Chapter 4: Chronic Myeloid Leukemia

Chronic myeloid leukemia (CML) is a myeloproliferative disorder of pluripotent hematopoietic stem cells, characterized by the molecular BCR-ABL1 rearrangement, which drives a proliferative and survival advantage of the leukemic clone. About 30% to 50% of patients with CML are asymptomatic at diagnosis and are incidentally diagnosed during routine examination. Patients may also present with characteristic clinical findings secondary to large numbers of myeloid circulating progenitors, leading to splenomegaly, leukocytosis, or even isolated thrombocytosis. The landscape of CML has had a dramatic course, with a host of findings that have elucidated the biology and molecular pathology of the disease. The understanding of the molecular events led to the creation of the first targeted therapy—imatinib mesylate. Its impact on therapy and survival propelled CML as a model for modern molecular medicine in the fast-developing era of personalized targeted therapy.

The incidence of CML is 1 to 2 cases per 100,000 adults with a slight male predominance and rising incidence with age, accounting for approximately 15% of newly diagnosed cases of leukemia in adults (¹). Chronic myeloid leukemia is uncommon in children, with a median age at diagnosis of 67 years. There are no known hereditary, geographic, familial, or ethnic associations. No chemical or infectious associations have been established, although an increased risk has been linked with exposure to ionizing radiation (²).

In 2014, in the United States, an estimated 6,000 cases of CML were diagnosed. Since 2000, the year of introduction of imatinib, the annual mortality in CML has decreased from 10% to 20% to 1% to 2%. Consequently, the prevalence of CML in the United States, which was estimated at about 25,000 to 30,000 cases in 2000, has increased to an estimated 80,000 to 100,000+ cases in 2015 and will reach a plateau of about 180,000 cases by 2030 (³). The overall survival (OS) of patients over the recent decade has greatly improved. The exact mechanism of initiation of the defining molecular event in CML—Philadelphia (Ph) chromosome translocation—remains elusive.

Chronic myeloid leukemia is defined by a unique cytogenetic and/or molecular abnormality—the Ph chromosome—originating in a pluripotent stem cell, with a balanced translocation between the long arms of chromosomes 9 and 22, t(9,22)(q34,q11.2) (⁴). Detectable by routine cytogenetics or by fluorescent in situ hybridization (FISH), the Ph chromosome is noted in 90% to 95% of patients with the clinical and laboratory features of CML. In the remaining 5% to 10% of patients, the molecular BCR-ABL1 rearrangement can be recognized with high-sensitivity reverse transcriptase polymerase chain reaction (RT-PCR). All remaining cases with unknown biology deemed as true Ph-negative CML or atypical CML carry a poor prognosis. The Ph chromosome joins the proto-oncogene c-abl from chromosome 9 to the breakpoint cluster region (BCR) gene in chromosome 22, generating a novel fusion BCR-ABL1 oncogene. Depending on the breakpoint region in BCR, several variants including those of 190, 210 (most common; 95% of Ph-positive cases), or 230 kDa molecular weight can be formed. The translation product is a chimeric protein with constitutively active tyrosine kinase activity that activates multiple downstream pathways, including PI3 kinase, NF-κB, JAK/STAT, RAS, RAF, ERK, MYC, and JNK.

Following the 2005 discovery of the JAK2 V617F mutation as the molecular abnormality common to polycythemia vera, essential thrombocytosis, and idiopathic myelofibrosis, CML has been segregated into a separate group of myeloproliferative neoplasms categorized as BCR-ABL1 positive/JAK2 V617F negative.

Roughly 10% to 15% of patients, typically at a more advanced stage of disease, experience clonal evolution with additional chromosomal aberrations including trisomy 8, monosomy 7, isochrome 17, a double Ph chromosome, or additional loss of material from 22q. Clonal evolution is considered a criterion of accelerated phase, particularly when it occurs during the course of the disease. Corresponding molecular alterations that follow these changes include deregulation of p53, RB1, C-MYC, and AML-EVI1.

There are three separate phases of CML: chronic phase (CP), intermediate or accelerated phase (AP), and terminal or blast phase (BP). Even though all three represent a stepwise progression of disease in terms of aggressiveness from CP to BP, the natural course of the disease may not progress from one to the other, nor will disease always include all three. The vast majority of patients are diagnosed as asymptomatic in the CP. Splenomegaly is the most consistent presenting sign, detected in 50% to 60% of cases. Previously defined criteria indicating progression from CP to AP include 15% or more blasts, 30% or more blasts plus promyelocytes, 20% or more basophils, platelets <100 × 10⁹/L unrelated to therapy, or cytogenetic clonal evolution (⁵). Other criteria have been proposed (Table 4-1). Compared to CP, median survival in AP is significantly shorter, although a significant improvement in survival has been observed with the availability of tyrosine kinase inhibitors (TKIs), particularly among early responders (⁶).

Most patients evolve into AP prior to BP, but 20% progress to BP without AP warning signals. Blast phase is considered an acute leukemia. It is defined by the presence of at least 30% blasts in the peripheral blood or bone marrow or the presence of extramedullary disease (chloroma or granulocytic sarcoma). Half of the patients in BP carry a myeloid phenotype, whereas the other half is split evenly between lymphoid and undifferentiated (⁷). Median survival in BP CML remains poor; a combination of TKI with chemotherapy followed by allogeneic stem-cell transplantation is recommended.

Initial evaluation aims to elicit signs and/or symptoms of disease, whereas physical examination evaluates for any presence of organomegaly or extramedullary hematopoiesis. Diagnosing typical CML requires documentation of the Ph chromosome abnormality by routine cytogenetic evaluation (karyotype), FISH, or molecular studies (RT-PCR) (⁸). Bone marrow aspiration and biopsy are required, because they will not only confirm the diagnosis (eg, cytogenetic analysis), but also provide information needed for staging and classification. Although FISH, which relies on co-localization of large genomic probes specific to the BCR and ABL genes, or RT-PCR from peripheral blood can both confirm the presence of the Ph chromosome or the BCR-ABL1 rearrangement, only bone marrow cytogenetics can reveal additional chromosomal abnormalities that are important in the initial diagnosis and staging.

True Ph- and BCR-ABL1–negative patients are considered to have Ph-negative CML or chronic myelomonocytic leukemia. In a few instances, a patient may be diagnosed with Ph-positive acute myeloid leukemia (AML) or acute lymphoblastic leukemia (ALL). It may be impossible to determine whether these cases represent de novo Ph-positive acute leukemias or a BP of a previously unrecognized CML.

The most common laboratory finding in CP CML is leukocytosis, with myeloid cells in all stages of maturation seen in the peripheral blood. Frequently, there is an increase of basophils and eosinophils. The bone marrow is markedly hypercellular, with the myeloid-to-erythroid ratio significantly increased (Figs. 4-1,4-2,4-3,4-4,4-5) (⁹).

A marrow aspiration and chromosomal analysis are needed to identify any variances in marrow blasts or basophils and to identify the Ph chromosome and possible cytogenetic clonal evolution (Fig. 4-6). Fluorescent in situ hybridization and PCR should be done at diagnosis in all patients. FISH may identify the presence of the BCR-ABL1 rearrangement, even when the Ph chromosome is not found by conventional cytogenetic analysis. Furthermore, FISH can be performed on peripheral blood interphase cells (Fig. 4-7). Reverse transcriptase PCR testing amplifies the region around the splice junction between BCR and ABL and is highly sensitive for detecting CML, identifying the initial type of BCR-ABL1 transcript (important for follow-up), and especially quantifying minimal residual disease (MRD). In general, simultaneous peripheral blood and marrow quantitative PCRs have a high level of concordance, although false-positive and false-negative results can be seen (false-negative results due to poor-quality material or failure of the reaction; false-positive results due to contamination). Quantitative PCR assesses the amount of BCR-ABL1 transcripts and is universally performed. Table 4-2 presents a summary of the proposed evaluation at diagnosis and during follow-up for patients with CML.

As the newer generation agents became available, the prognostic significance of certain clinical characteristics changed. Achieving a complete cytogenetic response (CCyR) is the most important prognostic factor for long-term survival (¹⁰). It has become evident that achieving a cytogenetic response, particularly during the first 3 to 6 months of therapy, translates into the best probability of long-term outcome.

Response to therapy is initially judged by measurement of hematologic response criteria. A complete hematologic response (CHR) is defined as normalization of white blood cell (WBC) counts to <10 × 10⁹/L with a normal differential, platelet count <450 × 10⁹/L, and disappearance of splenomegaly and other symptoms of CML. Patients who achieve a CHR are further classified according to the type of cytogenetic response attained (Table 4-3).

Cytogenetic responses are divided into complete, partial, and minor. Complete cytogenetic response corresponds to 0% of all metaphases remaining Ph positive; partial cytogenetic response is defined as 1% to 35% of metaphases being Ph positive; and minor cytogenetic response is defined as 35% to 95% Ph-positive metaphases. An analysis involving review of at least 20 metaphases is necessary to evaluate a cytogenetic response. Although FISH results correlate well with the karyotypic evaluation, cytogenetic response criteria have not been validated with FISH.

Molecular response is assessed by quantitative PCR (usually RT-PCR) in the peripheral blood or bone marrow (¹¹). A major molecular response (MMR) is considered when the BCR-ABL1/ABL1 ratio is ≤0.1% on the international scale (IS). A complete molecular response (CMR) represents achievement of undetectable transcripts of BCR-ABL1 in an assay with a sensitivity of at least 4.5 logs. For correlative purposes, BCR-ABL1 transcript levels (IS) of ≤1% are equivalent to a CCyR; levels of ≤10% are equivalent to a partial cytogenetic response.

Three commercially available TKIs are approved for the frontline treatment of CML: imatinib, dasatinib, and nilotinib. Available guidelines support all three as viable frontline options for the initial management of CP CML.

Imatinib Mesylate

Prior to the development of targeted therapy, CML was treated with busulfan or hydroxyurea for many years, with a poor prognosis and inability to delay disease progression. The introduction of interferon-alfa (IFN-α) improved survival in CML and resulted in CCyRs in 5% to 25% of patients with CP CML.

Imatinib mesylate (STI-571 or Gleevec) is a selective and potent competitive inhibitor of the adenosine triphosphate (ATP)–binding site of the Bcr-Abl oncoprotein, as well as c-kit, PDGFRα and PDGFRβ, and abl-related gene (ARG) (¹²). It is taken orally with 98% bioavailability and a half-life of 13 to 16 hours. It was first used in CML patients who developed resistance or intolerance to IFN-α and resulted in a CCyR of 60% and an estimated 5-year OS of 76% (¹³).

Based on these results, the large International Randomized Study of Interferon and STI571 (IRIS) trial evaluated imatinib versus IFN-α and low-dose cytarabine as frontline therapy in patients with newly diagnosed CP CML. Patients were randomized to receive imatinib 400 mg/d or INF-α plus low-dose subcutaneous cytarabine, which was standard of care at the time (¹⁰). After a median follow-up of 19 months, planned outcome analyses were significantly better in almost all categories for patients receiving imatinib. The rates of CCyR (74% vs 9%; P < .001) and freedom from progression to AP or BP at 12 months (99% vs 93%; P < .001) were improved. There was a high crossover rate to imatinib. The responses to imatinib were durable, as outlined in an 8-year follow-up of the original study (¹⁴). Estimated event-free survival was 81%, and OS was 93% when only CML-related deaths were considered. With an annual mortality of 2%, the estimated median survival of a newly diagnosed patient with CML may be in the range of 20 to 30 years.

Imatinib Dose

A phase I imatinib study in patients who had failed prior IFN-α therapy established a clear relationship between dose and response (¹⁵). No significant responses were observed at doses <300 mg daily. An arbitrary dose of 400 mg daily for CP was selected in phase II studies, despite the lack of dose-limiting toxicity at doses up to 1,000 mg daily (maximum-tolerated dose was not defined). The Rationale and Insight for Gleevec High-Dose Therapy (RIGHT) trial studied imatinib 400 mg twice a day as initial therapy in 115 patients with newly diagnosed CML (¹⁶). The rate of CCyR was 85% at 12 months and 83% at 18 months, with corresponding rates of MMR of 54% at 12 months and 63% at 18 months.

These results led to a randomized phase III open-label study, the Tyrosine Kinase Inhibitor Optimization and Selectivity Trial (TOPS), comparing 400 and 800 mg of imatinib in 476 patients. The trial showed significant superiority for the 800-mg dose in terms of MMR rate at 3 months (3% vs 12%), 6 months (17% vs 34%), and 9 months (33% vs 45%) but not at 12 months (40% vs 46%) (¹⁷). A final update of the trial showed no significant difference in MMR rates at 24 months (52% vs 50%) (¹⁸). Another European study also reported no benefit with imatinib 800 mg compared to 400 mg in high-risk CML (CCyR rates were 64% and 58% and MMR rates were and 40% and 33%, respectively) (¹⁹).

The French SPIRIT study evaluated the impact of adding IFN or cytarabine in a randomized study where 636 patients with untreated CP CML received one of the following: imatinib 400 mg daily alone, imatinib 400 mg daily with cytarabine (20 mg/m²/d on days 15-28 of each 28-day cycle) or pegylated IFN-α-2a (90 μg weekly), or imatinib 600 mg daily alone (²⁰). The rates of cytogenetic response at 12 months were similar among the four groups, whereas the rate of molecular response (a decrease in the ratio of BCR-ABL1/ABL1 of ≤0.01%) was significantly higher in patients receiving imatinib and pegylated IFN-α-2a compared with imatinib 400 mg alone arm (30% vs 14%, respectively; P = .001). This rate was also significantly higher in patients treated for more than 12 months compared to treatment lasting ≤12 months. However, this high rate of early and deep responses did not translate into long-term improvement due to the poor tolerance of pegylated IFN.

Imatinib 400 mg daily is the regimen of choice for newly diagnosed patients with CP CML, with an emphasis on maintaining adequate dose intensity with minimal treatment interruptions or dose reductions for the best outcome.

Management of Toxicity

Imatinib is well tolerated, although adverse events not requiring treatment interruptions or decrease in dosing may occur in 30% to 40% of patients. A list of some of the most frequently encountered side effects and suggestions for management are included in Table 4-4. Any grade 3 or 4 toxicities related to imatinib require treatment interruption and resumption upon resolution of toxicity or its decrease to grade 1 or less. Subsequent dose should be reduced if recurring or long-lasting adverse effects are encountered, keeping in mind that doses below 300 mg daily are not recommended due to lack of adequate activity. Only 2% to 3% of patients exhibit true intolerance to imatinib and require permanent discontinuation. Early recognition and intervention targeting toxicities greatly reduce the need for unnecessary treatment interruptions and dose reductions.

Myelosuppression is common and frequently seen within the first 2 to 3 months of therapy. It is generally self-limited, and dose interruptions are not recommended unless grade 3 neutropenia or thrombocytopenia (ie, neutrophils <1 × 10⁹/L, platelets <50 × 10⁹/L) develops. Anemia alone usually does not require interruptions or dose adjustments. Treatment is restarted when counts recover above specified thresholds. Following treatment interruption, WBC should be monitored at least once weekly, and if recovery occurs within 2 weeks, treatment would be resumed with the same dose at which myelosuppression occurred. If recovery takes longer than 2 weeks, the dose could be reduced in increments (eg, from 800 to 600 mg, from 600 to 400 mg, or from 400 to 300 mg). Hematopoietic growth factors may be beneficial with recurrent or prolonged myelosuppression (eg, erythropoietin or darbepoetin and filgrastim).

Dasatinib

Dasatinib (Sprycel) is an oral second-generation TKI that is a piperazinyl derivative. It has an excellent oral bioavailability and is 350 times more potent in vitro than imatinib (^21,22) (in vitro sensitivity of different BCR-ABL1 mutants to different TKIs is presented in Table 4-5) (²³). Dasatinib exhibits significant activity against most imatinib-resistant BCR-ABL1 mutations, with the exception of T315I, as well as a few others, including V299L and F317L (²⁴). In contrast to imatinib, dasatinib binds to both the active and inactive conformations of BCR-ABL1 and also inhibits the Src family of kinases, which may be important in suppressing critical cell signaling pathways (²⁵).

Following evaluation in the salvage setting after imatinib failure, dasatinib was assessed in the frontline setting. The DASISION trial was a phase III randomized study that compared imatinib 400 mg once daily to dasatinib 100 mg once daily in 519 patients with newly diagnosed CP CML (²⁶). The primary end point was confirmed CCyR at 12 months. The dasatinib arm resulted in higher confirmed CCyR at 12 months (77% vs 66%; P = .007). The rates of molecular response were significantly higher with dasatinib (MMR, 76% vs 64%, P = .002; molecular response with a 4.5-log reduction in BCR-ABL transcripts from baseline [MR^4.5], 42% and 33%, P = .025). Dasatinib induced deeper responses at early time points (3, 6, or 12 months) compared to imatinib. The rate of transformation to AP or BP was lower in patients treated with dasatinib (4.6% and 7.3%, respectively). There was no progression-free survival (PFS) or OS difference at 5 years. Relevant toxicities included pleural effusion rate of 29% with dasatinib (mostly grade 1 or 2; 15 patients discontinued dasatinib due to pleural effusion). Arterial ischemic events were slightly higher with dasatinib (5% vs 2%, respectively). Pulmonary hypertension was reported in 14 dasatinib-treated patients, with 6 discontinuing the drug. Comparison of the phase III trials in the frontline treatment of CP CML with imatinib, dasatinib, and nilotinib is outlined in Table 4-6.

A randomized phase II trial compared dasatinib 100 mg daily with imatinib 400 mg daily in 253 patients with newly diagnosed CP CML. Higher rates of CCyR (84% vs 69%) and 12-month MMR (59% vs 44%; P = .059) were reported in patients receiving dasatinib. No difference in PFS or OS was reported. Grade 3 and 4 toxicities were most commonly hematologic, including thrombocytopenia, which was more common with dasatinib (18% vs 8%) (²⁷).

The results of the SPIRIT-2 trial were recently reported (²⁸). More than 800 patients with newly diagnosed CML were treated in a phase III trial and were randomized to either dasatinib 100 mg daily or imatinib 400 mg daily (²⁸). The 12-month CCyR and MMR rates were higher with dasatinib (CCyR: 51% vs 40%, P = .002; MMR: 58% vs 43%, P < .001). Among 40 patients who discontinued therapy due to suboptimal response, only three patients (1%) received dasatinib. The PFS and OS were not significantly different. The rate of grade 3 or 4 thrombocytopenia was higher with dasatinib (13% vs 4%). Pleural effusions were observed in 78 patients (19%) treated with dasatinib (13 required drainage). The rate of cardiovascular events was slightly higher with dasatinib (2% vs 0.5%).

Dasatinib is otherwise well tolerated. Myelosuppression occurs frequently, with grade 3 or 4 neutropenia or thrombocytopenia in 20% of patients. The most common nonhematologic grade 3 or 4 toxicities at the same dose were pleural effusion (9%), dyspnea (6%), bleeding (4%), diarrhea (3%), and fatigue (3%).

Nilotinib

Nilotinib (Tasigna) is a structural analog of imatinib with 50 times more potent affinity for the ATP-binding site in vitro (²⁹) and more selective activity against unmutated and most mutated forms of BCR-ABL1 (^29,30). It is approved at a dose of 400 mg twice daily for patients with CP or AP CML who have resistance or intolerance to imatinib.

After approval for patients who failed imatinib therapy, nilotinib was evaluated in newly diagnosed CML in CP. The ENESTnd study was a randomized phase III trial comparing two different dose schedules of nilotinib (300 and 400 mg twice daily) to imatinib 400 mg once daily as initial therapy for patients with early CP CML (³¹). The primary end point was the rate of MMR at 12 months, which was higher with both doses of nilotinib compared to imatinib (44% and 43% vs 22%; P < .001). The rate of transformation to AP or BP by 12 months of therapy was significantly lower with nilotinib (<1%) compared to imatinib (4%). The adverse effect profiles showed a higher rate of cardiovascular events with nilotinib (10% with nilotinib 300 mg twice a day; 16% with nilotinib 400 mg twice a day; and 2% with imatinib). The 6-year follow-up continues to demonstrate higher rates of early and deeper sustained molecular response with nilotinib, a reduced risk of progression to AP and BP, and an acceptable safety profile (³²). Nilotinib is well tolerated, with grade 3 or 4 myelosuppression as the most common adverse event (neutropenia or thrombocytopenia observed in 10%-20% of patients). Nonhematologic toxicity includes liver function abnormalities in 10% to 15% of patients and asymptomatic elevation of lipase and amylase in 10% to 15% of patients. Vascular adverse events were reported at an cumulative rate of 10% over 6 years. Rare cases (<1%) of pancreatitis have been reported. Nilotinib has the potential for QTc prolongation, and a baseline electrocardiogram is required prior to the start of therapy. Diabetes may be exacerbated with nilotinib.

Selecting a Frontline Therapy

With multiple TKIs available for newly diagnosed CP CML, there are several considerations when choosing a starting agent, such as efficacy, patient status (eg, age; comorbidities; history of diabetes, hypertension, pancreatitis, chronic lung disease, pulmonary hypertension, cardiac history), and treatment value. The high prices of TKIs are of concern, given that patients can now remain on TKIs and expect to live normal lives (³³). The prices for the second-generation dasatinib and nilotinib are comparable, both costing more than $100,000 annually. Once imatinib becomes available as a generic drug, the choice of TKIs in relation to value (benefit-to-price) needs to be considered, particularly in patients with low-risk disease.

For patients with baseline cardiopulmonary comorbidities such as chronic obstructive pulmonary disease, congestive heart failure, or uncontrolled hypertension or pulmonary arterial hypertension, a TKI other than dasatinib may be favored, given the risk of pleural effusions. Dasatinib also impairs platelet function, and patients on concomitant anticoagulants may be at an increased risk for hemorrhagic complications (³⁴).

Nilotinib has been linked with hyperglycemia and QT interval prolongation, and should be used with caution in uncontrolled diabetics and in patients with baseline QT prolongation (routine monitoring of the QT interval is essential). Potassium and magnesium should be repleted to optimal serum levels prior to starting nilotinib, and the drug should be taken on an empty stomach twice daily. Recently, nilotinib has been associated with a low but significant incidence of peripheral artery disease, cerebrovascular accidents, and cardiovascular syndromes (³⁵). Therefore, it is reasonable to choose other TKIs for patients with cardiovascular morbidities. Nilotinib rarely causes pancreatitis and should be avoided in patients with prior history of pancreatic inflammations. Imatinib is associated with the development of peripheral edema as one of its major side effects. Among patients with significant baseline peripheral edema, nilotinib or dasatinib may be favored as first options; close monitoring and intermittent use of loop diuretics might mitigate the effects of fluid retention.

Monitoring involves routine blood counts with differentials, cytogenetics, and molecular testing for BCR-ABL1 transcript levels and for BCR-ABL1 kinase domain mutations. Blood counts should be performed every 1 to 2 weeks until CHR and at least every 3 months thereafter or more frequently as clinically indicated (³⁶). Cytogenetic analysis is the only test that gives reliable information regarding the presence of additional chromosomal aberrations.

At baseline, all patients undergo a marrow analysis to establish the diagnosis and provide a sample for cytogenetic testing. This also allows for proper staging in terms of the blast and basophil percentage. Presently, it is recommended that patients have a follow-up bone marrow study at 3, 6, and 12 months after starting therapy (³⁷). An alternative method to determine cytogenetic response is with the use of FISH on peripheral blood. If a patient is responding optimally, and the FISH study is negative at 6 or 12 months, it may be reasonable to omit further marrow exams, as the patient is likely to be in stable CCyR (^38,39). Once a patient achieves a stable CCyR, particularly if associated with MMR, bone marrow aspirations with cytogenetics are recommended only every 1 to 3 years or if there are significant changes in the transcript levels or peripheral blood counts.

For patients in durable CCyR, periodic molecular monitoring every 3 to 6 months using quantitative RT-PCR is acceptable and useful, but may lead to erroneous changes in treatment due to discordance in results between labs or even within the same lab. This is harmful to patients because it leads to potentially discontinuing a useful therapy that the patient may have been tolerating well. One strategy to minimize this is to use interphase FISH as a complementary diagnostic test along with the molecular test to detect possible false-positive or false-negative results generated by either assay (³⁹). For patients in CCyR, the achievement and maintenance of an MMR is questionable. Studies evaluating patients receiving imatinib or second-generation TKIs found that patients in CCyR have similar survival whether there is achievement of MMR or not (^40,41,42). For patients in MMR, periodic molecular monitoring every 6 months is useful.

The value of early molecular response has been shown in a number of studies to have strong prognostic value. This has been applied to each of the TKIs appropriate for use in the frontline setting. A BCR-ABL1 transcript level of less than 10% at 3 months has been shown to separate groups into high- and low-risk categories for long-term outcomes (ie, progression, survival) (^40,43,44). An important question is what to do with a patient who does not meet the 3-month benchmark. One option is to switch TKIs early, but there are currently no data showing this would alter long-term outcome. Several experts have suggested that a follow-up measurement at 6 months will help define patients clearly in need of a change in therapy (³⁷). This has been retrospectively analyzed with conflicting results (^45,46,47). At least two independent studies suggested that patients with BCR-ABL1 transcript levels greater than 10% at 3 months do not necessarily have an inferior outcome (^46,47). Patients who continued on therapy and achieved transcripts levels less than 10% by 6 months had the same long-term favorable outcome as patients with optimal molecular responses at 3 months.

Although achieving undetectable or the lowest possible transcript level is desirable, molecular positivity above the levels of MMR in the context of a CCyR is not an indication of failure of therapy. Some studies have suggested that increasing transcript levels may increase the risk of developing mutations or failure of therapy, but the magnitude of the increase that may predict for such events is variable, partly due to the variability of the testing in different laboratories. This may lead to erroneous changes in treatment, which is harmful, as it leads to potential discontinuation of viable and tolerable therapy. A single elevation in transcript levels should be confirmed in a subsequent determination 1 to 3 months later, and incremental increase in BCR-ABL1 transcript levels should be determined to be greater than the variability of the laboratory test. In such situations, compliance with therapy should first be revisited. The risk of relapse or emergence of mutations is mostly associated with a sizeable (five- to tenfold) increase in the BCR-ABL1 transcript levels that is associated with a loss of MMR or occurs in a patient who never achieved an MMR. Changes still below the level of MMR have little if any prognostic significance. Several studies evaluating patients receiving imatinib or second-generation TKIs showed that patients in CCyR have similar survival with or without MMR (^40,41). ABL kinase domain mutation screening should be performed in patients who have an inadequate initial response (defined as failure to achieve CHR at 3 months, partial cytogenetic response at 6 months, and CCyR at 12 months), in patients who show hematologic or cytogenetic relapse, or in patients with sustained 1-log increase in BCR-ABL1 transcript ratio (³⁶). All patients who progress to AP or BP CML should have ABL kinase mutations tested (Table 4-7).

When to Switch Therapy

Achievement of CCyR should be expected by 12 months of therapy, especially for standard-dose imatinib, whereas it may be reasonable to expect CCyR for second-generation TKIs within 3 to 6 months of treatment (⁴⁸). Patients who do not achieve a CHR by 3 months should be considered for a change in therapy. If considering a change in therapy at 3 or 6 months for BCR-ABL1 transcript level greater than 10% for patients on imatinib or second-generation TKIs, it is worth noting that very early switching has not yet been shown to influence the long-term outcomes (⁴⁹). As such, it can be advocated that if the transcript level at 3 months is greater than 10%, providers should perform serial molecular monitoring between 3 and 6 months for definitive treatment response evaluation (³⁷). If patients retain >10% transcript level at 6 months, change in therapy is indicated, because the chance of CCyR would be low. Patients who meet all of the relevant responses by the first 12 months are monitored periodically using FISH and PCR testing, and if there are clear signs of possible relapse or failure, bone marrow examination with conventional cytogenetics and kinase domain sequencing should be performed. Not achieving CCyR by 12 months or any extent of later cytogenetic relapse requires a change in therapy. Fluctuating molecular levels during concurrent CCyR should only prompt closer monitoring and a compliance assessment.

Response definitions recommended by the 2013 European LeukemiaNet guidelines are summarized in Table 4-8. Lack of MMR or CMR should not be interpreted as signal for change in TKI therapy. Achieving CMR allows for possibility of treatment discontinuation, which is only recommended in the setting of a clinical trial.

A subset of patients treated with imatinib may develop resistance. Among patients treated in CP, the rate of resistance is less than 4% per year and decreases after the first 3 years to approximately 0.5% to 1% per year. Following achievement of CCyR, the rate of resistance after year 3 of imatinib therapy is less than 1%, suggesting a durable CCyR on imatinib and the predictability of the CML course once such a response is obtained.

Mechanisms of resistance to imatinib can be BCR-ABL1 dependent or independent. The first, more common group includes amplification or overexpression of BCR-ABL1 or its protein product (⁵⁰) and point mutations of the ABL sequence (⁵¹). The second group includes multidrug-resistance expression and overexpression of Src kinases (⁵²). BCR-ABL1–dependent mutations have been identified in approximately 50% of patients who develop clinical resistance to imatinib (⁵³). More than 90 different mutations with varied significance and imatinib sensitivity have been described in relevant kinase domains, including the ATP-binding domain (P-loop), the catalytic domain, the activation loop, and amino acids that direct interact with imatinib. The “gatekeeper” mutation of particular importance is the T315I, which is resistant to all available TKIs except ponatinib. Most of the clinically relevant mutations develop in a few residues in the P-loop (G250E, Y253F/H, and E255K/V), the contact site (T315I), and the catalytic domain (M351T and F359V) (⁵⁴). A list of mutations following imatinib resistance with their half-maximal inhibitory concentration values is shown in Table 4-5 (²³). The long-term outcome of patients with CML treated with second-generation TKIs after imatinib failure is predicted by the in vitro sensitivity of BCR-ABL1 kinase domain mutations (⁵⁵). It is unclear if the identification of small mutated clones is clinically relevant (^56,57).

Although the sequential use of kinase inhibitor therapy often rescues a response, it can also result in further gain of mutations by the same (compound mutations) or different (polyclonal mutations) Ph clones, which represents a significant hurdle in the treatment of CML (⁵⁸). High molecular dynamics in particular yield compound BCR-ABL1 kinase domain mutations (polymutants), which represent two or more codon changes in the same BCR-ABL mRNA transcript (⁵⁹). Ultra-deep sequencing to resolve qualitative and quantitative complexity of mutated populations surviving TKIs has recently suggested that conventional Sanger sequencing might be inadequate for this evaluation (⁶⁰), although a recent report argues that many BCR-ABL1 compound mutations may actually be artifacts due to PCR-mediated recombination (⁶¹).

Before labeling a patient as having TKI resistance and modifying therapy, treatment compliance and drug-drug interactions must be evaluated. Lower adherence rates observed more commonly in younger individuals, patients with adverse effects on treatment, and patients who have required dose escalations correlate with worse outcomes (⁶²). Mutation analysis should be carried out only in instances of failure criteria.

Imatinib Dose Escalation

Imatinib escalation was the main option for managing suboptimal responses and treatment failures before the era of second-generation TKIs. In a phase II study that reported a 2-year follow-up after high-dose imatinib (400 mg twice daily; n = 49) for patients with CP CML and resistance to imatinib at doses from 400 to 600 mg, the major cytogenetic response (MCyR), CCyR, and MMR were 33%, 18%, and 12%, respectively, with an estimated PFS of 65% at 2 years (⁶³). In another dose-escalation study from the University of Texas MD Anderson Cancer Center, 84 patients with CP CML were dose escalated to imatinib 600 to 800 mg/d after developing hematologic failure (n = 21) or cytogenetic failure (n = 63) to standard-dose imatinib (⁵⁵). Among patients who met the criteria for cytogenetic failure, 75% (47 of 63 patients) responded to imatinib dose escalation. In contrast, in patients in whom imatinib was dose escalated because of hematologic failure, 48% achieved a CHR and only 14% (3 of 21 patients) achieved a cytogenetic response. Patients more likely to respond to imatinib dose increase are those who have previously achieved a cytogenetic response and then lost it and who have not developed any mutations unresponsive to imatinib. With mutations, a switch to a second-generation TKI is preferable.

Dasatinib was first approved by the US Food and Drug Administration at a dose of 70 mg orally twice daily based on its efficacy and safety in a series of phase II trials in patients with all stages of CML (and those with Ph-positive ALL) who were resistant to, or intolerant of, imatinib (^64,65,66). Over 50% of patients treated with dasatinib in CP after imatinib failure achieved a CCyR. Responses to dasatinib, nilotinib, bosutinib, and ponatinib among patients in CP, AP, and BP after imatinib failure or resistance are summarized in Table 4-9.

A randomized trial of dasatinib versus higher dose imatinib (800 mg daily) among patients who failed prior therapy with imatinib (400-600 mg) showed a significantly higher rate of response and PFS for patients receiving dasatinib, particularly among those who were already receiving imatinib 600 mg, those with BCR-ABL1 mutations, and those who had never achieved a cytogenetic response with imatinib, establishing second-generation TKIs as the preferred approach after failure to imatinib standard-dose therapy (⁶³).

In a long-term, 6-year follow-up of a phase III trial of 670 CP CML patients who were resistant (74%) or intolerant (24%) to imatinib, patients were randomized to receive dasatinib 100 mg once daily, 50 mg twice daily, 140 mg once daily, or 70 mg twice daily (⁶⁷). At 6 years, PFS rates were 49%, 51%, 50%, and 47%, respectively, and OS rates were 71%, 74%, 77%, and 70%, respectively. The 6-year MMR rates were 43% and 40% in patients treated in the 100 mg once daily arm and all other arms combined, respectively. Dasatinib 100 mg once daily retained high activity and was associated with less toxicity, particularly pleural effusions and myelosuppression (grade 3 or 4 neutropenia or thrombocytopenia, 30% each), and the lowest rate of drug discontinuation for toxicity. Based on these results, the standard dasatinib dose for patients in CP became 100 mg daily. Similar results established 140 mg once daily as the preferred dose in advanced-stage disease.

In a phase II study, 321 CP CML patients who were resistant or intolerant to imatinib were treated with nilotinib 400 mg twice daily (⁶⁸). At the 48-month follow-up, 45% of the patients achieved CCyR; the PFS and OS rates were 57% and 78%, respectively (⁶⁹). Deeper levels of molecular responses at 3 and 6 months correlated with improved long-term outcomes. In the expanded access ENACT trial, 1,422 patients in CML CP or AP after imatinib failure were treated with nilotinib 400 mg twice daily (⁷⁰). After a median follow-up of 18 months, the CCyR rate was 50% and the PFS rate was 80%.

Bosutinib

Bosutinib (SKI606) is an orally available dual Src/Abl inhibitor that is 30 to 50 times more potent than imatinib against unmutated BCR-ABL1. It has activity against most imatinib-resistant BCR-ABL1 mutants with the exception of T315I. In contrast to other available TKIs, bosutinib has minimal inhibitory activity against C-Kit and PDGFR. It was initially studied in patients resistant or intolerant to imatinib (⁷¹). Among 288 patients treated in a phase I/II trial, more than two-thirds of the patients had imatinib-resistant CML. The primary end point, MCyR at 6 months, was achieved in 31%. Overall, 41% of patients achieved a CCyR, and 64% of them achieved MMR. The 2-year PFS and OS rates were 79% and 92%, respectively. Responses to bosutinib in CP, AP, and BP after imatinib resistance or failure are summarized in Table 4-9.

Bosutinib also was assessed as third- or fourth-line therapy in 118 patients with CP CML who had been previously treated with imatinib followed by nilotinib and/or dasatinib (⁷²). Bosutinib resulted in a CCyR rate of 24%. The 24-month PFS and OS rates were 73% and 83%, respectively. Clinical efficacy in the relapse setting led to the evaluation of bosutinib as frontline therapy.

Treatment with bosutinib has been generally well tolerated with no pleural effusions and modest myelosuppression. The most common adverse events were gastrointestinal (nausea, vomiting, diarrhea) and were usually grade 1 or 2, manageable, and transient and diminished in frequency and severity after the first 3 to 4 weeks of treatment (⁷²). A multicenter phase III randomized trial is ongoing to compare a lower dose (400 mg daily) of bosutinib and standard-dose imatinib (400 mg) in patients with newly diagnosed CP CML.

Ponatinib

Ponatinib (formerly AP24534) is a rationally designed third-generation TKI that efficiently inhibits Bcr-Abl as well as FLT3, PDGFR, VEGF, and C-KIT (⁷³). It is more than 500 times more potent than imatinib at inhibiting BCR-ABL1 and is the first compound in the class to inhibit T315I mutation (⁷³). Ponatinib was approved following the phase II PACE trial, in which 449 patients with heavily pretreated CML or Ph-positive ALL, resistant to or intolerant to dasatinib or nilotinib or with the T315I mutation, were enrolled (⁷⁴). The dose of ponatinib was 45 mg once daily. Among the 267 patients who received ponatinib in CP, 56% achieved an MCyR by 12 months (MCyR rate of 70% in patients with a T315I mutation). Patients responded more favorably if they had received fewer TKIs. After a median follow-up of 3.5 years, 59% of patients achieved MCyR; 83% of those remained in MCyR at 3 years. Furthermore, 39% of patients achieved an MMR or better. The 3-year PFS and OS rates were 60% and 81%, respectively (⁷⁵). Arterial occlusive events occurred in 28% of patients (23% serious). The most common all-grade treatment-emergent adverse events were abdominal pain (46%), rash (46%), thrombocytopenia (45%), headache (43%), constipation (41%), and dry skin (41%) (⁷⁵). Grade 3 or 4 toxicities included myelosuppression (48% overall, observed less commonly in CP CML patients), hepatotoxicity (8%), pancreatitis (5%), hemorrhagic events associated mostly with grade 4 thrombocytopenia (5%), treatment-emergent symptomatic hypertension (2%), neuropathy (2%), and cranial neuropathy (<1%).

As of early 2014, ponatinib labeling included a revised warning regarding risk of thrombotic events (13% per year), vascular occlusions, heart failure, and hepatotoxicity; revised dosing information; and indications limited to adults with T315I mutation and those for whom no other TKI is indicated (⁷⁶). Vascular occlusion adverse events were more frequent with increasing age and in patients with prior history of ischemia, hypertension, diabetes, or hyperlipidemia (⁷⁷). Factors associated with increased risk of vascular occlusion events include older age, higher dose, history of myocardial infarction or prior vascular events, and longer duration of CML (⁷⁸).

Choosing a Second- or Third-Line Option

When faced with treatment failure, a bone marrow with cytogenetic assessment should be completed to determine disease phase and possible clonal evolution. Mutational testing for BCR-ABL1 kinase domain mutations should also be performed, because it helps guide the decision on TKI selection. Dasatinib, nilotinib, and bosutinib are effective against most mutations known to elicit resistance to imatinib (^71,79). For dasatinib, bosutinib, and nilotinib, in vitro and in vivo data have identified mutations that have differential responses to different agents: dasatinib and bosutinib perform better with Y253H, E255K/V, or F359C/V mutations, whereas nilotinib has activity with V299L and F317L mutations, which confer resistance to dasatinib and bosutinib (^80,81). If mutational information is not available, one can resort to considerations regarding toxicity. Bosutinib is a valid choice for patients who fail imatinib and who have pulmonary and vascular risk factors. Ponatinib should be considered the agent of choice for any patient with T315I mutation or compound mutations, in patients who have failed prior second-generation TKIs, and in patients with advanced-phase disease. Vascular thrombotic events represent a serious risk but are likely outweighed by the risks of T315I-mutated disease, for which effective treatment options are scarce.

Omacetaxine mepesuccinate is a cephalotaxine ester and a derivative of homoharringtonine that has a mechanism of action independent of tyrosine kinase inhibition. It is a multitarget protein synthesis inhibitor with an excellent bioavailability through the subcutaneous route. Unaffected by the presence of mutations, it has been in clinical development for several years with activity against CML (⁸²). It has recently also been found to affect the leukemic stem cell compartment, making it an attractive option for the potential total elimination of the leukemic burden and potential cure.

Omacetaxine was approved in the United States for the treatment of patients with CP or AP CML with resistance and/or an intolerance to two or more TKIs on the basis of pooled data from two phase II, open-label, international, multicenter studies (CML-202 and CML-203), in which patients were treated with omacetaxine 1.25 mg/m² twice daily for 14 days every 28 days until response, followed by maintenance for 7 days every 28 days (^83,84). Initial results showed that 20% of patients with CP CML achieved a durable MCyR with a median response duration of 17.7 months (⁸³) and 27% of patients with AP CML achieved a major hematologic response (MHR) that was maintained for a median of 9 months (⁸⁴). After a minimum follow-up of 24 months, 18% of patients in CP CML achieved a MCyR with a median duration of 12.5 months and 14% of patients in AP CML achieved or maintained an MHR for a median of 4.7 months (MCyR was not achieved); median OS times for CP CML (n = 50) and AP CML (n = 14) patients who received more than three cycles of treatment were 49 and 25 months, respectively (⁸⁵). Grade 3 or higher hematologic toxicities (including thrombocytopenia, anemia, or neutropenia) were the major side effects (79% and 73% for CP CML and AP CML, respectively), with discontinuation due to toxicity in 10% of CP and 5% of AP patients. Further analyses of CP and AP CML patients with the T315I mutation (n = 16 and n = 2, respectively), showed that three CP patients achieved MCyR, whereas one AP patient achieved no evidence of leukemia (⁸⁵).

After the successful introduction of TKIs, there has been a paradigm shift in the approach to CML therapy. An allogeneic stem cell transplant (ASCT) is no longer recommended as a first-line therapy and is instead reserved as a third-line or later strategy. It still has an important role in patients who evolve to AP or BP. Transplantation carries significant risks of acute and chronic graft-versus-host disease (GVHD), veno-occlusive disease, life-threatening infections, secondary malignancy, and poorer overall quality of life, although recent advances in the field have significantly improved some of these risks. Current recommendations for an ASCT are restricted to patients who are in AP or BP CML, those in CP who have failed at least two TKIs and acquired compound mutations, and patients harboring the T315I mutation after a trial of ponatinib therapy (⁸⁶). Prior exposure to TKIs does not have a negative impact on the transplant outcome; in fact, patients referred to transplant may have a better outcome if undergoing transplant with less CML burden (⁸⁷). Recommendations for the role and timing of ASCT CML are outlined in the Table 4-10 (⁷⁶).

Few studies have addressed the issue of discontinuing TKI therapy. The Stop Imatinib (STIM) trial evaluated the risk of relapse in patients who stopped treatment with imatinib after being in CMR (MR^4.5) for longer than 2 years. At the most recent follow-up of 50 months, among 100 patients monitored, 61% developed molecular relapse, with most of these events occurring within 7 months of imatinib discontinuation. Nearly all patients regained CMR on retreatment with imatinib (one patient lost CCyR and was treated with dasatinib) (^88,89). This study suggested that approximately 40% of patients with durable CMR might be cured with TKI alone. Overall, low-risk Sokal score and duration of imatinib therapy greater than 60 months predicted for preservation of CMR following discontinuation of therapy.

The TWISTER study followed 40 patients who stopped imatinib after greater than 2 years of undetectable MRD (MR^4.5) (⁹⁰). During the median follow-up of 43 months (minimum, 15 months), 22 patients (55%) became molecularly positive, and nearly 70% of molecular relapses occurred within the first 6 months of treatment cessation. Following the resumption of TKIs, patients regained molecular responses. None of the relapsing patients developed kinase domain mutation, AP, or BP.

A pan-European Stop Tyrosine Kinase Inhibitor trial (EURO-SKI study) aimed to define factors associated with durable deep MR after stopping TKI. An interim analysis of 200 patients with 6-month follow-up of molecular events was reported (⁹¹). Adults in CP CML on TKI treatment in confirmed deep molecular response (molecular response with a 4-log reduction in BCR-ABL transcripts from baseline [MR⁴]; BCR-ABL <0.01%) for at least 1 year (>4 log reduction on TKI therapy for >12 months confirmed by three consecutive PCR tests) and on TKI treatment for at least 3 years were eligible. Median duration of TKI treatment was 8 years (range, 3-12.6 years), and median duration of MR⁴ before TKI cessation was 5.4 years (range, 1-11.7 years). Overall, 123 of the 200 patients remained without relapse in the first 6 months. Recurrence of CML, defined as loss of MMR, was observed in 47 (47%) of 114 patients treated for <8 years, as compared to 27 (27%) of 86 patients treated for >8 years (P = .003). The duration of MR⁴ >5 years versus <5 years was predictive for a lower relapse rate (P = .03).

The feasibility of second-generation TKI discontinuation has also been tested in the French STOP 2G-TKI study for patients treated with nilotinib or dasatinib as frontline therapy or after imatinib failure or intolerance (⁹²). Twenty-four months of persistent MR^4.5 was a requirement for discontinuation of TKI. Interim outcome of 52 patients with a median follow-up of 32 months (range, 12-56 months) reported treatment-free survival of 54% (majority of relapses occurred early, with a median duration of 4 months).

At present, given the uncertainty in selecting the best candidate patients for discontinuation of TKI therapy and the lack of long-term follow up, TKI discontinuation should not be recommended outside the context of a clinical trial.

Patients in AP or BP CML may receive initial therapy with newer generation TKIs (preferred over imatinib) to reduce the disease burden and may be considered for early ASCT (^93,94,95). Nonlymphoid BP response rate is 40% with a combination of TKIs and chemotherapy, with a median OS of 6 to 12 months. With TKI and antilymphoid therapy, the response rate in lymphoid BP is 70% to 80%, and the median OS is 12 to 24 months (^96,97). Tyrosine kinase inhibitors provide hematologic responses in 80% of patients, with an estimated 4-year OS of 40% to 55% in AP, but only a 40% response rate with a median OS of 9 to 12 months in BP. De novo AP responds better to frontline TKI therapy compared to AP transformed from antecedent CP, with a 6- to 8-year OS on TKI therapy of 60% to 80% (⁹⁸). These patients may continue on TKI therapy indefinitely, provided they attain CCyR. Currently, the only curative therapy for AP or BP CML is ASCT, with a cure rate of 15% to 40% for AP and 5% to 20% for BP (³⁷). Patients in both phases should be encouraged to participate in clinical trials.

Chronic Phase

Standard-of-care, category 1 recommendation for the treatment of patients with newly diagnosed CP CML includes any of the three TKIs: imatinib, dasatinib, or nilotinib. Comorbidities, disease risk factors, side effect profiles, and cost may help in choosing one over another. Second-generation TKIs have demonstrated higher rates of early optimal responses; however, their impact on long-term OS remains to be evaluated. Achievement of CCyR is the primary goal of TKI therapy, and it may be desired by 3 to 6 months of therapy with dasatinib or nilotinib. Patients should be carefully monitored with an emphasis on compliance as well as treatment of side effects to minimize unnecessary treatment interruptions and dose reductions. Allogeneic stem cell transplantation or other chemotherapy agents are no longer recommended as frontline treatments given the excellent responses and long-term OS achieved with TKIs. Kinase domain mutation profile is of relevance in the CML cytogenetic or hematologic relapse setting only and should be considered for patients who are failing imatinib or second-generation TKIs or who progress to AP or BP.

Accelerated and Blast Phase

Newer generation TKIs are preferred over imatinib as the frontline therapy. Allogeneic stem-cell transplantation, ideally in second CP, should be considered early for all AP patients based on response to TKI therapy. De novo AP may be treated long-term with TKI therapy alone provided CCyR is achieved. Tyrosine kinase inhibitor monotherapy or combination with chemotherapy (hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone [hyper-CVAD] for lymphoid, AML-type for myeloid phenotype) may be considered for patients who are poor transplant candidates or as a bridge to ASCT.

In 2016, multiple TKIs are available for the treatment of CML, including imatinib, dasatinib, nilotinib, bosutinib, and ponatinib, in addition to omacetaxine and other older therapeutic agents. Following the dramatic refinement of therapeutic options for CML over the last 15 years, the majority of patients with CML are expected to have a normal life expectancy, provided that compliance with therapy is maintained and optimal monitoring is closely implemented, modifying therapy for early signs of resistance or treatment failure. The prevalence of CML will continue to increase over the next two decades. This will offer an opportunity to celebrate this success, but also will be a burden in terms of potential long-term side effects and costs of care. Future efforts will address improving complete molecular eradication of CML and achieving durable CMRs, even after therapy discontinuation (molecular cure). New-generation TKIs in novel combinations with available agents (eg, omacetaxine, decitabine, pegylated IFN) or with new investigational therapies (eg, JAK2 inhibitors, hedgehog inhibitors, stem cell toxins or vaccines, bcl2 inhibitors, other immune approaches) are likely to play a key role in the ultimate cure of CML.