF.

Dr. Klein has a concern apart from clinical issues, and that is intellectual property. A patent is pending on the correlation between the presence of a JAK2 mutation and myeloproliferative diseases. “This is one of the early markers,” he says, “but it will happen that large numbers of genetic markers will help us define clinical entities and form the basis of therapeutic interventions. The idea that one person or institution could control the relationship between a marker and a disease phenotype is problematic and likely to encumber our ability to optimally perform testing for these markers. It is also likely to increase the cost of testing.”

All of the specialists interviewed for this story have already instituted testing for JAK2 mutations or are in the process of doing so. Dr. Mariappan notes that the discovery of JAK2 mutations has been translated into clinical use in just a few years, in contrast to one of the first oncogenic mutations, the serine tyrosine kinase associated with the BCR-ABL translocation in chronic myelogenous leukemia, which is targeted by imatinib and related small molecules. “It is a pretty simple test,” he says. “It doesn’t need reference lab expertise. A good molecular diagnostic lab can do it in house. Many midsize hospital labs are doing it.”

Dr. Mariappan raises the possibility that a panel approach, including MPL and other JAK2 mutations in addition to V617F, may be helpful. In a subset of patients with ET the MPLW515L mutation is found, but only in a small subset of cells. It is an activating mutation of the thrombopoietin receptor that causes uncontrolled proliferation of megakaryocytes.

Dr. Mariappan came to Boston from a fellowship at Stanford, which he calls a “forerunner” in recognizing the importance of the JAK2 mutation and instituting testing for it. “They see a lot of chronic myeloproliferative disorders there and the volume of testing is very high,” he says. Dr. Arber of Stanford says they test for JAK2 mutations by PCR but do not do other mutations. Some PDGFR mutations are translocated, which can be picked up on routine karyotypes or by FISH analysis. “Right now, if we suspect a translocation specifically, we test for the most common abnormalities and send the others out,” Dr. Arber says. He agrees that a JAK2 mutation can be found in at least 90 percent of true PV cases.

At M.D. Anderson, Dr. Jones’ laboratory performs a quantitative assay for the common JAK2 mutation site and has been doing it for several years. He chose a quantitative assay because of the initiation of clinical trials for small-molecule inhibitors of JAK2. “We want to be able to monitor a patient’s course, analogous to doing BCR-ABL PCR in CML.” Dr. Jones is concerned that some of the current JAK2 molecular assays are very sensitive and some are less so. “So if a lab is reporting a qualitative, highly sensitive assay, it may find JAK2 mutations that are clinically irrelevant and push the diagnosis into one disease category, whereas another lab with a less-sensitive assay may not detect a mutation and so end up with a different diagnosis,” he says. Commercial suppliers currently offer both qualitative and quantitative assays, he notes, “because the hematopathology community itself has not decided whether quantitative measurement is necessary.”

Some papers have strongly recommended quantitative testing, Dr. Mariappan points out. “But the WHO document does not suggest quantitative testing, and I agree with this for now,” he says. “Negative or positive is more important than quantitative testing, especially since targeted therapy is not yet available. Once we have specific inhibitors, quantitative measurement may become more important.” He says a group from Quest Diagnostics is promoting a plasma-based quantitative assay. “Dr. [Maher] Albitar asserts that quantitative testing in plasma may characterize various myeloproliferative disorders as well as give prognostic information,” Dr. Mariappan says, referring to Quest’s medical director of hematopathology (Ma W, et al. Blood. 2008;111: 3906– 3907). This issue is still open, in his view.

Dr. Wang’s laboratory at Weill Cornell/New York-Presbyterian does an in-house qualitative assay. “There has been no clear demonstration for the clinical utility of quantitative testing,” she says, verifying Dr. Jones’ contention that experts in the field do not yet agree.

“Molecular has its own problems,” she says. “It depends on which method you use. And there is no standardization. WHO didn’t define the assay parameters.” Agreeing with Dr. Jones, she says, “Certain tests are either overly or insufficiently sensitive.” For JAK2 molecular testing, she has found false-negative results to be of particular concern. “True PV can still have a mutation rate of less than five percent,” Dr. Wang says, referring to the fractions of cells carrying the mutation. Her group compared the results of 33 PV cases using assays of different sensitivities and found assays of 0.1 percent sensitivity to be adequate (Wang YL, et al. Leukemia. 2008;22:1289). She has seen reports from other labs reporting five to 10 percent sensitivity. In such cases, after consulting with her clinicians about patients’ clinical information, Dr. Wang may advise that the specimens be retested to see if they are truly negative.

At the University of Rochester Medical Center, the hematologists order molecular testing whenever there is a suspicion of a myeloproliferative disorder, Dr. Rothberg says. “They don’t just require a strong suspicion, but any suspicion in a patient with a high hematocrit or high platelet level. So it is a pretty frequently done test. We’re probably doing 30 to 50 per month.” Dr. Rothberg says this mutation was discovered and moved to clinical use faster than he has ever seen before with a molecular assay. “It was less than one year from the first Internet report to adoption by clinical labs.” He attributes this rapid adoption to the test having filled a need for better diagnostic tools for myeloproliferative disorders. “It is very difficult to diagnose polycythemia vera,” he says, “so a molecular test has a lot of value. Before it often took months to say definitively whether the patient had a neoplasm.” For instance, in a smoker with a high hematocrit it might have been necessary to do extensive testing. “Previously a definitive diagnosis might depend on measurement of the red cell mass, which is difficult, expensive, and available only at major medical centers. And the patient must be present; you can’t send it out.” But with the availability of mutation testing, a definitive diagnosis can be reached rapidly in most polycythemia patients. Dr. Rothberg foresees red cell mass measurement being done much less frequently, and possibly not at all.

As a step toward standardization, Drs. Rothberg and Wang are part of an eight-laboratory Centers for Disease Control and Prevention/ Association for Molecular Pathology working group that is doing a JAK2 sample exchange. “We are looking at comparative analytic sensitivity and seeing whether we can agree on positives and high-level positives,” he says. The eight laboratories do several different types of assays.

“There is a continual flow of data coming out on this topic,” Dr. Klein says.

Incorporating mutation analysis for JAK2 into the workup of possible myeloproliferative neoplasms not only improves diagnosis but also might make the workup more cost-effective, in Dr. Mariappan’s estimation. “Once the mutation is positive, the clinician can probably stop the workup,” he says. “Some hematologists feel that any genetic-based testing may end up as an expensive workup. I answer: If you follow the WHO-based algorithm and do a JAK2 test and EPO level to work up a suspected patient with polycythemia vera, the estimated cost is $600 combined. The advantage is you made a definitive diagnosis in about 98 percent of patients, answering lots of clinical questions and avoiding additional workup.” Some of the older assays, he says, such as red cell mass measurements, cannot be done reliably under current CLIA guidelines, and other tests such as erythroid colony formation are more expensive, time-consuming, and laborious, and are not offered widely. “The algorithm also suggests that in a small proportion of cases you might avoid a bone marrow biopsy.” Dr. Mariappan compares the myeloproliferative neoplasms situation to that of chronic myelogenous leukemia. In both, a baseline biopsy is recommended. However, he notes, the number of bone marrow biopsies for patients who are followed up with imatinib therapy has fallen significantly. “Clinicians started following CML patients on Gleevec by peripheral blood,” he says, “less frequently doing bone marrow biopsy.”

Moreover, Dr. Mariappan says, JAK2 mutation analysis can be done on peripheral blood, so in some cases the clinician doesn’t even get a baseline bone marrow biopsy. “I’m seeing a trend of not doing what has been routine in the past,” he says. In fact, he notes, the decrease in bone marrow biopsies is causing “a little consternation” among pathologists. “Pathologists are basically good morphologists,” he says. “Pathologists prefer a tissue-based diagnosis in conjunction with molecular-based testing.” Right now, Dr. Mariappan explains, when a definitive diagnosis of polycythemia vera is made with the revised WHO criteria, if the patient is doing well, the clinician will probably not repeat a bone marrow biopsy until there is clinical suspicion of progression. Previously, because the diagnosis was somewhat presumptive, bone marrows were done at frequent intervals to assess morphology and follow progression.

However, bone marrow biopsy is valuable for distinguishing among the various myeloproliferative neoplasms and for monitoring disease progression. Currently, doing JAK2 mutation testing alone cannot reveal progression. “Progression has to be based on a good bone marrow biopsy and clinical parameters,” Dr. Mariappan says. “We may have to work out followup intervals and an acceptable bone marrow biopsy interval.”

Dr. Rothberg agrees that bone marrow biopsy numbers are likely to decrease. “Everyone is sort of agreeing that you don’t have to do a bone marrow if there is a JAK2 mutation,” he says. If the mutation is absent, you still might have myeloproliferative disease, especially for ET and PMF, where JAK2 mutations are not as common. “In those cases, bone marrow biopsy is still very important,” he says. Biopsy might also be useful where the clinician is concerned about progression to myelofibrosis or leukemia, he agrees. “It will still be done pretty frequently.”

Probably the main reason that mutation testing was adopted so quickly for myeloproliferative neoplasms is the example of CML, where an underlying mutation is now targeted by specific chemotherapy. However, at this time there is not yet an imatinib-like drug for polycythemia vera.

There have been trials of small-molecule inhibitors of JAK2. But it is not yet clear whether JAK2 is an essential part of the myeloproliferative disease process. “It may end up being a surrogate diagnostic marker, but its exact role in the pathogenesis of the disease is still an unanswered question,” Dr. Mariappan says. He does not think that lack of an inhibitor negates the value of JAK2 testing. “At the end, giving an accurate diagnosis is the job of pathologists and clinicians,” he says. At the least, once a specific diagnosis is made, some patients become candidates for interferon therapy, he notes, which can be helpful.

Dr. Jones says many small-molecule inhibitors of JAK2 are being developed now. “We expect that soon one with therapeutic effect will arise. At that point, you would definitely want to know the JAK2 status of a given myeloproliferative tumor before starting therapy.” The experience at M.D. Anderson suggests that cancers with overexpression or dysregulation of a specific kinase can eventually be successfully targeted, he says. But two criteria must be fulfilled. First, the inhibitor with the best kinase target range and the lowest toxicity profile must be found. Second, “the molecule you are targeting has to be very important to driving the disease process,” Dr. Jones says. Both conditions were met in CML and thus targeted therapy was successful. “With JAK2 mutation,” on the other hand, “it is not clear that it is the only thing driving the disease condition,” he says. “So even if we get a successful inhibitor, that may not be equivalent to therapeutic control.”

Dr. Wang shares Dr. Jones’ view, and feels even more strongly. “JAK2 is not a disease-causing mutation,” she says. “Through the last few years of study that has become clearer.” She bases her conclusion on the fact that, unlike other tumor markers that are present in almost every single malignant cell, only a fraction of malignant cells in PV patients harbor the JAK2 mutation(s).

Dr. Tefferi headed a trial of an inhibitor at Mayo Clinic in patients with “huge” spleens—one of the clinical signs of disease. Dr. Wang says, “Patients sometimes felt a lot better and had smaller spleens. But JAK2 burden didn’t come down concomitantly with the shrunken spleens, so the inhibitor had a beneficial clinical effect but perhaps not through JAK2.” Various biotechnology and pharmaceutical companies have developed several JAK2 inhibitors, a few of which have already entered early-phase clinical trials. “This is an area of active research—the final verdict is still pending,” Dr. Wang says.

To draw conclusions at this point about whether a small-molecule inhibitor of JAK2 will emerge would be premature, Dr. Tefferi says. “Give us another two or three years to understand the precise pathogenetic contribution of these mutations and whether or not they are useful drug targets.”

While not as revolutionary as the introduction of JAK2 mutation testing, the change in platelet count for diagnosing essential thrombocythemia from 600 to 450 x 10(9)/L “will lead to a whole lot of clinical change,” Dr. Mariappan says. “Some clinicians think it is not necessary and may lead to increased testing and may pick up a lot of patients with subclinical disease for whom we don’t have a treatment.”

Dr. Arber suggests that the change in the criterion for platelet count will make it even more difficult for pathologists to sort out disorders. “But it is really important for people to understand that platelets are not the sole defining criterion for separating myelofibrosis from essential thrombocythemia.” In the past, he says, platelet count was used as a more rigid criterion, and some cases were probably misclassified in both directions. “If it was greater than 600, we assumed it was ET when it could be primary myelofibrosis, and sometimes you can have ET at a lower count.”

Dr. Wang says oncologists have mixed feelings about whether the platelet count criterion should be lowered. To some extent, the lowering of the threshold is justified by the presence of a JAK2 mutation in some cases of essential thrombocythemia.

Will the change in platelet count result in overdiagnosis? “Not if you are a good clinician,” Dr. Tefferi says. “If one reads the WHO document carefully, it requires either the presence of a molecular marker or bone marrow histological features to make a diagnosis of ET. In addition, the criteria specifically underscore the need to rule out reactive causes of thrombocytosis.”

In summary, one might say: If you have traditional criteria plus a JAK2 mutation, that’s a strong indicator for PV diagnosis; if you have a mutation without traditional criteria, be cautious.

“Keep in mind,” Dr. Arber says, “that as we define more genetic changes, we will move toward a genetic classification. In the future, rather than putting diagnoses into buckets, rather than using labels like PV or ET, for example, we may refer to ‘JAK2-mutated myeloproliferative neoplasms.’” Such labels are now “a bit premature,” Dr. Arber concedes. But it might happen in the near future. “Certainly if we get drugs to target these genetic abnormalities,” he concludes, “it would be more logical to classify them by how they should be treated.”