Measuring response to BCR-ABL inhibitors in chronic myeloid leukemia


  • Jerald P. Radich MD

    Corresponding author
    1. Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, Washington
    • Clinical Research Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, D4-100, Seattle, WA 98109-1024

    Search for more papers by this author
    • Fax: (206) 667-2917


In patients with chronic myeloid leukemia (CML), the hallmark Philadelphia chromosome is the marker of disease that can be detected by conventional metaphase cytogenetics, fluorescence in situ hybridization, or polymerase chain reaction. The current “gold standard” of treatment response is cytogenetic response. Cytogenetic response to imatinib is strongly associated with disease progression and survival. Various strategies aimed at improving cytogenetic response have been explored, such as escalation of imatinib and switching to the newer breakpoint cluster region/v-abl Abelson murine leukemia viral oncogene (BCR-ABL) inhibitors dasatinib and nilotinib. Data from recent randomized trials of dasatinib and nilotinib as first-line therapy of newly diagnosed chronic-phase CML suggest that these agents are more effective than imatinib in achieving 6-month and 12-month complete cytogenetic responses. However, it is still too early to know whether or not this early response will translate into a long-term survival advantage. In addition, more sensitive assays to detect residual disease also may be associated with improved long-term outcomes. The deepest measure of response—a complete molecular response—may help identify patients who can stop taking imatinib for the short term, although the long-term consequences of this strategy remain unknown. Cancer 2011;. © 2011 American Cancer Society.


Chronic myeloid leukemia (CML) is a clonal myeloproliferative disease characterized by the oncogenic Philadelphia chromosome, which is formed by a reciprocal translocation between chromosomes 9 and 22.1, 2 This translocation causes 5′ portions of the breakpoint cluster region (BCR) gene from chromosome 22 to juxtapose with the 3′ tyrosine kinase domains of the v-abl Abelson murine leukemia viral oncogene (ABL) from chromosome 9. The novel chimeric protein BCR-ABL dictates the pathophysiology of CML.3-6

CML accounts for approximately 20% of all adult cases of leukemia.7 The age-adjusted incidence rate in the United States is 1.5 per 100,000 population per year, and the median age at diagnosis is 66 years.8 Clinically, CML is divided into 3 phases, although molecular data suggest that the disease actually comprises 2 phases: chronic phase (CP and advanced phase (AP).9 The initial CP often has a relatively benign presentation with an expansion of normal myeloid cells; however, if left untreated, unopposed BCR-ABL activity will foster genetic instability, causing genetic changes that lead to advanced-phase disease with an AP and a blast phase (BP).10 The treatment of advanced-phase disease is woefully inadequate compared with that for CP, no matter which therapy is used.

The treatment of CML with the BCR-ABL inhibitor imatinib has revolutionized management of the disease, relegating the role of allogeneic bone marrow transplantation to that of salvage therapy. The more potent and newer BCR-ABL inhibitors, nilotinib and dasatinib, were approved first to treat patients with imatinib-resistant or imatinib-intolerant disease, but recent randomized trials have led to approval by the US Food and Drug Administration for nilotinib and dasatinib as first-line therapy. The National Comprehensive Cancer Network (NCCN) guidelines recommend both dasatinib and nilotinib as first-line therapy options.11 Given this new embarrassment of riches, how do we decide which BCR-ABL inhibitor to start patients on and when to switch?

Measuring Response: Tools and Endpoints

The earliest and most easily measured disease response is normalization of the peripheral white blood cell (WBC) count or complete hematologic remission (CHR), which generally occurs within 3 months in the vast majority (>90%) of newly diagnosed patients with CML-CP who receive treatment with a BCR-ABL inhibitor. The next disease-response level hinges on the cytogenetic response from a bone marrow sample. Cytogenetic response currently is considered the “gold standard” of treatment response. Typically, cytogenetic responses are divided into several categories, which usually are determined by metaphase chromosome analysis (using ≥20 metaphases) of the bone marrow (Table 1).12 Responses range from none; through minimal, minor, and partial; and up to a complete cytogenetic response (CCyR). In addition, the category of major cytogenetic response (MCyR) includes both CCyR and partial cytogenetic responses (PCyR). Cytogenetic testing also can detect secondary clonal evolution, which is indicative of disease progression to accelerated phase. Although fluorescence in situ hybridization on blood cells also can be used to assess the level of cytogenetic response, it is not a recommended replacement for cytogenetic testing of bone marrow, because it does not detect secondary cytogenetic abnormalities. However, fluorescence in situ hybridization is an acceptable alternative if insufficient numbers of bone marrow cell metaphases are available for analysis or if bone marrow sampling is not possible.12 The most sensitive tool for assessing disease status is real-time quantitative polymerase chain reaction (QPCR), which measures the fusion BCR-ABL messenger RNA transcript. Not only is it exquisitely sensitive, but QPCR also has the added advantage of analyzing peripheral blood rather than a bone marrow specimen.

Table 1. Definitions of Cytogenetic Responsea
Degree of Response% Ph+ Metaphases Detected
  • Abbreviations: Ph+, Philadelphia chromosome positive.

  • a

    Adapted from Baccarani 2009.12

  • b

    Complete and partial responses are often grouped under the category “major” (see National Comprehensive Cancer Network 201011).


Cytogenetic Response to Imatinib Therapy

A treatment recommendation for newly diagnosed CML-CP is oral imatinib 400 mg once daily.11-13 Substantial data justify this recommendation. The International Randomized Study of Interferon versus STI571 (IRIS), a large phase 3 randomized trial (N = 553) in patients with CML-CP, demonstrated a superior 18-month CCyR rate (76% of patients who received imatinib vs 15% of those who received interferon plus cytarabine14). An 8-year follow-up study indicated rates of event-free survival (EFS) and freedom from progression to advanced disease of 81% and 92%, respectively.15 It should be noted, however, that, in the definition of EFS in the IRIS trial, discontinuation of therapy was not counted as an event. After 8 years of follow-up, 55% of patients remained on imatinib therapy.15 Thus, clinical use surveys and population studies have indicated that resistance to or intolerance of imatinib prevents many patients from gaining the full benefit of treatment; CCyR rates in patients who are treated in nontrial settings appear to be lower than in the controlled setting of clinical studies.16-19 After 8 years of follow-up, 6% of patients had discontinued because of adverse events or other safety reasons, 16% had discontinued because of an unsatisfactory therapeutic outcome, 3% had undergone stem cell transplantation, 3% had died, and 17% had discontinued for other reasons.15

Data from numerous imatinib trials have established CCyR as an important predictor of progression-free survival (PFS) and overall survival (OS). Five-year follow-up and data analysis from the IRIS trial indicated that patients who achieved a CCyR at 12 months were more likely than others not to progress to AP or BP by 60 months (97% vs 81%).20 Additional analyses revealed that patients who achieved a CCyR also had superior OS compared with those who did not reach this milestone.21 Six-year follow-up data and analysis from the same trial also suggested that achieving a CCyR within 6 months of the initiation of therapy may be an important predictive factor for PFS.22 An analysis of IRIS data after a minimum follow-up of 8 years indicates that patients who achieved less than a CCyR at 18 months had a probability of only 29% of achieving a CCyR without a subsequent event (defined as death, transformation to advanced disease, loss of CHR or MCyR, or increased WBC count to >20 × 109/L).15

Five-year data from the Hammersmith group demonstrated similar results to those from the IRIS study: Patients who achieved a CCyR at 12 months demonstrated a benefit in terms of both PFS (with progression defined as the development of advanced disease) and OS compared with patients who did not achieve a CCyR at 12 months (PFS, 96% vs 74%; OS, 98% vs 74%).18 Other studies also have demonstrated a significant association between CCyR at 12 months and improved PFS.23, 24

Indicators of Response: The More Sensitive, the Better?

The IRIS study investigators defined a major molecular response (MMR) as a 3-log reduction in the BCR-ABL messenger RNA level compared with levels in an aggregate of baseline samples.25 Several studies have borne out the positive clinical significance of achieving an MMR, although a few others have not. A 12-month landmark analysis of the IRIS trial demonstrated that the 2-year PFS rate was 100% in patients who achieved a CCyR and an MMR, 95% in patients who achieved a CCyR but no MMR, and 85% in patients who achieved neither a CCyR nor an MMR (with progression defined as death, the development of advanced CML, increasing WBC count, or loss of CHR or MCyR in the initial publication).25 In a study of 90 patients who achieved a CCyR (median follow-up, 49 months), it was observed that relapse-free survival was significantly shorter in patients who did not achieve an MMR than in those who attained this landmark (median time not reached).26 In addition, an analysis of follow-up data from a phase 2 study suggested that the estimated 6-year EFS rate was significantly higher in patients who achieved both an MMR and a CCyR (100%) than in those who achieved only a CCyR (44%; P = .0001).27 However, an 18-month landmark analysis of the IRIS study indicated that 5-year PFS did not differ significantly between patients with both CCyR and MMR (100%) and those with CCyR only (98%; with progression defined as the development of advanced CML).20 Moreover, in a retrospective analysis of patients with CML-CP who received imatinib (n = 276), landmark analyses at 6 months, 12 months, 18 months, and 24 months failed to identify any significant association between OS and achieving an MMR while in CCyR. Durable MMRs (documented continuously for >12 months) were associated with longer PFS (defined as loss of hematologic or cytogenetic response, death, or development of advanced CML) but not with longer OS.28

A recent follow-up of the IRIS trial29 suggested that achieving an MMR was associated with superior EFS than achieving a CCyR but without an MMR. Serial molecular studies were evaluated based on the International Scale, in which a value of ≤0.1% represents MMR and values from >0.1% to 1% represent CCyR.13 Patients who had BCR-ABL transcripts >1% at 12 months had inferior EFS and higher rates of progression to AP/BP compared with patients who had lower levels of molecular response. Patients who achieved an MMR at 18 months had no progressions to AP/BP and had a 95% EFS rate at 7 years. The probability of losing a CCyR was 3% for patients who achieved an MMR at 18 months compared with 26% for patients who achieved a CCyR but no MMR. Similar near absence of progression in patients who achieved an MMR also has been reported in front-line studies of nilotinib30 and dasatinib31 and in US and Canada Intergroup trial S0325 comparing dasatinib and imatinib.32

Thus, the majority of the data for MMR point to a “safe haven” with a very low risk of progression to advanced-phase disease or loss of CCyR. It is not surprising that achieving an MMR is not clearly associated with superior OS compared with achieving a CCyR. Fortunately, patients who fail initial imatinib therapy can be salvaged with a newer BCR-ABL inhibitor and, thus, can survive despite initially failing front-line therapy.

Currently, there is an effort to encourage all laboratories to adopt the International Scale to standardize QPCR results because of issues with interlaboratory variability.33 Without this scale, interpreting molecular data is more difficult, although it is not impossible if the laboratory 1) fully understands the characteristics of their test and 2) has an idea of what a baseline test result is in the aggregate of patients with newly diagnosed CML who are referred for their assay. Without these parameters, physicians and patients will struggle with the meaning of the molecular test and certainly will have difficulty in judging where they stand with regard to MMR.

The achievement of MMR has been associated with the eventual achievement of a complete molecular response (CMR).34 Thus, patients who achieved an MMR by 12 months had a cumulative incidence of subsequently obtaining a CMR of approximately 75% compared with <10% for patients who were not in MMR. CMR has become an interesting endpoint, because it appears that some patients in prolonged CMR may discontinue therapy and remain in CMR. A study by Mahon et al assessed imatinib discontinuation in patients who had maintained a CMR for at least 2 years. Those investigators observed that 61% of patients relapsed after discontinuing imatinib therapy (most within 6 months), but all could be retreated with response, although not all patients achieved a second CMR. At a median follow-up of 17 months, 39% of those patients remained off BCR-ABL inhibitor therapy.35

CMR is obviously an attractive endpoint for both patients and physicians, but several caveats should be noted. First, with current therapy, CMR is relatively unusual, occurring in <5% of patients who receive treatment with imatinib at 12 months.25, 30 Although this prevalence may increase with time, it still is expected that few patients will attain a persistent CMR. Second, although patients who relapse have been retreated effectively, it should be emphasized that, even in CMR, there likely resides a pool of CML “stem cells.” Because these cells are spared BCR-ABL inhibition, there may be clones that begin to take on genetic changes, which put them on the pathway to progression. If so, then patients who appear to be doing well with retreatment may be at risk several years down the line. Thus, discontinuation of BCR-ABL inhibitors should only be done in the context of a clinical trial. The third issue concerning CMR is the evolving concept of what it actually means. If a CMR is defined as an “undetectable” BCR-ABL signal, then this must be interpreted in the context of the sensitivity of the assay and the adequacy of the sample: That is, an undetectable test can occur from a poor assay (poor sensitivity) or a poor sample (not enough cells sampled).

How Should Patients Be Monitored in Routine Practice?

The European LeukemiaNet (ELN) and the NCCN have developed guidelines and recommendations for the treatment and monitoring of patients with CML; these include criteria defining optimal responses, suboptimal responses, and failure of response to imatinib treatment.11-13 The “optimal” response to imatinib therapy is attaining a CCyR after 12 months of treatment.12, 18, 20, 21, 24 Treatment failure calls for a change in therapy.12 According to ELN and NCCN criteria, a patient who fails to achieve either a CHR by 3 months, or any cytogenetic response by 6 months, or an MCyR by 12 months, or a CCyR by 18 months of therapy is deemed an imatinib failure. Likewise, loss of cytogenetic response also necessitates a change in treatment and appropriate mutation analysis.12

To monitor therapeutic responses over time, cytogenetic assessment should be performed at diagnosis, 3 months and 6 months after treatment initiation, and every 6 months thereafter until a CCyR is achieved.12 The need for further cytogenetic testing after the achievement of a CCyR is under debate. The ELN prefers yearly cytogenetic assessment; new clonal abnormalities in Philadelphia chromosome-negative cells may occur in some patients (approximately 5%) in CCyR.13 Most of these clonal abnormalities involve trisomy 8 and appear to be benign.13 However, some cases involving additional clonal abnormalities have the cytogenetic and clinical characteristics of a secondary myelodysplastic syndrome or acute myeloid leukemia.36-38 The NCCN recommends that mutational analysis would be useful to identify a subset of patients at increased risk of progression as a result of cytogenetic mutation.11 Furthermore, if there is a persistent and unexplained drop in blood counts, then the NCCN guidelines recommend bone marrow cytogenetic analysis to rule out non-Philadelphia chromosome-related chromosomal abnormalities or myelodysplasia.11

Imatinib Resistance and the Importance of Prolonged Response

Although it is clear that BCR-ABL inhibitor treatment of newly diagnosed CML-CP radically changes the natural history of the disease, it is unclear whether a “cure” is a reasonable expectation. Thus, the more immediate and useful clinical goal is to prevent transformation to CML-AP/BP as long as possible, because advanced disease carries such a dire prognosis (Table 2).14, 39-48

Table 2. Complete Cytogenetic Response Rates for Approved Tyrosine Kinase Inhibitorsa
 Complete Cytogenetic Response Rates: Phase of CML (Reference)
CompoundCP, %AP, %BP, %
  • Abbreviations: AP, accelerated phase; BP, blast phase; CML, chronic myeloid leukemia; CP, chronic phase.

  • a

    In key trials across different phases of CML.

  • b

    Phase 3 trial.

  • c

    Phase 2 trial.

First-line treatment   
 Imatinib74 (O'Brien 200314)b21 (Kantarjian 200944)c10 (Saglio 200843)c
 Dasatinib98 (Kantarjian 200948)b  
 Nilotinib98 (Mauro 200649)b  
Second-line treatment   
 Dasatinib50 (Palandri 200942)b32 (Cortes 201046)b14-38 (Saglio 201030)b
 Nilotinib46 (Branford 200350)c20 (Cortes 201047)c29 (Palandri 200841)c

The chief obstacle to achieving prolonged responses to BCR-ABL inhibitor therapy is resistance. Primary (intrinsic) resistance is defined as failure to achieve the landmark outcomes (CHR and CCyR), whereas secondary (acquired) resistance is defined as loss of therapeutic response.49

Development of resistance is a major concern in the treatment of CML. An intent-to-treat analysis revealed that 26% of patients discontinued imatinib after a median follow-up of 15.5 months because of lack of efficacy or progressive disease.18 A survey of clinical use reported disease progression and/or loss of response to imatinib in 12% of patients with CML, and a population study indicated that approximately 50% of patients with newly diagnosed CML-CP who were receiving imatinib failed to achieve or maintain a CCyR and/or progressed to advanced disease.17, 19

Targeting Resistance

The most common mechanism of imatinib resistance is the acquisition of a point mutation in the ABL tyrosine kinase domain. BCR-ABL mutations occur in approximately 50% of patients who develop resistance to imatinib.50, 51 The phosphate-binding loop (P-loop) (residues 248-256) is the most common mutation site (Table 3).52, 53 Certain P-loop mutations confer a high level of insensitivity to imatinib and are associated with a poor prognosis, although some of these patients can benefit from the use of new BCR-ABL inhibitors.50, 54-57 Studies conducted before the availability of the newer BCR-ABL inhibitors suggested that patients with P-loop mutations had poorer outcomes, with rapid progression to advanced CML disease.50, 55, 56 The most worrisome mutation is the threonine-to-isoleucine mutation at codon 315 (T315I), because none of the currently available BCR-ABL inhibitors blocks its ABL activity.7, 50, 57 Patients who harbor this mutation should be offered stem cell transplantation or entry into a clinical trial of a novel agent.12 Other mechanisms that may affect imatinib sensitivity include altered expression of drug influx and efflux proteins (ie, phosphoglycolate phosphatase [Pgp] and octamer-binding protein [OCT-1]), pathologic activation of SRC family kinases, clonal evolution, and over expression of BCR-ABL.7, 57-62

Table 3. Most Frequent Breakpoint Cluster Region/V-Abl Abelson Murine Leukemia Viral Oncogene (BCR-ABL) Mutations After Imatinib Treatment and Their Sensitivity to BCR-ABL Inhibitorsa
  IC50-Fold Increase Over Native BCR-ABL
  • Abbreviations: A-loop, aromatic residue interacting with the adenine ring of adenosine triphosphate; ATP, adenosine triphosphate; E, glutamic acid; F, phenylalanine; G, glycine; H, histidine; I, isoleucine; IC50, 50% inhibitory concentration; K, lysine; L, leucine; M, methionine; NA, not applicable; P, proline; P-loop, phosphate-binding loop; R, arginine; T, threonine; V, valine; Y, tyrosine.

  • a

    Adapted from Quintas-Cardama & Cortes 2009.53

  • b

    Residues are presented in descending order of frequency of mutation (see Khorashad 200856).

T315IATP-binding region>24.6>153.9>250
M351TCatalytic domain3.41.21.4
F359VCatalytic domain713.52.8
E355GCatalytic domain8.9NA2.3

Second-Generation BCR-ABL Inhibitors and Resistance

Dasatinib and nilotinib were approved first for the treatment of patients with CML who were resistant to or intolerant of first-line therapy with imatinib. Dasatinib and nilotinib are treatment options for patients who experience failure of imatinib therapy either by not achieving a CCyR by 18 months or through loss of a previously attained CCyR.12

Both dasatinib and nilotinib are highly active in patients with CML. A phase 3 trial of second-line dasatinib 100 mg once daily in patients with CML-CP (n = 167) demonstrated a CCyR rate of 50% by 24 months (best response; minimum follow-up, 2 years); 89% of patients who achieved a CCyR within 2 years maintained it at 24 months.40, 63 In another phase 3 second-line trial, 32% of patients with CML-AP who received dasatinib 140 mg once daily (n = 158; median follow-up, 15 months) achieved a CCyR.44

In a phase 2 study of patients who received second-line nilotinib (n = 321), including those with CML-CP, the CCyR rate was 46% at a minimum follow-up of 19 months; 84% of those CCyRs were maintained at 24 months,48 and the 24-month PFS and OS rates were 64% and 87%, respectively (with progression defined as the development of advanced disease, discontinuation because of advanced disease, or death). In patients with CML-AP (n = 137), the CCyR rate after a minimum follow-up of 11 months was 20%; of those responses, 83% were maintained at 12 months.45 A CCyR rate of 29% also has been reported in patients with CML-BP (n = 136).39

Two studies of dasatinib have demonstrated the clinical significance of achieving a CCyR in the second-line treatment of CML-CP. In a phase 3 dose-optimization study, the PFS rate at 36 months was greater in patients who achieved a CCyR at 6 months (93%) than in those who achieved only a PCyR (76%) or less than a PCyR (54%). Similarly, the 36-month PFS rate was greater in patients who achieved a CCyR at 12 months (90%) than in those who achieved only a PCyR (77%) or a lesser cytogenetic response (63%).64 A retrospective study of 1067 patients with CML-CP indicated that those who received dasatinib and achieved a CCyR at 12 months had a higher 24-month PFS rate than patients who did not attain a CCyR or an MMR (97% vs 78%); progression in that analysis was defined based on the phase 3 optimization study (eg, progression to AP/BP, loss of CHR or MCyR, increasing WBC count, or death).65

Mutation Data: What Do They Mean?

The type of BCR-ABL mutation potentially can influence which BCR-ABL inhibitor should be given to patients who fail on imatinib.12, 66, 67 The most common BCR-ABL mutations and their sensitivity to approved BCR-ABL inhibitors are listed in Table 3. It should be noted, however, that sensitivity values in the table were derived from artificial in vitro settings and, thus, may not always correlate with the actual sensitivity to a mutant clone in a given patient. However, there appear to be cases in which particular mutations can influence the decision-making in choosing the best “fit” with the activity of a particular second-generation BCR-ABL inhibitor. For example, in a phase 2 study of second-line nilotinib in imatinib-resistant and imatinib-intolerant patients, there were no patients with the T315I, tyrosine-to-histidine at codon 253 (Y253H), glutamic acid-to-valine at codon 255 (E255V), or phenylalanine-to-glutamine/valine at codon 359 (F359C/V) mutations who achieved a CCyR.68, 69 In that study, the most common mutations associated with disease progression during nilotinib treatment were the glutamic acid-to-lysine/valine at codon 255 (E255K/V), F359C/V, Y253H, and T315I mutations.70 A retrospective analysis of dasatinib clinical data according to mutation status after imatinib failure (n = 1043) indicated that no patients with T315I mutations and only 7% of patients with phenylalanine-to-leucine at codon 317 (F317L) mutations achieved a CCyR.71 In a phase 3 dose-optimizing study of patients with CML-CP, newly developing mutations associated with loss of response to dasatinib were T315I, F317L, V299L, and, rarely, E255K.64 The study by Branford et al66 assessed how often patients had mutations that would specifically direct subsequent, newer BCR-ABL inhibitor therapy. Those authors observed that 43% of resistant patients who harbored mutations carried mutations that were clinically relevant and that 28% carried mutations that were clinically relevant for either dasatinib or nilotinib. It was further estimated from previous data that the percentages of imatinib-resistant patients in CML-CP or CML-AP who carried clinically relevant mutations were approximately 18% and 31%, respectively. In patients with CML-BP, 29% to 52% of patients may carry a clinically relevant mutation.66

Cytogenetic Response, Resistance, and the Role of Transplantation

Allogeneic transplantation is curative in CML; however, given the success of BCR-ABL inhibitor therapy, it is now generally reserved to salvage patients with resistance to imatinib who do not respond to a second-generation BCR-ABL inhibitor, for patients who have the T315I mutation or advanced-phase disease, or for those rare patients who have intolerance to all BCR-ABL inhibitors. The most common clinical question is how to determine which resistant patients should proceed to undergo transplantation: How long should they wait for a response to the second-generation BCR-ABL inhibitor?

A cytogenetic response to second-line therapy is a useful parameter for this particular decision. Two studies have suggested measures of response to guide second-line BCR-ABL inhibitor therapy, allowing physicians to declare failure early and expeditiously proceed to transplantation. These include achieving at least a minimal cytogenetic response after 3 months of therapy with a secondary BCR-ABL inhibitor and an MCyR by 12 months.72, 73 Thus, 1 strategy would be to start a donor search at the first sign of resistance. Because the process of human leukocyte antigen typing and donor identification can take 3 to 6 months, patients who have an inadequate response at that time point can be moved efficiently to transplantation.

Can We Improve Response to First-Line Imatinib?

Several avenues are being explored that may result in increased CCyR rates and durability. One approach is to simply increase the starting dose of imatinib. Several reports suggest that higher dose imatinib (>400 mg daily) produces faster responses but may have no long-term impact on either CCyR rates or survival. Data from 2 randomized, phase 3 trials indicate that patients who received imatinib 800 mg daily achieved a significantly higher CCyR rate at 6 months than those who received imatinib 400 mg daily, but there were no significant differences in CCyR rates at 12 months and beyond.74, 75 The increasing dose of imatinib also increased drug toxicity, leading to more dose interruptions in the higher dose imatinib arm, which could have undermined response.

A second suggested approach in the context of imatinib treatment is to titrate patient doses according to trough plasma drug concentrations. It has been reported that imatinib serum concentrations were significantly lower in patients who did not achieve a CCyR (694 ± 556 ng/mL; n = 12) compared with patients who did (1123 ± 617 ng/mL; n = 56).76 In a subanalysis of patients entered into the IRIS and Tyrosine Kinase Inhibitor Optimization and Selectivity (TOPS) trials (n = 526), patients who had the lowest trough levels of imatinib at day 29 of treatment had inferior rates of CCyR at 12 months.77 However, data from a different study (n = 78) indicated no correlation between cytogenetic response rates and trough imatinib plasma levels at 12 months (CCyR: trough imatinib plasma level, 1010 ng/mL; no CCyR: trough imatinib plasma level, 1175 ng/mL).78 Moreover, low imatinib plasma levels may be a result of patient drug metabolism or poor adherence, and the role of drug monitoring and dose modification in these 2 events probably differs greatly and is problematic to distinguish.

A third approach is to administer imatinib in combination with another compound. One combination partner being explored is interferon-alfa. In the STI571 Prospective Randomized Trial (SPIRIT) (N = 636), the 12-month MMR rate (57% vs 38%; P < .001; unadjusted) and the 24-month MMR rate (64% vs 43%; P = .006; unadjusted) were significantly higher for imatinib 400 mg daily plus pegylated interferon (n = 159) compared with imatinib alone (n = 159). However, 12-month CCyR rates did not differ significantly (66% vs 58%; P > .05; unadjusted). The endpoint of survival without progression to advanced disease within 24 months produced rates of 98% for imatinib 400 mg daily alone and 97% for imatinib 400 mg daily plus pegylated interferon.79 However, 2 other large, prospective studies suggested no efficacy advantage for imatinib plus interferon compared with the same dosage of imatinib alone.80, 81 High interferon discontinuation rates also suggest that clinical applications of imatinib combined with this compound may be limited.81, 82

Finally, patient compliance should be examined in patients who do not achieve optimal responses. In 1 study, adherence to imatinib was assessed in 169 Belgian patients with CML.83 Patients who had ELN-defined suboptimal responses to imatinib had a higher degree of nonadherence than who achieved optimal responses (23% vs 7%; P = .005).83 In a study of patients who received imatinib 400 mg daily for a median of 60 months and who had achieved a CCyR (n = 87), adherence measured over 3 months was an independent predictor of the probability of achieving a 6-year MMR (P = .001).84

Are Second-Generation BCR-ABL Inhibitors as Front-Line Therapy Better Than Imatinib?

The “million-dollar question” is obviously whether the more potent, newer BCR-ABL inhibitors are better than imatinib as front-line therapy for newly diagnosed CML. Three phase 3, randomized trials suggest that they are, at least in very early reports of efficacy. Initially, both dasatinib (100 mg daily) and nilotinib (800 mg daily) were studied in 2 concurrent, phase 2 studies in patients with untreated CML-CP.45, 46, 85, 86 Dasatinib initially was administered in 2 schedules: 50 mg twice daily (n = 31) and 100 mg once daily (n = 31). In the nilotinib trial, all patients received nilotinib 400 mg twice daily (n = 61). Data after a median follow-up of 24 months (dasatinib) and 17 months (nilotinib) indicated that both compounds induced high rates of CCyR, which were higher than those anticipated from historic data from standard-dose or high-dose imatinib therapy (Table 4).45, 46, 85, 86 In a separate phase 2 study of nilotinib in untreated patients with CML-CP, the 24-month rates of CCyR and MMR (92% and 82%, respectively) and the 30-month rate of OS (99%) also suggested high activity for this compound in this setting.87

Table 4. Complete Cytogenetic Response Rates Over Time for First-Line Dasatinib and Nilotinib Therapy Compared With Historic Data for Imatinib
 CCyR Rate by Compound (References), %
Time on Therapy, moDasatinib, 100 mg/d (Cortes 201047; Cortes 200985)aNilotinib, 800 mg/d (Cortes 201046; Cortes 200986)aImatinib, 400 mg/d (Cortes 200985,86)Imatinib, 800 mg/d (Cortes 200985,86)
  • Abbreviations: CCyR, complete cytogenetic response.

  • a

    There were ≥28 patients at each time point for dasatinib and nilotinib.


Dasatinib 100 mg daily and nilotinib 600 mg daily recently have been approved by the US Food and Drug Administration for the first-line treatment of patients with CML-CP on the basis of data from 2 randomized phase 3 comparisons of nilotinib or dasatinib with imatinib. In the Evaluating Nilotinib Efficacy and Safety in Clinical Trials-Newly Diagnosed Patients (ENESTnd) trial, the MMR rate at 12 months was significantly higher in the nilotinib arms compared with imatinib (nilotinib 600 mg daily, 44%; nilotinib 800 mg daily, 43%; imatinib 400 mg daily, 22%; P < .001).30 After a median follow-up of 18 months, cumulative CCyR rates in the nilotinib arms were higher compared with imatinib (nilotinib 600 mg daily, 85%; nilotinib 800 mg daily, 82%; imatinib 400 mg daily, 74%),88 and cumulative MMR rates in the nilotinib arms were significantly higher compared with imatinib (66%, 62%, and 40%, respectively; P < .0001).88 Finally, fewer progressions to advanced-phase disease occurred in the nilotinib arms (<1% on either nilotinib schedule vs 4% for imatinib).88

The Dasatinib versus Imatinib Study in Treatment-Naive CML (DASISION) trial compared dasatinib and standard-dose imatinib (400 mg) in newly diagnosed patients (N = 519).31 After a minimum follow-up of 12 months, the rate of confirmed CCyR by 12 months was 77% with dasatinib and 66% with imatinib.31 The CCyR rate by 18 months was superior in the dasatinib arm compared with the imatinib arm (78% vs 70%; P = .0366). The MMR rate also was greater in patients who received dasatinib compared with those who received imatinib (57% vs 41%; P = .0002). In total, 2% of patients who received dasatinib and 4% of patients who received imatinib transformed to advanced disease.89

Both dasatinib and nilotinib are well tolerated in the first-line setting. Most drug-related adverse events were grade 1 or 2 in severity, and grade 3 and 4 nonhematologic events were rare.89

Recently, data from the US and Canada Intergroup phase 2 study of dasatinib versus imatinib in newly diagnosed patients with CML-CP (S0325) became available. In that trial, 253 patients were randomized to receive either dasatinib 100 mg daily or imatinib 400 mg daily. Preliminary analyses are impaired by an incomplete assessment of all cytogenetic data from the plethora of contributing centers. All outcomes, however incomplete, suggest a striking similarity to results from the DASISION trial. For example, CCyR rates at 12 months had the same trend between the dasatinib and imatinib arms (82% vs 69%).32

However, the DASISION, ENESTnd, and Intergroup trials have reported only short-term results. Although those results appear to be encouraging, there is no evidence to date that either BCR-ABL inhibitor offers a long-term difference in OS compared with imatinib. In addition, imatinib will gain generic status in many countries in a few years, thereby presenting it with an economic advantage over newer BCR-ABL inhibitors. Until long-term data are available, it is unclear whether the best way to treat newly diagnosed patients in CP, indeed, is to initiate and maintain therapy with a newer BCR-ABL inhibitor. Plausible alternative strategies include initial treatment with imatinib and switching to a newer BCR-ABL inhibitor if aggressive milestones of treatment response are not met or initial treatment with a newer BCR-ABL inhibitor followed by the use of generic imatinib as a cheap “maintenance” therapy in optimally responding patients.

In conclusion, the treatment of CML focuses on prolonging freedom from progression, which translates to long-term survival. CCyR is a landmark response and a positive predictive indicator of improved PFS and OS and can be used to assess the adequacy of treatment response in the front-line and second-line settings. Dasatinib and nilotinib are effective for treating patients who are resistant to imatinib, and there are data suggesting that these agents are more efficacious than imatinib in the first-line setting. In particular, molecular response rates appear to be higher with these newer BCR-ABL inhibitors compared with imatinib. This is potentially important, because several ongoing studies are addressing whether patients in CMR can discontinue therapy. In general, in these studies, roughly half of patients who were removed from BCR-ABL inhibitor therapy retained a CMR with follow-up of 1 to 2 years. Although the long-term durability of these off-therapy remissions is unknown, the results suggest that a subset of patients with prolonged CMR may be “cured.” If so, then the more potent BCR-ABL inhibitors, coupled with careful molecular monitoring, may play increasingly important roles in the management of CML. However, the unclear, long-term advantages of dasatinib and nilotinib over imatinib and the future generic status of imatinib complicate this scenario. Careful studies of these issues should both improve the care of CML patients and build the framework of targeted therapeutic approaches for other malignancies.


Editorial support was provided by StemScientific, funded by Bristol-Myers Squibb.


The author made no disclosures.