A 41-Year-Old Woman With Chronic Myelogenous Leukemia

DR BURNS: Mrs P is a 41-year-old woman with chronic myelogenous leukemia (CML). She previously worked as a hairdresser but has not worked for several months. Mrs P lives with her husband and children. She has commercial insurance.

For a number of years, Mrs P noted that she felt fatigued and had headaches. She sought medical care, but no explanation was found. In June 2002, she went to a new primary care physician and was found to have a white blood cell count of 58 × 10³/µL, hematocrit of 41.4%, and a platelet count of 211 × 10³/µL. There were 51 polys, 2 monocytes, 1 eosinophil, 1 promyelocyte, 13 myelocytes, 14 bands, and 5 metamyelocytes. A bone marrow biopsy was performed and she was found to have a myeloproliferative disorder; cytogenetic testing showed classic Philadelphia chromosome (Ph).

Mrs P was seen at another medical center and imatinib mesylate (Gleevec; Novartis, East Hanover, NJ) was recommended. After starting imatinib (400 mg/d), she developed severe occipital headaches and an abdominal rash. The drug was stopped and her symptoms improved; however, her white blood cell count rose to higher than 120 × 10³/µL. She was subsequently restarted on a lower dose of imatinib (200 mg/d) with a plan to slowly increase the dose to 400 mg/d. She tolerated this regimen better, although she continued to note ongoing occipital headaches and fatigue. Her white blood cell count improved to 88 × 10³/µL.

Mrs P has a past medical history significant for kidney stones and migraines and a past surgical history significant for an appendectomy and hysterectomy. Her only current medication is imatinib mesylate (400 mg/d). She is allergic to penicillin and sulfa. She quit smoking in October 1994, and she does not drink alcohol.

Mrs P recently sought a second opinion and an allogeneic transplantation was recommended. She decided to proceed with the transplant as she believed that it offered a reasonably good chance for cure and because she felt that she could neither afford imatinib for the long-term nor tolerate its adverse effects. Mrs P and her siblings underwent histocompatibility testing and were not a match. She has considered her options and is planning a matched, unrelated donor stem cell transplantation.

I first noticed that things were different for me a few years ago. That's when I started getting very bad migraines and was just not feeling myself. The doctors kept looking at me like I was, you know, just saying things were wrong, and they tried making me out to be a hypochondriac or something.

I was [feeling] very tired. They thought that my tiredness was due to the headaches, they checked me for sleep apnea, and they put me on a lot of different medications that helped me sleep so I wouldn't be so tired and get the headaches. But they didn't help.

My new doctor checked out the blood cells and found that my white blood count was high, and he found out that I had a positive chromosome level. So then, we went on to a bone marrow biopsy, and it confirmed that I had CML.

I was very scared when I found out exactly what CML was because I had no idea how my previous doctor did not catch it. I didn't know how far along in my body the CML was. So that made me more nervous than anything.

Gleevec [imatinib] is an option, but it's very expensive. Insurance would pay only 60% of it—I would have to come up with 40%, which is more than $800 a month. I could stay on medicine the rest of my life to keep my numbers [down], but that would only keep them suppressed. It would never cure the problem. I would always have it. And due to insurance, it's very expensive, and I really couldn't afford all that medication. The only cure was the bone marrow transplant, and since I am young—I'm only 41 years old, and I've got a long life ahead of me—I'd like to see my kids grow up. There's a little chance that I could go into relapse, but there's a better chance that the bone marrow transplant will actually cure me. My insurance is paying for the full bone marrow transplant this time.

I told the doctors that when they found a donor that no matter what it took I was ready to go. I didn't want to wait. I wanted that chance to have my extra life now, not later. And the sooner I got this going the sooner I would heal from it.

What is the epidemiology of CML? How is CML defined and classified? How does CML usually present? How is the diagnosis established? What is the natural history and prognosis for patients with a new diagnosis of CML? What are the current treatment options and the risks and potential benefits of each? What are the long-term outcomes of imatinib and bone marrow transplant? What does the future hold for treatment? What do you recommend for Mrs P?

DR ANTIN: Mrs P has a dilemma shared by other young people with CML. The brief interval between the initial trials of imatinib and its clinical availability have outstripped our understanding of the long-term effects of the drug. The substantial coverage in the media and medical journals has established imatinib as an important and desirable therapy for CML, yet its relative newness does not allow accurate comparison with stem cell transplantation. Mrs P needs to decide whether to commit herself for months or years to a new agent (with unknown long-term benefits) that would cost her $800 per month out of pocket (approximately 40% of the retail cost of the drug) or to undergo stem cell transplantation, a dangerous therapy known to cure CML.

Chronic myelogenous leukemia is an uncommon disease. It affects all races equally with an incidence of approximately 1 per 100 000. Men are twice as likely to be affected as women.¹ There are no geographic or exposure associations, with the possible exception that there appears to be an increased risk of CML among nuclear weapon survivors.² It is less common than chronic lymphocytic leukemia, acute myelogenous leukemia, and myelodysplastic syndrome.¹ Like most hematologic malignancies, the incidence increases with age. The median age at onset is 67 years, and it is distinctly uncommon in children.¹

Chronic myelogenous leukemia is an indolent illness; approximately half of patients are asymptomatic at diagnosis and the rest have nonspecific symptoms and signs that provide few explicit clues. Often, the diagnosis is suspected based on the complete blood cell count obtained at a routine annual checkup or during an evaluation for an unrelated disorder. Fatigue is the most common manifestation, occurring in 80% of patients.³ Mrs P had fatigue and headaches, neither of which was apparently enough to trigger her previous primary care physician to perform a blood cell count. Other common symptoms are weight loss (60%), abdominal discomfort (40%), and easy bruising (35%) despite normal or high platelet counts.³ Leukostasis, priapism, or thrombosis may occur but are rare.^{3 - 4} Leukocytosis is a characteristic finding of the disease; it is higher than 100 000 cells/µL in about 30% of patients, and a left shift is uniformly observed.³ Basophilia is common and should always raise the suspicion of CML in a patient with an elevated white blood cell count, as basophilia does not occur as a normal consequence of infection. Mild anemia (hemoglobin <12 g/dL) is common (65%), and the platelet count is usually normal or high normal. High platelet counts (>700 × 10³/µL) are seen in 25% of patients and low platelet counts (<150 × 10³/µL) are seen in only 5% of patients.³ Low platelet counts may suggest more advanced disease.

The hallmark of CML is a balanced translocation between chromosome 9 and chromosome 22 [t(9;22)] first observed in the 1960s by Nowell and Hungerford⁵ and first recognized to be a translocation by Rowley⁶ (Figure 1). It was called the Philadelphia chromosome (Ph). The translocation links the proto-oncogene c-abl to BCR (breakpoint cluster region). An 8.5-kb mRNA is transcribed into a hybrid P210 bcr-abl protein (less commonly P190 or P230). BCR-ABL is required for leukemic transformation. ABL is a tyrosine kinase (TK) with transforming ability. The BCR motif results in tetramerization of the ABL, which binds to F-actin (required for activity) and results in resistance to apoptosis, cell growth, reduced adhesion to stromal tissue (loss of contact inhibition), and accumulation of cells in the blood. It has been recognized since 1990 that the human BCR-ABL fusion gene can induce a myeloproliferative disorder in mice that resembles CML.⁷ Interestingly, the cells are not growth factor independent. A simplified view of the distinction between CML and acute leukemia is that there is a proliferative thrust due to the Ph chromosome, but the cells are able to mature. Thus, they are capable of transporting oxygen, fighting infection, and the like.

It is quite possible that Mrs P had the disease for months or even years before the diagnosis was established. The ability of the cells to fully mature and function allows the maintenance of a reasonably normal life in the face of very active disease. Weight loss and fatigue are probably due to the metabolic demands of maintaining a marrow compartment that may be 10-fold expanded. In acute leukemia, a failure of cell maturation results in leukopenia, anemia, thrombocytopenia, and their respective complications. Late in the course of CML, new mutations may occur that prevent differentiation and ultimately result in an acute or blast phase that looks much like acute myelogenous leukemia, although it is much harder to treat.

Typically, the bone marrow is hypercellular, but the diagnosis is established by finding the t(9;22) translocation on a metaphase spread of bone marrow cells. Other techniques such as polymerase chain reaction (PCR) and fluorescence in situ hybridization (FISH) can also be used to make a diagnosis, but those techniques are unable to detect additional cytogenetic abnormalities. Failure to detect other chromosome changes is a significant problem since the more advanced stages of the disease are often associated with the acquisition of new cytogenetic abnormalities that will not be apparent with FISH or PCR.⁸

The disease has 3 phases. At diagnosis, 85% to 90% of patients are in the chronic phase,³ which is the most indolent phase of the disease and the most amenable to therapy. Most patients with symptoms are in the stable phase, and symptoms will tend to resolve with control of blood cell counts. Mrs P noted that her headaches and at least some of her fatigue were ameliorated as the leukemia burden diminished in response to therapy. In time, some patients will pass through an accelerated phase associated with weight loss, fever, bone pain, extramedullary disease, increased drug requirements, increasing blasts, increasing basophils, anemia or thrombocytopenia, marrow fibrosis, and additional chromosomal abnormalities (eg, second Ph chromosome, iso [17]).⁹ At this point, higher doses or new agents are required to control leukocyte counts. Ten percent to 15% of patients present in accelerated phase or blast crisis,³ which carries a poor prognosis. Accelerated phase eventually evolves into blast transformation, characterized by a rapid course that is fatal without stem cell transplantation. The blasts are myeloid about two thirds of the time and lymphoid one third of the time, as assessed by cell surface markers. Blast phase is an ominous event with a median survival of 3 to 6 months.¹⁰

Several systems to estimate prognosis have been established. A complex calculation based on laboratory and physical findings at diagnosis, such as spleen size, white blood cell count, platelet count, basophil count, and others, allows the assignment of patients to good, intermediate, and poor prognostic groups,^{11 - 12} and this calculation would place Ms P in the good group. Based on conventional therapy (untreated or hydroxyurea), the good prognosis group has a median survival of 8 years, while the median survival of the poor prognosis group is only 3 years. The intermediate group has a median survival of 5 years, which also is the overall median survival for CML.¹³ Typically, patients die of CML because of development of blast crisis. Particularly for patients in intermediate phase or blast crisis, discussion of advance directives is appropriate.

Routine clinical care of patients with CML requires monitoring of peripheral blood cell counts on a frequent basis, although no clinical studies have established exactly what that frequency should be.¹⁴ Early in therapy, blood cell counts may need to be obtained weekly to titrate drug dose, but as the disease comes under control, the frequency can be reduced to monthly, every 3 months, or biannually. If the patient's blood cell count normalizes, symptoms resolve, and the spleen shrinks, the patient is considered to be in hematologic remission. As shown in Figure 2, this may be achieved with all of the effective therapies.^{15 - 21} In many patients, routine cytogenetic analysis will continue to demonstrate t(9;22). If the proportion of abnormal cells declines from its baseline frequency (usually 75%-100%) to more than 35% more than of all metaphases, it is considered a major cytogenetic response.^{22 - 25} A complete cytogenetic response requires no apparent Ph-positive metaphases. However, cytogenetics will detect Ph-positive metaphases only when 3% to 5% of all cells are affected.¹⁴ The value of monitoring cytogenetics resides in its lack of specificity, ie, it will identify the acquisition of new chromosomal abnormalities other than t(9;22).⁸ If a more sensitive assessment of residual disease is desired, marrow or blood can be assessed using FISH on either interphase or metaphase cells. FISH will detect masked translocations that are not apparent on cytogenetic analysis and, depending on the technique, it can detect t(9;22) that appear in as few as 0.1% to 1% of cells. The most sensitive technique is reverse transcriptase polymerase chain reaction (RT-PCR) for the BCR-ABL mRNA, which can detect t(9;22) appearing in only 10^-4 to 10^-6 cells. A complete molecular response generally indicates undetectable BCR-ABL mRNA by a PCR technique. However, as noted in Figure 2, even a complete molecular response is not tantamount to cure. A slightly less sensitive variant of this assay—real-time quantitative PCR—allows a quantitative assessment of BCR-ABL mRNA and thus can be used to monitor responding patients for relapse before it becomes clinically apparent.¹⁶ As noted, the disadvantage of FISH and PCR is that they do not detect new chromosomal abnormalities. Thus, my recommendation is to assess cytogenetics every 6 months. For those patients who have no apparent Ph-positive metaphases, monitoring quantitative PCR may allow early assessment of treatment failure.

Hydroxyurea, Interferon, and Busulfan. For Mrs P and patients like her, the principal challenge in the management of CML is selection of therapy that either cures the disease or prolongs survival. Busulfan is no longer used in the management of CML, since randomized trials demonstrated its inferiority to hydroxyurea.^{12 ,26 - 27} Hydroxyurea is an inexpensive oral agent with few adverse effects. It rapidly controls blood cell counts, but it induces remission in only 1% to 4% of patients even when used intensively,^{18 ,22} and thus is rarely used alone. It is now primarily useful to lower blood cell counts rapidly before introducing another therapy. Interferon alfa is a remission-inducing agent, although still only a minority of patients respond.^{22 ,28} In approximately 20% of patients, the percentage of Ph-positive metaphases in the marrow declines to less than 35%.^{22 - 24} Individuals who have this response have a median survival of 7 to 10 years compared with 3 to 5 years in nonresponders.²² The 3- to 5-year median survival is essentially identical to what is observed in patients taking busulfan or hydroxyurea.²² Interestingly, a small cohort of patients with complete cytogenetic responses may even clear the Ph chromosome by PCR. These patients may have very long survivals,²⁹ suggesting that at least in some patients, interferon may suppress the leukemic stem cell.³⁰ However, interferon does not appear to cure the disease in most people.²⁹ When cytosine arabinoside (Ara-c) is added to interferon,²⁴ the response rate may increase, but the effect on survival is controversial and myelosuppression toxicity increases.²³ Interferon has substantial disadvantages. It is expensive and toxicity includes flu-like syndrome, insomnia, autoimmune manifestations, depression, alopecia, and neurotoxicity. Due to toxicity, 20% of patients cannot tolerate interferon.^{22 - 24} In addition, improvement in survival is seen primarily in low- or intermediate-risk patients, whereas high-risk patients have an outcome similar to those taking hydroxyurea. Figure 2 shows an approximation of the relative ability of these treatments to control the disease. It is clear that the residual tumor burdens are usually quite high.

Imatinib. Imatinib mesylate is a competitive inhibitor of several cytoplasmic TKs, including the BCR-ABL fusion peptide. Because the fusion TK appeared to be sufficient to cause a CML-like myeloproliferative disorder, investigators tested TK inhibitors for their specific ability to inhibit cellular proliferation based on excessive TK activity, and imatinib was identified in this way.^{31 - 34} Imatinib was the first rationally designed therapy to target a specific genetic change in leukemia and the second agent to specifically target a known mutated gene responsible for leukemia (the first was all-trans retinoic acid for promyelocytic leukemia).³⁵

Clinical studies demonstrated unequivocally that imatinib could induce rapid hematologic remissions.²⁵ For patients with accelerated or blast phase disease, 20% to 40% will have hematologic improvement and 10% to 20% may have a major cytogenetic response.^{36 - 37} In contrast, complete hematologic responses were seen in 95% of patients with stable phase disease. Hematologic control is similar in all 3 risk groups, although, as expected, patients with early disease appear to do somewhat better. Suppression of the Ph chromosome occurs in up to 70% of patients; however, 95% of patients continue to have demonstrable BCR-ABL mRNA by PCR.²¹ A randomized trial in 1106 patients (IRIS [International Randomized IFN vs STI571 Study Group]) was performed comparing imatinib (n = 553) with the combination of interferon and cytarabine (n = 553) in patients with early, previously untreated, stable-phase CML. This study demonstrated a striking advantage in complete cytogenetic responses (67.8% vs 7.4%) and 1-year progression-free survival (97.2% vs 80.3%). The likelihood of entering blast crisis was low in both groups in the first year after diagnosis. As yet it is too early to determine whether imatinib improves survival.²⁵ However, as depicted in the Figure 1, even an excellent drug like imatinib is probably far from eradicating the disease.

Imatinib is a well-tolerated oral agent and most patients are able to take it long term with manageable adverse effects, but nausea, vomiting, diarrhea, fluid retention, edema, and muscle cramps occur in 30% to 50% of patients.^{25 ,38} About 5% of patients have trouble tolerating this agent, and Mrs P found the headaches, nausea, and persistent fatigue to be too debilitating. This prevalence of adverse effects requiring drug discontinuation contrasts with toxicity from interferon alfa, which requires discontinuation of the drug in as many as 20% of patients.^{22 - 24} Since the median age at onset of the disease is in the seventh decade, most patients with CML are not candidates for transplantation, and imatinib would seem to be the most effective and least toxic alternative. Unfortunately, imatinib is expensive—a serious consideration for anyone like Mrs P who may not be able to afford the drug and a particular concern for older patients who may be most likely to benefit from it. Some of this problem may be partially offset by a prescription assistance plan from the manufacturer providing discounts based on financial need.

There are several concerns and limitations in the use of imatinib. As noted above, 95% of patients continue to have detectable BCR-ABL transcripts while receiving imatinib therapy, and there is no evidence that this proportion is likely to change with longer follow-up.²¹ It appears that committed progenitors but not stem cells are inhibited by imatinib, suggesting that there will be a continued pool of leukemic stem cells.^{19 ,39} If the genetic defects were stable, the continued presence might be of little concern; however, new point mutations in the active site of the TK do occur, endowing the cells with resistance to TK inhibition.^{37 ,40 - 43} Furthermore, some patients have as much as 20-fold amplification of BCR-ABL, which reduces sensitivity to the drug.^{40 ,42} Finally, new cytogenetic abnormalities may occur, sometimes in cells that do not even contain the Ph chromosome.^{20 ,44}

Stem Cell Transplantation. Stem cell transplantation is the only therapy known to cure CML, but with the potential for significant morbidity and mortality. It also can be extremely expensive.⁴⁵ The development of transplantation as curative therapy for CML has evolved dramatically since the early 1970s. Stem cell transplantation was initially used as a rescue or antidote for the administration of what was otherwise lethal chemoradiotherapy (conditioning regimen). The conditioning regimen was designed to be maximally toxic to the marrow with less intense (but appreciable) damage administered to nonhematopoietic organs. The marrow allowed reconstitution of lymphohematopoiesis after the leukemia was destroyed, along with any normal lymphopoiesis and hematopoiesis.⁴⁶ Subsequent retrospective observations strongly suggested that the graft did more than replace hematopoiesis—there was a measurable graft-vs-leukemia (GVL) effect,^{47 - 48} in which the T cells of the donor recognize host antigens on the leukemic cells and suppress or destroy the disease. Ultimately, T cells alone were shown to be able to eradicate the leukemia in patients who had relapsed after transplantation.⁴⁹ Hematopoeitic stem cell transplantation can reduce the leukemic cell burden from 10¹² cells at diagnosis to less than the lower limit of detectability of the BCR-ABL mRNA by nested primer RT-PCR (<10⁶ cells),⁵⁰ and many patients receiving donor lymphocyte infusions appear to be cured (Figure 2).The next logical step was to administer reduced-intensity chemoradiotherapy simply to prevent rejection of the graft and to infuse large numbers of donor T cells plus stem cells to maximize the GVL response. The stem cells are required for hematopoietic recovery since the recipient's hematopoiesis is largely leukemic, and it is destroyed by the lymphocytes in the graft. This procedure is called nonmyeloablative transplantation. Because the intensity of the regimen is lower, there is substantially less transplant-related toxicity, and the procedure is applicable to older patients and patients with underlying medical problems. Thus, 2 transplantation procedures are commonly used: (1) stem cell transplantion following conventional high-dose drug therapy and (2) reduced intensity, or nonmyeloablative, transplants. Moreover, the graft can consist of cells collected directly from the marrow cavity, cells collected by leukapheresis after stimulation with hematopoietic growth factors (peripheral blood stem cells), or umbilical cord blood cells. The cell source can be a family member or an unrelated donor, and the donor can be fully matched or a partial mismatch. However, this therapy is new enough that long-term data are lacking, and the total number of patients treated is relatively small.^{51 - 52} Furthermore, the variety of transplantation techniques makes a simple analysis of outcomes difficult.

Young patients (defined as <50 years of age) like Mrs P with fully matched donors who receive a transplant during stable phase using conventional high-dose therapy generally do well with 5-year disease-free survivals of 60% to 80%, depending on the series.^{53 ,56} Many of these remissions appear to be permanent.⁵⁴ This outcome ultimately attracted Mrs P. Patients older than 50 years who receive transplants have greater morbidity and mortality.⁵³ Unrelated donor transplantation outcomes are asymptotically approaching those of matched sibling donor transplantation as HLA-typing technology has improved. With allele-level matching, outcomes of 50% to 70% 5-year disease-free survival can be achieved. Again, older patients have proportionately worse outcomes.⁵³ Failure of allogeneic transplantation is usually due to conditioning regimen-related toxicity, opportunistic infections, and graft-vs-host disease (GVHD), with rates varying by center.⁴⁶ Neither relapse nor graft rejection is a major concern in this illness. In contrast to acute leukemia, in which a remission is generally desirable before transplantation, marrow grafting has been performed while the patient with CML had active disease. The incidence of GVHD can be substantially diminished by removing the T cells from the stem cell product prior to infusion; however, this almost invariably results in a loss of GVL. Newer T-cell depletion strategies attempt to maintain the benefits of T-cell depletion while reducing the risk of relapse. A recent example involves depleting T cells at the outset and restoring GVL by adding a donor lymphocyte infusion several months after transplantation.⁵⁵ Interferon use prior to transplantation appears to influence neither the toxicity nor the relapse rate of matched sibling or unrelated donor transplantation.⁵⁶

Mrs P felt more comfortable with conventional dose than nonmyeloablative transplantation because more of these transplants have been performed and the follow-up is more mature. However, at least one study of nonmyeloablative transplantation reported an actuarial overall and disease-free survival at 5 years of 85% (95% confidence interval, 70%-100%).⁵⁷ The incidence of acute and chronic GVHD after reduced-intensity transplantation is similar to the incidence observed in conventional dose conditioning and requires a skilled team for prophylaxis and therapy for transplant-related complications. The role of nonmyeloablative transplantation in the overall management of this disease is not defined and should be considered part of an investigational approach. For those patients who relapse after transplantation, infusion of donor lymphocytes is an effective salvage therapy, resulting in 5-year disease-free survivals of 60% to 80%.^{50 ,58} The use of imatinib in patients who have relapsed after transplantation is also promising.⁵⁹

The introduction of imatinib has transformed care for patients with CML. A relatively nontoxic, rationally developed therapy, it controls the disease in most patients. It has superceded interferon therapy, despite interferon's long track record and lack of evidence of long-term genetic injury to stem cells. However, imatinib is too new to assess its effects on long-term survival. The continued detection of BCR-ABL transcripts indicates that the disease is not cured, and the relative resistance of primitive hematopoietic cells suggests that cure may be elusive. In addition, the observation that new cytogenetic abnormalities may be discovered in Ph-negative cells requires careful long-term evaluation of all patients receiving this drug. However, the use of imatinib in combination with other therapies with different mechanisms of action may prove to be more effective. There are now ongoing trials of imatinib plus interferon and imatinib plus cytarabine. The use of imatinib as an adjunct to transplantation, either before the procedure to reduce leukemia burden or after grafting to reduce relapse risk, has not been assessed. T-cell depletion plus imatinib therapy with or without donor lymphocyte infusion might be a good strategy for clinical investigation.⁶⁰ Fortunately, ongoing work in immunology, stem cell research, infectious diseases, and molecular biology continues to progressively improve outcomes in stem cell transplantation. Some combination of pharmacologic therapy and immunologic attack may control or cure CML in the majority of patients. Other malignancies, such as gastrointestinal stromal tumors⁶¹ and hypereosinophilic syndrome,⁶² are particularly sensitive to the TK inhibition and also can be managed effectively with imatinib. A family of similar small-molecule TK inhibitors may be useful adjuncts in the therapy of acute myeloid leukemia.^{63 - 65}

Since Mrs P was unable to tolerate imatinib, I recommend that she undergo a matched, unrelated donor stem cell transplantation. This recommendation is based on her age, the availability of an allele-level HLA-compatible donor, and her excellent understanding of the risks and benefits of transplantation (she expressed these at greater length than space permits here). While it is hard to infer prognosis for a given patient, decision analysis suggests that this strategy is a good one to maximize quality-adjusted survival.⁴⁵ Based on current knowledge, this strategy is beneficial for young patients and patients who cannot tolerate imatinib. Mrs P has chosen to participate in a clinical trial with the goal of minimizing GVHD and maintaining GVL. Transplant-related toxicity and mortality increase with age,⁴⁶ and her precise risk is difficult to quantify with certainty. The challenge to clinicians is to assess their patients' tolerance for risk and uncertainty and integrate those factors with outcome data to help patients make an informed decision.

When treating patients older than 50 years or so, imatinib (or interferon) is an appropriate initial choice, and cytogenetics and quantitative PCR should be monitored if possible. If a complete cytogenetic response is observed, these patients can be monitored on a 6-month basis. If the quantitative PCR shows increasing BCR-ABL mRNA, if the cytogenetics do not clear, or if new cytogenetic abnormalities are discovered, transplantation should be strongly considered. With the advent of nonmyeloablative procedures, patients who are beyond the usual age limitations of stem cell transplantation may be eligible for this unique form of immunotherapy. If the disease appears to be stable, such monitoring should be continued indefinitely. In the next few years, as experience with imatinib matures, our ability to make firm recommendations will improve.

A PHYSICIAN: Is there any role for screening of the Philadelphia chromosome?

DR ANTIN: Screening for Philadelphia chromosome is not the necessary first step. Patients with CML do eventually get sick—it can take 6 months or a year, or sometimes 2 years. Mrs P was not feeling well for about a year or two before the diagnosis was made. This raises an interesting point: one thing that primary care physicians have been told is to be cautious about doing blood tests. However, if I were in an emergency department or primary care office and a patient came in and said, "I don't feel right. I haven't felt right for the past 2 or 3 months. I'm very tired," I would definitely obtain a complete blood cell count.

A PHYSICIAN: Can imatinib cure anybody? Is it being overtouted?

DR ANTIN: We do not know the answer to that question. It may cure some people; it is not going to cure everybody. Is it possible that there is a cohort of people, as with interferon, that will be alive 10 or 20 years later on imatinib and are PCR negative and are fine? I suspect that there will be. What we don't know is what proportion it will be. For the moment, I think that we have to make inferences from what we understand about the biology and from a very small amount of data. The way to go may be to try to develop additional therapies that will enhance the effects of imatinib, where you can reduce the dose to a certain point and then use some other form of therapy or specifically stem cell toxic therapy that will kill that leukemic progenitor.

A PHYSICIAN: From your standpoint, do you think imatinib will be overused like granulocyte colony-stimulating factor (filgrastim)?

DR ANTIN: I think it will be far less overused. Filgrastim can be used for all sorts of different things, some of which are relevant to the disease, some are not, some are useful, some aren't. Imatinib has a very specific effect in hematologic disorders that are due to isolated abnormalities of a TK gene^{62 ,66} and a role in a rare solid tumors.⁶¹ But it's really not a question of whether it's going to be abused—the question is how do you decide how long to keep people on it, or whether they should go on it. From a practical standpoint, we are putting all the patients on it, and seeing how they tolerate it. If they are young, we try to take them to transplantation with a reduced leukemic burden. If they are older, we wait and see what happens and hope that we catch them before something bad happens. The economic issues are significant, but I don't think it's subject to the same abuse potential as filgrastim.

A PHYSICIAN: Is all PCR-detectable evidence tantamount to a poor short-term outcome? In other words, is it conceivable that imatinib plus something can reduce but not eliminate it?

DR ANTIN: That is correct. The question is whether CML is like diabetes now. That is, that CML is not something that you are cured of but you can live with. Everyone is asking this question. The answer will be forthcoming, but not for several years.