From the time chronic myeloid leukemia (CML) was recognized as the first form of leukemia in 1845, it has probably become one of the best understood human malignancies. Its pathogenesis began to unravel in 1960 with the discovery that CML cells have a consistent cytogenetic abnormality, later termed the Philadelphia (Ph) chromosome.[1,2] In 1986, researchers discovered that the Ph chromosome carried a BCRABL fusion gene, and by the early 1990s, the encoded oncoprotein (P210-Bcr-Abl) was generally accepted as the initiating event in chronic phase CML as a consequence of its enhanced tyrosine kinase activity.[3,4] Thereafter, much effort was directed at inhibiting the kinase activity of this oncoprotein, culminating in the recent introduction into clinical practice of imatinib(Drug information on imatinib) mesylate (Gleevec), a Bcr-Abl tyrosine kinase inhibitor (Figure 1).[5-7] Most remarkably, within 3 years of its introduction, imatinib has proved to be the drug of choice in the management of patients with CML.[8] Prior to this important therapeutic milestone, other noteworthy advances had been made. The disease was generally considered incurable as re- cently as 25 years ago; since then, the use of allogeneic stem cell transplant (allo-SCT) has resulted in long-term remissions and almost certainly "cures" in selected patients.[9] However, it proved impossible to extend allo-SCT to all CML patients due largely to a lack of suitable donors and the increased risk of potentially lethal graft-vs-host disease (GVHD) in older recipients. The recent intro- duction of use of less-intensive transplant conditioning regimens has facilitated greater use of allo-SCT, but its precise role in the therapeutic algorithm for patients with CML remains to be defined.[10] Other important therapeutic advances have been the introduction of interferon-alfa in the early 1980s and the discovery that adoptive immunotherapy with donor-derived lympho- cytes could restore durable remissions in patients who relapsed following allo-SCT.[11-13] Despite these major developments, many important issues remain unresolved. In this paper, we will briefly address some of these questions and controversies. Molecular Biology The chronic phase of CML appears to arise as a consequence of a single pluripotential hematopoietic stem cell acquiring a BCR-ABL fusion gene associated with a Ph chromosome, which somehow confers a proliferative advantage to this stem cell over normal hematopoietic stem cells and thereby allows the BCR-ABL-containing cells to displace normal hematopoiesis. The Ph chromosome results from the reciprocal translocation of chromosomal material involving the long arms of chromosome 9 and chromosome 22, referred to as t(9;22)(q34;q11). Although more than 90% of CML patients have a Ph chromosome and a BCR-ABL fusion gene, about 8% of patients with hematologically "acceptable" CML lack the Ph chromosome and are described as having Ph-negative CML. About half of such patients have a cytogenetically occult BCR-ABL gene and are thus Ph-negative, BCR-ABL-positive cases; the remainder are BCR-ABL-negative; some of these have mutations in the RAS gene.[14] The mRNA molecules transcribed from the BCR-ABL fusion gene usually contain one of two possible BCRABL junctions, designated e13a2 and e14a2, respectively (Figure 2). Rarely, CML patients demonstrate an alternative consistent chromosomal translocation such as t(5;12)(q33;p11) and t(8;13)(p11;q12), both of which are associated with different oncoproteins with enhanced tyrosine kinase activity, namely platelet-derived growth factor receptor B and fibroblast growth factor receptor 1. Functional studies performed on cells from these leukemias suggest that the signal pathways activated are very similar to those activated in the Bcr-Abl-positive leukemias. It is generally accepted that the BCR-ABL fusion gene is the initial molecular abnormality in chronic phase; somewhat perversely, the successful application of imatinib is considered proof of this notion. Notably, Bcr-Abl transcripts can also be detected at very low levels in normal people, and the vast majority of such individuals do not develop CML.[15,16] The origin of these specific transcripts in normal people remains unclear. The precise mechanisms underlying the above observations, notably the cause of the chromosomal translocation and the precise nature of the proliferative advantage it confers, remain unknown. Interestingly, a 76-kb "duplicon" (a low copy DNA repeat sequence) has been found in close proximity to the ABL, BCR, and other genes.[17] Exposure to ionizing radiation increases the risk of developing CML; in vitro studies assessing the effects of high-dose irradiation on myeloid cell lines have demonstrated the development of Bcr-Abl transcripts indistinguishable from those that characterize CML.[18] However, the possible contribution of cosmic radiation in "causing" the typical sporadic case remains unknown. The Bcr-Abl oncoprotein appears to confer a number of key cellular changes including a reduced apoptotic response in mutated cells, decreased proteasome-mediated degradation of ABL-inhibitory proteins,[19] deregulation of cellular proliferation, and decreased adherence of CML cells to the bone marrow stroma and extracellular matrix. A number of different signal transduction pathways are known to be activated in the presence of a functioning Bcr-Abl oncoprotein, but precisely how it induces the leukemic phenotype is still largely a mystery (Figure 3). The progression of the chronic phase to the more advanced phases is presumably due to acquisition by the leukemia clone of one, or more probably, a series, of additional molecular changes, often in conjunction with recognizable new cytogenic abnormalities. In some cases, specific genes have been implicated in disease progression, notably p53, p16, RB, EVI-1, and possibly LYN. The nonrandom cytogenic changes that occur in advanced phase disease, principally +8, +Ph, +19, and iso (17)q, should eventually help to identify other new molecular events.[20] Prognosis Currently, several available methods may help predict survival for individual patients. Some of these methods, such as the Sokal prognostic index, are based on criteria definable at diagnosis and correlated with duration of survival for subgroups of patients treated predominantly with busulfan(Drug information on busulfan) (Busulfex, Myleran).[21] This technique was useful during the busulfan era and may still have value. The Euro system, introduced by Hasford and colleagues, is an analogous system for predicting survival of patients treated with interferon-alfa and may more accurately discriminate prognosis for patients treated with interferon-containing regimens.[22] Other possible prognostic factors are the presence of genomic deletions in the region of the reciprocal ABL-BCR on the derivative 9q+ chromosome and the rate of shortening of telomeres in the leukemia clone.[23,24] It is likely that DNA microarray studies will also play a future role in the staging of patients with CML. Treatment Although allo-SCT is the only treatment strategy that currently results in long-term molecular remission and probable "cure" in CML patients in chronic phase, this procedure is only available to less than onethird of patients.[25] Until recently, the standard treatment for newly di- agnosed CML patients in chronic phase not eligible for an allo-SCT was interferon-alfa, either alone or in combination with low-dose cytarabine(Drug information on cytarabine).[ 26] Interferon-alfa largely replaced hydroxyurea in the mid- 1990s, when it was demonstrated that it induces major cytogenetic responses in about one-third of patients and an overall survival advantage of 1 to 2 years compared to hydroxyurea.[27] Interferon-alfa treatment was associated with a wide range of side effects including flu-like symptoms, lethargy, depression, and weight loss. In an attempt to reduce this toxicity, clinicians have begun using a longacting form of the drug-pegylated interferon-alfa. The notion of adding cytarabine to interferon-alfa appeared attractive on the basis of a recent study demonstrating superior survival for the combination compared to interferon- alfa alone, but this result has more recently been called into question.[27] Today most hematologists would regard imatinib alone or in combination with other agents as the treatment of choice for patients not destined for immediate allogeneic SCT.[28] Imatinib Mesylate
Imatinib entered clinical trials for patients with CML in chronic phase as well as those in more advanced phases in 1998.[29] The drug caused a rapid reversal of clinical and hematologic abnormalities and major cytogenetic responses in over 50% of chronic phase patients. It was administered orally, and side effects were relatively minor, with nausea, headache, rashes, and fluid retention being the most common; significant cytopenias and hepatotoxicity were less frequent. The toxicity, in general, seems to be appreciably less than that associated with interferon-alfa. Current studies confirm the initial impressive results of imatinib.[8,30,31] Notably, over 95% of patients in chronic phase who are refractory or resistant to interferon-alfa achieved a complete hematologic response, and 55% of these patients achieved a major cytogenetic response, but thus far very few patients have achieved convincing molecular remissions.[32,33] To what extent such patients obtained survival benefit could not immediately be ascertained, though it does now appear that patients who obtained cytogenetic responses survive longer than matched controls.[34] A prospective, randomized phase III trial designed to answer this question therefore started in 2000, and the interim results were published recently[35]; they revealed that 74% of the patients treated with imatinib achieved a complete cytogenetic remission (CCR). Progression-free survival was significantly better in the imatinib-treated cohort compared to the interferon-alfa and cytarabine cohort (97.2% vs 90.3%, P < .001), but it is too early to expect evidence of prolonged surviv- al. It is therefore not possible to conclude that imatinib as a single agent cures substantial numbers of patients, but it may well offer the prospect of "operational cure" to a significant proportion. Moreover, whether imatinib is superior to interferon-alfa plus cytarabine in previously untreated chronic phase CML patietns remains to be determined. In the molecular analysis of the cohort achieving a CCR, 3.6% of these patients achieved a complete molecular remission (defined by a complete absence of detectable BCR-ABL transcripts in the blood; the investigators considered this to represent > 4.5 log reduction in BCR-ABL/BCR level when compared to the median pretreatment level).[36] Although imatinib appears to be quite safe, caution must be exercised in light of several recent reports. Gratwohl and his colleagues have reported a potentially fatal side effect- cerebral edema-soon after the initiation of imatinib therapy.[37] An interesting nonsinister effect-hair repigmentation-has been reported in a small cohort of responders.[38] The mechanisms for these unique effects remain speculative, although the possible inhibition of the platelet-derived growth factor receptor has been suggested for the former effect. Imatinib is known to be a potent competitive inhibitor of the tyrosine kinase associated with the platelet-derived growth factor receptor, and a recent report has confirmed its usefulness in patients with chronic myeloproliferative disorders with rearrangements of this receptor.[39]

• Imatinib in Advanced Phases of CML-CML-Studies have also confirmed impressive (although less durable) clinical activity of imatinib in advanced phases of CML.[40-42] This is remarkable because imatinib targets mainly the Abl kinase activity of the Bcr-Abl oncoprotein, and as other (additional) genetic events underlie disease progression, the drug might have been expected to have little or no activity in the advanced phases. It is now likely to be tested in combination with various chemotherapeutic agents. Notably, some in vitro evi- evidence demonstrates that the combination of imatinib and agents such as interferon-alfa and hydroxyurea may result in anatagonism.[43,44]
• Imatinib-Refractory Patients-Acquired resistance to imatinib among patients in chronic phase appears to be rare and can often be overcome by increasing the dose of the agent.[45] In contrast, resistance has been seen in up to 70% of those in myeloid blast crisis, and all patients in lymphoid blast crisis relapse within 6 months of responding to imatinib. This reaction appears to result from a variety of diverse mechanisms, including acquired mutations in the Abl kinase domain, Bcr-Abl overexpression, Pglycoprotein overexpression reducing the cellular uptake of imatinib, selection of preexisting mutant cells, and possibly excessive degradation of the Bcr-Abl protein.[46-49] The intrinsic production of some proteins, such as alpha-1 acid glycoprotein or a P450 enzyme, may neutralize imatinib and render it ineffective.[50] Of great interest is the recent finding of "acquired" mutations, which result in structural changes that prevent imatinib binding but do not prevent pathologic phosphorylation of the relevant substrates by the oncoprotein. Currently, at least 18 different mutations have been described and are associated with some degree of resistance to imatinib. Of particular interest is the recent report that mutations in the P loop of the Abl kinase domain predict for disease progression, whereas mutations not involving the P loop are less ominous.[ 51] Various lines of evidence suggest that these "acquired" mutations reflect selection by imatinib of mutant clones already present at low levels before inititation of treatment rather than de novo acquisition during imatinib therapy.[52-55] These observations mean that even when multiple additional genetic events predominate in the advanced stages of CML, the original molecular event still appears to play some role in maintaining the aggressively transformed phenotype, emphasizing the importance of BCR-ABL in the pathogenesis of CML.[56]

Interferon-Alfa
Interferon-alfa treatment was until recently the mainstay of treatment for CML. It is associated with a wide range of side effects including flulike symptoms, lethargy, depression, and weight loss. In an attempt to reduce this toxicity, clinicians have begun using a long-acting form of the drug-pegylated interferon-alfa. The notion of adding cytarabine to interferon- alfa appeared attractive on the basis of a recent study demonstrating superior survival for the combination compared to interferon-alfa alone, but this result has more recently been called into question.[28] It is remarkable that despite being in clinical use for 2 decades, the precise mechanism of interferon-alfa's action remains unknown. Although it was undoubtedly a valuable drug, it did not appear to produce any durable molecular responses. If one defines "cure" as complete eradication of all leukemia cells (which would require persisting failure to detect BCR-ABL transcripts by reverse transcription polymerase chain reaction), then interferon- alfa did not result in cure, but in a small percentage of patients, an "operational cure"-low numbers of cells persisted but appeared unable to reestablish clinical disease. Allogeneic Stem Cell Transplant
Cure by allo-SCT depends on the combined effects of chemotherapy or chemoradiotherapy conditioning before transfusion and the graft-vs-leukemia effect mediated by allogeneic T lymphocytes.[57] Current results using human leukocyte antigen (HLA)-identical sibling donors suggest the probability of event-free survival at 5 years of about 60%; the probability of relapse at 5 years is 15%.[58] In contrast, the results of allo-SCT performed in more advanced phases of the disease are generally poor.[59] Because only about one-third of all patients considered for an allo-SCT have an HLAmatched sibling donor, many efforts have been directed toward the identification of suitable alternative donors, ie, either partially matched family members or unrelated volunteers. Clinical results with these alternative donors appear to be slightly inferior or comparable to HLA-matched sibling transplants.[60-62] The major determinants of survival, other than the phase of the disease, include the patient's age at transplant, the duration of disease from diagnosis, the cytomegalovirus status of the patient, acute and chronic GVHD, and the sex of the donor. Thus, survival appears to be best among patients who are transplanted within 1 year of diagnosis, are less than 40 years of age, and have a male donor, and in cases where both patient and donor are cytomegalovirus- seronegative.[63] For such a cohort, 5-year disease-free survival is around 70% to 80%, and the relapse rate, 10% to 20%.[64] The precise details of the transplant procedure and the choice of stem cells (marrow vs peripheral blood) also influence outcome.[65] Engraftment appears to be more rapid following a peripheral blood stem cell transplant, but the incidence of chronic GVHD may be increased.[66,67] It remains unclear whether the risk of relapse differs significantly following transplantation of allogeneic blood cells rather than marrow cells.[68,69] Acute and chronic GVHD continues to be a significant cause of transplant- related mortality. The best approach to prophylaxis of GVHD, usually a combination of cyclosporine and methotrexate(Drug information on methotrexate), remains controversial. Over a decade ago, it was demonstrated that alloreactive T lymphocytes cause GVHD reactions and that T-cell depletion of the graft substantially reduces the incidence of acute and chronic GVHD but also increases the risk of relapse.[70] Allo-SCT using T-cell depletion with the CD52 monoclonal antibody Campath-1 results in an actuarial relapse rate (after transplantation in chronic phase) of 60% to 70%. Other methods of T-cell depletion, including use of other antibodies and E-rosette formation/soybean lectin agglutination, also reduce the incidence of GVHD but increase the incidence of relapse to varying degrees.[71] Recently, it was shown that transplantation of highly purified CD34+ peripheral blood stem cells with Tlymphocyte add-back does not appear to be associated with an increased risk of relapse; this strategy is being tested further.[72] Prior use of interferonalfa may be associated with an inferior survival, although reports are conflicting.[73-75] Preliminary data from the European Group for Blood and Marrow Transplantation (EBMT) suggest that treatment with imatinib mesylate does not have a deleterious effect on subsequent allogeneic transplantation,[ 76,77] but this will require further observation.